Podcasts about ETL

We've been running a bit of an Agent Cloud series surveying all the top inference/compute/cloud providers, from Databricks to Daytona to Railway and, even further back, E2B, but we're excited to conclude this series returning to Modal, which has just raised a monster $355M Series C.The cloud was built for developers. But agents are now changing that.The old infra stack was designed for a human who could read docs, reason through YAML, and understand dashboards to figure out what they need when something broke. While this was painful for developers, it worked since they could fill in missing context in their heads.However, agents don't have that luxury. Now in this new era of agents, everything has to be tighter.They need a place to write code, run it, inspect the output, change the environment, debug failures, and try again. Fast iteration and feedback loops with all the necessary context are crucial for agents to operate properly. Furthermore, sandboxes are a clear representation of this shift as agents can easily spin up isolated environments. This programmatic infra even extends to research:Two years ago, we were one of the first to cover Modal with CEO Erik Bernhardsson and Alessio designed our favorite LS thumbnail of all time:At the time, Modal was just a teeny little company with a $17M Series A.Today, fresh off their $355M Series C, Modal is one of the clearest examples of the agent cloud future being built in real time: a cloud platform moving past traditional web app assumptions toward the workloads AI actually creates such as elastic inference, sandboxes, GPU burst, post-training, background agents, and infrastructure that agents themselves can operate.In this episode, Modal CTO Akshat Bubna joins swyx and Vibhu to unpack why AI applications don't fit traditional cloud assumptions, why Kubernetes was never designed for bursty compute-heavy workloads, and why Modal is now shifting from developer experience to agent experience.We go deep on Modal's AI infra stack: serverless functions, decorator-based infrastructure, elastic inference for custom models, GPU snapshotting, DeFlash, speculative decoding, Auto Endpoints, sandboxes, persistent storage, networked containers, private IPv6, RDMA, multi-node training, and Modal's capacity pool across 17 cloud providers. Akshat also explains why RL rollouts can require 100,000 sandboxes, why production agents need hard guardrails, why observability may matter more than reading code, and why AI has made infrastructure exciting again.We discuss:* Why Kubernetes wasn't built for bursty AI workloads* How Modal started as a better runtime before becoming an AI cloud* Why Modal added GPUs before ChatGPT* The shift from developer experience to agent experience* Why observability matters when agents are writing the code* Elastic inference for custom models across audio, video, robotics, and comp bio* GPU snapshotting, cold starts, and why inference workloads are so bursty* Why RL rollouts can require 100,000 sandboxes* DeFlash, speculative decoding, and frontier-level inference performance* Auto Endpoints and making optimized inference easier to deploy* What Modal adds beyond vLLM, SGLang, and raw GPU rental* Modal's 17-cloud capacity pool and supercloud strategy* Networked sandboxes, sidecars, private IPv6, and RDMA* Serverless multi-node training for post-training and research workloads* Auto-research, model-guided sweeps, and agents launching GPU experiments* Compute strategy, capacity planning, and batch tiers* Why production agents need specialized sandboxes and hard guardrails* Modal's take on managed agents, CI, Gitpod/Ona, Python, TypeScript, and Modal BenchAkshat Bubna* LinkedIn: https://www.linkedin.com/in/akshat-bubna-188885103* X: https://x.com/akshat_bModal* Website: https://modal.comTimestamps00:00:00 Introduction00:00:39 Modal's origin and why Kubernetes wasn't enough00:04:32 Developer Experience → Agent Experience00:06:21 Modal's AI cloud primitives00:09:14 Sandboxes, agent loops, and proto-Cognition00:12:12 Elastic inference, GPU snapshotting, and 100,000 sandboxes00:15:24 DeFlash, speculative decoding, and Auto Endpoints00:19:59 Production-grade inference beyond raw GPUs00:22:00 Background agents, Ramp Inspect, and the agent lifecycle00:24:08 Modal's 17-cloud supercloud strategy00:26:40 Networked sandboxes, private IPv6, and RDMA00:32:48 Multi-node training, post-training, and auto research00:37:36 Compute strategy, capacity planning, and batch tiers00:40:55 Open models, real-time AI, and production agent infra00:43:06 Hard guardrails, managed agents, and specialized sandboxes00:46:06 Why AI made infrastructure exciting again00:48:30 Model APIs, differentiated products, and agentic video00:51:50 CI, coding-agent infra, SDKs, and Modal Bench00:57:28 Closing ThoughtsTranscriptIntroduction: Modal, Series C, and the Art PartySwyx [00:00:00]: We're here with Akshat, CTO of Modal, together with Vibhu. Congrats on your Series C.Akshat [00:00:10]: Thank you.Swyx [00:00:11]: Your party yesterday was amazing.Akshat [00:00:15]: Yeah.Swyx [00:00:15]: From all the photos and all the swag.Akshat [00:00:17]: We had a bunch of art installations, which was fun, seeing, like, our products on pedestals next to, like, Rodin.Swyx [00:00:25]: Very nice. Very nice. When you started, it was not the GPU inference company. Maybe it was in your mind. Take us back to the origin story.Modal's Origin: A New Runtime Beyond KubernetesAkshat [00:00:39]: I first met Eric, who's the CEO, through an investor. Back then Eric was already thinking about building, a new runtime, and he got there thinking through why are workflow orchestration products so hard to use. It's because you have to run them on Kubernetes. Kubernetes is hard to manage. It's not built for burstiness and, custom images,Swyx [00:01:03]: YeahAkshat [00:01:03]: It has a terrible developer experience.Swyx [00:01:05]: And I'll, I'll interjectAkshat [00:01:06]: YeahSwyx [00:01:07]: For listeners, who are new, we interviewed Eric two years ago, and there's a bit more of the story there from Spotify and all those things.Swyx [00:01:14]: And I came across Eric through Data Council because he did that talk on the serverless container stack that you guys did, which was like, that was my first like, “Okay, I need to take Modal very seriously” moment.Akshat [00:01:26]: Yeah.Swyx [00:01:26]: But it was still very unclear, like, do I need all this for just my data pipelines?Akshat [00:01:33]: Yeah. initially what we were thinking about was if we build a better runtime, it's a very useful primitive in itself. It's There's a lot of things that, get solved by serverless functions, like you can do, ETL stuff, you can do job queues, you can do all this, like, bursty processing, which it turns out every company had needs for. but then we also were thinking about this as like, this is a primitive that we can build a whole collection of products on, which are very verticalized. So perhaps data engineering would've been the first one, but we were thinking about inference. Back then it was more classical inference, like computer vision stuff and running XGBoosts and whatnot. But we added GPUs to the product a year before ChatGPT came out.From Serverless Containers to GPU WorkloadsSwyx [00:02:19]: Nice.Akshat [00:02:19]: We just didn't think it would be that big of a deal.Swyx [00:02:22]: Yeah, just like add A100.Vibhu [00:02:23]: Was there any, like, early key problem that really sparked off why you built it?Akshat [00:02:28]: Yeah. Primarily it's just, none of the tooling that was out there was built for, one, a really great developer experience, and also there's a general trend of, a lot of the workloads that we were seeing were very. I wish there was a better word for it, but compute-heavy. Like, they need, one, like, need a lot more resources, so you need to burst up and down a lot, versus like Kubernetes designed for, like, slow scaling and, more for, like, web server use cases. And also there's just a lot more specialization in, like, what kinds of environments these workloads run in. Like, we had sometimes they need accelerators, sometimes they need different kinds of images, and this is just like a consistent thing that we saw across a lot of companies. That would be the next step.Software-Defined Infrastructure and Decorator-Based DXSwyx [00:03:13]: Yeah. Yeah. Be nice. I don't know how much this factored into the early story, but I wrote a post when I was at Temporal about infrastructure, software-defined infrastructure or something like that.Akshat [00:03:22]: Yeah, the self-provisioningSwyx [00:03:23]: Self-provisioning.Akshat [00:03:24]: Yeah.Swyx [00:03:24]: Yeah. I can't even remember my own post.Swyx [00:03:26]: And then you put me on the landing page.Akshat [00:03:28]: Yeah. We really like, the term and so we stole it.Swyx [00:03:32]: Because you had the insight that everything can just be in decorators co-located with the code, right?Akshat [00:03:37]: Yeah.Swyx [00:03:37]: Was that a big part of the originalAkshat [00:03:39]: YesSwyx [00:03:39]: Story or it was just like a DX layer?Akshat [00:03:41]: That was, really important because we really didn't want people to spend, so much time, writing YAML, and it seemed like you could really condense the surface area of what you're doing, put it in code so you can operate on it just like you operate on other code, and like build stuff that's more expressive and dynamic. and so yeah, that was always a very important part.Swyx [00:04:04]: Then the pushback is this is a DSL.Akshat [00:04:07]: Yeah.Swyx [00:04:07]: It's you're closed source. I am locked into Modal.Akshat [00:04:11]: Yeah. We never really got pushback for that because the nice thing about Modal is you can bring whatever code you have, and sure, the DSL is at the configuration layer for, what hardware you're using, how you're scaling things up, but you still own the code.Akshat [00:04:27]: And that's, that's been an important, part of our story, even as we do inference now.Swyx [00:04:32]: Yeah.Vibhu [00:04:32]: How much of do you think still stays the same today? Like if you were to build something today, DevX very important, but I feel like, a lot of this has been changed with just hook it up to an agent, have Claude Code, have Codex implement a tool. there's very agent native primitives that are different than if I'm doing this myself, right?Developer Experience → Agent ExperienceAkshat [00:04:54]: We've changed our SDK team to think about agent experience instead of, developer experience and we think that the same benefits that apply for DX also apply for AX, which is why would you have an agent read through hundreds of Kubernetes files and like write YAML that's not even typed when it can make a couple of changes in a decorator and it gets this self-provisioning runtime of, being able to see its changes live in action? yeah, it just seems from the customers we talk to, they find Modal is much faster for agents to use versus operating on a different substrate.Swyx [00:05:34]: Yeah, because like you, again, you co-locate the infrastructure requirements to the code that runs it.Akshat [00:05:38]: Yeah.Swyx [00:05:38]: Well, the negative thesis now is that nobody's looking at their code anymore, so there's no point.Akshat [00:05:44]: Yeah, people aren't looking at code. one thing we still see is really important is observability.Swyx [00:05:51]: Yeah.Akshat [00:05:51]: Like how good is your dashboard? And of course, like we have, we push a lot of it to the CLI so the agents can do their own investigation, but you still need humans to go interpret what's going on and, make judgment calls and whatnot. and that's I feel like, Maybe more important now than looking at the code itself.Swyx [00:06:11]: Yes, because like, you can try to treat the code as a black box and then use, see the observable action that comes out of it, and then just prompt a change.What Modal Is For: AI Cloud PrimitivesAkshat [00:06:21]: Yeah.Swyx [00:06:22]: So I think it takes a bit of restraint to not specialize, to say, “I want to ship a new primitive,” and then just be general purpose.Swyx [00:06:31]: People ask you, “What are you for?” You're like, “ I don't know. We can do this, we can do that.”Vibhu [00:06:36]: Well, I'd be curious to see, like, okay, if we were to ask you, like, what is Modal for even at a high level? There's a lot you guys do, sandboxes, GPUs, everything. How do you answer?Akshat [00:06:46]: Modal is a cloud platform that's built for, where we've built the primitives from scratch for AI applications. and right now it covers, inference, training, batch processing, and sandbox workloads.Akshat [00:07:00]: But we're building a lot moreSwyx [00:07:02]: I noticed you didn't say web server, so there is still a role for, like, the always-on large-scale Kubernetes type things.Akshat [00:07:09]: Yeah, absolutely. We're, we're not trying to compete with the renders of the world, because yeah, we think the differentiator for us is the, are the workloads that need specialized compute, need to scale up and down a lot. yeah, they're, they're, they're just shaped differently.Working Alongside Frontier StartupsVibhu [00:07:26]: I think you're building a lot of it alongside the startups, right? They're innovating quite a bit, even in your, like, latest blog post. Like, even in the series C, the customers that you mention here, the cognitions, technical ones, ramps and whatnot, they're, they're innovating with you, right? And that's not something AWS is doing directly with.Akshat [00:07:45]: Yeah, absolutely. I think, this is again classic. We're a small team. We can move really fast. our engineers are working with our customers and figuring it out. Yeah.Swyx [00:07:54]: So my first week at Cognition, I walked in, there was someone wearing a Modal shirt. I was like, “What are you doing here?” They're like, “Yeah, I just. I am embedded inside of Cog.”Akshat [00:08:05]: Yeah, I think that was Peyton. We sent him overSwyx [00:08:07]: Yeah.Akshat [00:08:07]: Because, the latency of communication was too high otherwise.Swyx [00:08:12]: Yeah, distributed node, you have to - you have to place one and collocate.Vibhu [00:08:16]: Yeah.Swyx [00:08:16]: So I had a, I had direct personal experience, right? So I worked on smol developer three years ago. it was inspired by Claude 1. I think you onboarded me at some point, like, just before, and I was like, “Oh, like, I need some bursty compute. Like, I was just gonna try using Modal.” And it was a, it was a pretty pleasant experience. apparently, I showed up in the board meeting, like the analytics.smol developer, Sandboxes, and Proto-CognitionAkshat [00:08:39]: Yeah, you blew up on Hacker News and,Swyx [00:08:41]: YeahAkshat [00:08:41]: We got a big traffic spike. I. I think the way you used smol developer was Modal functions for running stuff, which was. Like, the, that was a good use case. but then, yeah.Swyx [00:08:53]: Yeah. That - So to me, that was proto-cognition.Akshat [00:08:55]: Right.Swyx [00:08:56]: If only I had, like, stuck to it.Swyx [00:08:58]: Like, that was like, if - did you say draw the tech treeAkshat [00:09:00]: AbsolutelySwyx [00:09:00]: You're just like, “Yeah, like, probably this will happen.”Akshat [00:09:02]: Yeah. Like, he was so close. You were just rebuilding upon usSwyx [00:09:04]: I just didn't realize.Akshat [00:09:05]: But the funny story there is at the same time, we were talking to a bunch of customers who needed something like sandboxing.Swyx [00:09:14]: Yeah.Akshat [00:09:14]: This is like twenty-three.Swyx [00:09:15]: Yeah.Akshat [00:09:16]: So we builtSwyx [00:09:17]: You introduced a new API right after that.Akshat [00:09:18]: Yeah.Swyx [00:09:19]: Yes.Akshat [00:09:19]: Like, we built sandboxes in May of twenty-three before anyone was even knew this was gonna be a thing. And the first example we published was, we took smol developerSwyx [00:09:28]: Smol developerAkshat [00:09:28]: And put it in a loop, so the agent can iterate on itself.Swyx [00:09:33]: Loops are hot these days.Vibhu [00:09:34]: It's the looper.Akshat [00:09:34]: Yeah.Vibhu [00:09:35]: Loops in. When was this, twenty-three?Akshat [00:09:38]: Yeah.Vibhu [00:09:39]: A small check.Akshat [00:09:39]: Yeah.Swyx [00:09:39]: It's like twenty-three. so the. the, those for listeners, like, the problem was the models are not built for any of this, right?Swyx [00:09:46]: Like, you're just trying to like. They're not post-training to understand, like, looping and, like, self-correction and tool calling was there, but, like, also not that great.Akshat [00:09:55]: Yeah.Akshat [00:09:55]: I don't remember if you used tool calling in this one, but yeah, the models would just diverge after like ten iterations and not produce anything meaningful.Swyx [00:10:03]: Yeah. But like, then. So okay, like now talking to myself three years ago, the answerVibhu [00:10:08]: Of course they will get betterSwyx [00:10:09]: Collect all the failures, build benchmark, and then collect all the, examples, build the RL environmentAkshat [00:10:15]: RightSwyx [00:10:15]: Sell it for like ten billion dollars to Meta.Swyx [00:10:17]: And then also train a model and then sell that for sixty billion dollars to Elon. And this isAkshat [00:10:23]: Yeah, of courseSwyx [00:10:23]: The funny machine. Like, it's like, it's about the hardware.Akshat [00:10:28]: It's hard to have that inherent conviction that the stuff will get that much better.Swyx [00:10:33]: In retrospect, it's so f*****g obvious.Akshat [00:10:36]: Fair enough.Swyx [00:10:37]: Like, what else were we doing back then? I don't know. anyway. Yeah. So this. That was the start of your sandboxing journey, right? I feel like it didn't blow up until, like, last year.Akshat [00:10:49]: Yeah.Swyx [00:10:50]: So there was like a couple years of quietness.Akshat [00:10:52]: Exactly, yeah. We wereVibhu [00:10:53]: I think very underrated product value. Like, my experience with Modal, Charles, before he had joined Modal, met this guy at a hackathon, and he really insisted we wanted to run some small model, not hosted anywhere, and he's like, “ there's this cool company, Modal. They'll like spin up a GPU sandbox, we can throw it on there. They'll take a Hugging Face link.” And like there's so much value just right there, right? Like instant hosting, spin it up, spin it down. It'll stay cold, but we run the demo a few days later, it'll come back up and like all this stuff in retrospect, like it's still what we needed like today.Akshat [00:11:27]: Yeah, it's still needed today. workload shapes have changed a lot as, we run stuff for people with really massive production scale and, there it's it's not about scaling from zero to one, but it's how do we scale really elastically, from like thousand to fifteen hundred GPUs very quickly in a given region. It's the same shape problem.Elastic Inference, GPU Autoscaling, and Custom ModelsVibhu [00:11:50]: Okay. So you look at, say, Cursor Composer, right?Akshat [00:11:53]: Yeah.Vibhu [00:11:53]: They had a. “We'll do RL on a model every couple hours.” you guys have a whole version of RL inference gym and whatnot.Vibhu [00:12:01]: When you look at workloads like that, you're doing train runs where you need to scale up, scale down every hour thousands of GPUs, right? That's the example for we do need it, right?Akshat [00:12:12]: Yeah. Well, so I'll, I'll take a step back and, maybe talk about like how people use Modal today. because our biggest use case is, elastic inference. And the thing we first found product market fit, with was inference for custom models. So we stayed away from the LLM space, and we were serving companies like Suno for audio, Runway for video, robotics, comp bio companies that train their own model elsewhere. But Modal is the best black box that for deployment, scaling to however many GPUs you need as your traffic pattern changes. And we saw all of them like have a very unpredict- predict- predictable, traffic pattern. it's like diurnal. It's Some days, like the company will do a launch and, they'll need like, way more. And it's not just one model that they deploy. They-- all these companies deploy, lots of different models in different regions, and so the autoscaling problem becomes even harder because then you have to scale within a certain region, and those cycles are offset. So different times you scale up in different regions.Akshat [00:13:20]: So that's like our sortVibhu [00:13:22]: And thatAkshat [00:13:22]: YeahVibhu [00:13:22]: That in and of itself is a huge category. There's a bunch of inference providers which, provide this fireworks, does this as a service together, whatnot, Base10. that's carved into its own niche for language models, at least right now.Akshat [00:13:36]: Yeah. the thing that we have specialized in is the autoscaling aspect.Vibhu [00:13:41]: Yeah.Akshat [00:13:41]: Because we found that it's not universally true that everyone else can autoscale, and we've gone deeper into it on the tech side by, we've incorporated GPU snapshotting into the product so we can take the GPU state, like your torch.compile model, snapshot it, and the next cold start is way faster. And so going back to your question, it's That's why you need a lot of burstiness for inference. But then people also do a lot of demand training, like for RL stuff, your rollouts are bursty, as you said. People also do a lot of batch jobs. So we'll see, a lot of companies, before they have a training run, they'll need thousands of GPUs to run encoding or something like that. And I think those things are much more bursty than. I agree that agents are not that bursty. sandboxes are, except when you're doing RL. RL is justRL, Batch Jobs, and 100,000 SandboxesVibhu [00:14:28]: Or commerceAkshat [00:14:28]: Insanely bursty.Vibhu [00:14:29]: Yeah.Akshat [00:14:30]: Yeah. Like when you're doing, rollouts, you sometimes need a hundred thousand sandboxes in your sandboxes.Vibhu [00:14:37]: Yeah. I'm curious if you've seen early sparks of continual learning. There are some people, like our friends, ngram, recently announced thisAkshat [00:14:45]: YeahVibhu [00:14:45]: They're, they're trying to do training. That also seems like a different workload, right? If you're doing training twenty-four/seven per se, there's a very weird dynamic of how you're using GPUs between people and whatnot, but seems like something you guys would work for.Akshat [00:15:00]: As you said, we're, we're fortunate to work with a number of, customers at the frontier and grab some of our customers. and they are taking the primitives we have, and trying to use them in very interesting ways, like continual learning. It's possible as the stuff gets better, some of that will be part of, our offering as well if, more people need it. but we're, we're just waiting to seeVibhu [00:15:23]: YeahAkshat [00:15:23]: How it shakes out.Vibhu [00:15:24]: Is there a primitive that you added after sandboxing that was the next step in the story?LLM Inference, DeFlash, and Speculative DecodingAkshat [00:15:32]: I guess we've been going much deeper into LLM inferenceVibhu [00:15:35]: YeahAkshat [00:15:35]: Because we realized that some of the advantages we have with like autoscaling, again, especially in different regions and whatnot, are, not present elsewhere. and the place where we had a gap was we weren't, working on the model layer itself. Like we were a black box. And, we realized that, we can get to frontier-level model performance, with, by having great people who work on this. And, we've been open sourcing a lot of our work, in terms of, Recently, we, shared our work on DeFlash, which is a block-based, speculator, and we've open sourced, all of it. So, you can - By using open source DeFlash, you can get the same performance as you would with one of the proprietary providers. And the next thing we're thinking about hereVibhu [00:16:23]: I thought this wasAkshat [00:16:24]: YeahVibhu [00:16:24]: An interesting blog post as well, right? Like, I think in here you make a claim that. Not a claim, just that how effective speculative deco-decoding really just get to.Akshat [00:16:33]: Yeah.Vibhu [00:16:33]: Anything you wanna point out from this around, what people should know?Akshat [00:16:39]: Yeah, absolutely. the high-level summary is, it would help to describe what speculative decoding is.Vibhu [00:16:44]: Yes.Akshat [00:16:44]: I will, yes.Vibhu [00:16:45]: I think, likeAkshat [00:16:46]: YeahVibhu [00:16:46]: So we've covered like Eagle and all thisAkshat [00:16:47]: YeahVibhu [00:16:47]: Like Hydra and all those things, but it was like two years ago.Akshat [00:16:51]: Yeah.Vibhu [00:16:51]: I think it doesn't hurt, right?Akshat [00:16:52]: Yeah. Speculative decoding is you have a smaller model, called a draft model, predict tokens ahead of the bigger model, and then you have the bigger model, verify all of this, all the tokens are predicted. And the reason it's faster is if you're predicting, one token at once, you're bound by memory bandwidth. But if you can batch the verification of, the draft model, then you're much more efficient using compute, and it's faster, and as long as your draft model is producing a lot of tokens that can get accepted, which is called the accept length, you can get a speed up that's, multiple times of, the original model speed. and well, that's what we highlight here. It's Like people talk a lot about we made these kernels faster and whatnot, but improving kernel will only give you like few percentage points of improvement, and, increasing accept length, literally is a multiplicative decreaseVibhu [00:17:47]: Like two to four X.Akshat [00:17:48]: Yeah, exactly.Vibhu [00:17:48]: Without much head-on performance.Akshat [00:17:50]: Yeah. I think it may - you are running a second model, right? So it may be something more expensive in the compute,Vibhu [00:17:57]: I meant quality performanceAkshat [00:17:58]: Probably not by muchVibhu [00:17:58]: But yeah. I thinkAkshat [00:17:59]: So there's no drop in quality performanceVibhu [00:18:01]: YeahAkshat [00:18:01]: Because you're always. You're never accepting a token that the big modelVibhu [00:18:04]: It's strictly betterAkshat [00:18:05]: YeahVibhu [00:18:05]: Or it's same.Akshat [00:18:06]: Exactly.Vibhu [00:18:07]: Right. Yeah.Akshat [00:18:08]: And so we've been working a bunch on DeFlash, which is a block-based speculator. so it's instead of predicting, one token at a time, it's predicting a block. And we've been open sourcing our work with it. The next thing for us here is for helping people train speculators and custom models. it's it's something that traditionally is very forward-deployed engineering driven, support deployed, engineer driven, like you work with customers and help them do that. And our vision for. This is why we launched Auto Endpoints, is we want to make frontier-level performance available to everyone. And so, we mentioned this in the announcement, we teased it. The next thing we're, we're launching is, as you run an auto endpoint, we shadow trafficAuto Endpoints and Frontier-Level PerformanceVibhu [00:18:54]: Do you want to explain what auto endpoints are?Akshat [00:18:57]: Yeah.Vibhu [00:18:57]: I lovely, yeah.Akshat [00:18:58]: Yeah. So, this is, I guess, going back to your Modal is you touch the code, but, sometimes people don't wanna touch the code, and they wanna get started with an endpoint that works and has all the great performance and, scalability that Modal has. So we've made that easier with, a way to create an endpoint from our UI, from the CLI, that has all of our optimizations that we talked about, like the DeFlash stuff already baked in, and there's full transparency. So we give you the code, you can go run it yourself, and if you want, you can eject out into the full Modal experience, which we see as people get sophisticated, they do wanna tweak the models, they wanna, fine-tune stuff. You can still do all of that. It's it's not a black box. And yeah, the next thing, as we teased later in the post, is how do we give you value even beyond this in terms of having your draft models evolve as your data distribution evolves, again, without having to talk to a person and, yeah.Vibhu [00:19:59]: I guess just to understand it directly, you have the GPUs, you have an endpoint that's compatible, you serve open model. If someone was to do this themselves, what's the delta that you guys provide? So you do a lot of open source great work on effective inference. how does it compare to, say, I take the same model, 5.2 FP8, take shelf inference engine, vLLM, SGLang, get compute of similar capacity, similar cost. What's the delta that plugging into something this, like this offers outside of the benefit of, scaling?Production Inference Beyond Raw GPUsAkshat [00:20:34]: It's interesting because we've taken the approach of open sourcing our contributions and upstreaming them. we work closely with the SGLang team. We want the improvements that our team, comes up with to be, there in open source for others to use, even outside of Modal. The benefit to us is we have a team that has significant expertise in terms of if you do have something that is not there, our team can help you get that performance, first. the other thing is with these endpoints, we are way more elastic, as you said, than, anyone else, and you have true scaling to zero. you have true, burstiness, and in practice, that matters a lot more to people than just finding, the GPU and, running Modal code on something.Vibhu [00:21:20]: Yeah. And I will say it's not that straightforward to just. like what I said is easier said than done, right?Akshat [00:21:26]: Yeah.Vibhu [00:21:27]: It's I think still for the average person, still hard to just gut check using different. There's, there's quite a bit of combinations you can make there. the trade-offs aren't really known at face value.Akshat [00:21:40]: Yeah. it's it's not just that. I think it's it's that running production-grade inference is a hard infer problem.Vibhu [00:21:49]: YeahAkshat [00:21:49]: Even if you subtract out the autoscalingVibhu [00:21:50]: YeahAkshat [00:21:51]: Is controlling things like tail latency and, making sure every, request is delivered at least once and whatnot.The Model and Agent LifecycleVibhu [00:22:00]: There's a lot of innovation that you can do here. I think, it's very interesting that you're starting to encroach on, like as you become a full cloud, you're starting to encroach on other people's turf.Vibhu [00:22:09]: What will you not do?Akshat [00:22:13]: Well, we wanna follow our users and, make sure they get like a platform that has everything that works well together. so right now we're focused on the model lifecycle and the agent, lifecycle. so both like going from data prep to training to inference, and then also if I want to deploy a background agent, let's say, sandbox, do persistent storage, a whole bunch of other stuff.Vibhu [00:22:38]: We talked to Cole, who did, OpenInspect. Yeah.Akshat [00:22:42]: Yeah.Vibhu [00:22:42]: And RealInspect also is on Modal.Akshat [00:22:44]: Yeah. So Ramp Inspect was a great example of a background agent that was really successful because they, were able to use some of the primitives like snapshotting and fast scaling to just have something that feels really reactive and works well.Ramp Inspect and Background AgentsVibhu [00:23:02]: Yeah. That's the new CTO of, Ramp right there.Akshat [00:23:05]: Yeah, Rahul.Vibhu [00:23:08]: It was really fun. yeah, okay, I think, all very bullish. Like, one of my reflections was also I did not originally. So when I met you guysThe Inference Inflection: CPU, GPU, and Co-LocationVibhu [00:23:19]: You weren't that much in the GPU game, and now you're all about, inference. And one of the points that I hinged on for Jensen's keynote at GTC this year was, what we're calling like the inference inflection, right? That let's say in AI workloads or machine learning workloads, it used to be like, let's call it eight to one GPU to CPU, and now it's more like one to one, which is like a interesting. Like, - because of how much agents are blocked or call out to this, to CPU heavy stuff the actual, like, limiting factor, like, swings back and forth from GPU to CPU a lot more than it used to be all GPU and then occasional CPU.Akshat [00:24:01]: Yeah.Vibhu [00:24:02]: GPU, CPU. And now it's like just constantly, and you just have to locate everything.Seventeen Clouds and the Supercloud StrategyAkshat [00:24:08]: Yeah. And that's one of the things that, again, we see as, something appealing about Modal, which is we've built this capacity pool that spans, 17 cloud providers, so we're, we're very good at Running on various kinds of cloud capacity across the worldSwyx [00:24:24]: You don't have your own data centers?Akshat [00:24:25]: We don't have our own data centers. We just run across a lot of neo cloudsSwyx [00:24:29]: Yeah. AreAkshat [00:24:30]: Metal providers.Swyx [00:24:30]: Yeah. Question mark.Swyx [00:24:31]: Yeah. You're, you're running the math, and you're like, “What's the cutover point where you're like.”Akshat [00:24:36]: Yeah, it's a good question. part of it is we see our differentiator in the software layer, and, being capital light and focusing on the software helps us move really fast. so far it's worked out well because there are so many other people building data centers that we're able to work effectively with them, and again, focus on what makes us, special.Swyx [00:24:55]: Yeah.Swyx [00:24:56]: 17 gets you into, like, the local providers sometimes. LikeAkshat [00:25:00]: The,Swyx [00:25:01]: Which was the most interesting one?Akshat [00:25:02]: There are a lot more neo clouds than you expect, and they all have various degrees of, various levels of reliability. And, that's why it's something we've invested a lot of time in, is building our own reliability layer on top. so if the GPU falls off the bus or something happens, we user workloads are not affected, and that lets us use a lot more capacity than,Swyx [00:25:30]: YeahAkshat [00:25:30]: You as a user would be able to.Swyx [00:25:32]: It's a useful thing to have because like now everyone knows, like, what layer you are and, like, you optimize for being the super cloud of all clouds.Akshat [00:25:41]: Yeah. That's, that's, that's the idea. and so I guess when you mentioned colocation, that's, that's another interesting thing where, one thing we've seen is people come to us when they want, very specifically located, CPUs or GPUs, like they wantSwyx [00:25:57]: Oh, they pin it in likeAkshat [00:25:58]: YeahSwyx [00:25:58]: EU?Akshat [00:25:59]: Exactly. Or EU, US.Swyx [00:26:01]: Right. Data resiliencyAkshat [00:26:02]: AustraliaSwyx [00:26:02]: Locality thing or performance or what?Akshat [00:26:04]: It's either data locality or latency, yeah.Swyx [00:26:07]: Yeah.Akshat [00:26:07]: Like, you want your. They're running sandboxes and model. They want them to be right next to aSwyx [00:26:10]: Yeah, it's easy thenAkshat [00:26:11]: YeahSwyx [00:26:12]: To. That is important in all those things. and so, like, you've accidentally, I don't know if it's accident, but, like, you've built the perfect primitive for agents to express themselves. And then, like, it's almost very funny how every extra development just involves more file system, just involves more CPU.Akshat [00:26:30]: Yeah.Swyx [00:26:31]: Just like the things that you already have. I don't know much about, if there's any, like, networking usages that are interesting, but you've also done some good work on networking.Networking, Sidecars, Private IPv6, and SandboxesAkshat [00:26:40]: Yeah, that's exactly right. Like, we're just taking compute storage and networking and building stuff on that layer, for, again, the stuff people need.Swyx [00:26:49]: YeahAkshat [00:26:50]: We see a few interesting networking things coming up. one is people want networked sandboxes. so we haveSwyx [00:26:57]: For like a Docker cluster type thing.Akshat [00:26:59]: Yeah.Swyx [00:26:59]: Sorry, Docker Swarm. Oh, f**k. What is it called?Akshat [00:27:02]: Compose.Swyx [00:27:03]: Compose type thing.Akshat [00:27:04]: Yeah. So if you want Docker Compose, our sandboxes now support, this thing called sidecars. So you can. A sandbox is a pod of containers, and you can run multiple containers in, a sandbox. also useful because, going back to networking, people want a lot of control over, outbound networking from a sandbox.Swyx [00:27:23]: Yeah.Akshat [00:27:23]: Like, they might wanna run a middle proxy for, like, maybe logging stuff for RL or, controlling how egress can happen to a domain, injecting credentials. and yeah. So we've, we've had to build a lot of that stuff ourselves.Swyx [00:27:38]: Yeah.Akshat [00:27:39]: But then also sometimes people want, sandboxes spanning multiple nodes to talk to each other, which is an emerging thing we're seeing. We have support for that for a different reason, and yeah, we'll see if that becomes stable.Swyx [00:27:52]: Like, just an open socket. It's a. This is directly like mTLS.Akshat [00:27:56]: We do support that, which is you can, expose a tunnel inside a sandbox.Swyx [00:28:01]: Yeah.Akshat [00:28:01]: And then you can either expose it to public internet or it can be, you can add like a HTTP, auth layer above it. But we have this thing called I6PN, which we haven't talked about, which is this, like, overlay network using IPv6 addresses. so if Modal containers, within the same workspace, when this is enabled, can address each other using this private IPv6 address, and no one else can.Akshat [00:28:28]: So it's like private networking, for containers. We built it because we needed it as a primitive for our distributed training product. so we have this other feature, which is you can add a decorator to a function, and you get a cluster of GPUs. and they have RDMA networking. so you can run a distributed training job, that's truly serverless. and we did the overlay network for that. But then we've seen that people are using it for other reasons, and, I'm intrigued to yeah, what would people do with it.Swyx [00:28:59]: Build primitives and let people figure it out, right?Akshat [00:29:01]: Yeah, exactly.Swyx [00:29:02]: You put out a pretty interestingAkshat [00:29:03]: They're like, they read the docs webpage. Let me use thatSwyx [00:29:06]: YeahAkshat [00:29:06]: Something they never intended to work. This is literally not even in our docs page. People somehow found it, and they're using it.RDMA, Memory Movement, and Distributed TrainingSwyx [00:29:12]: Huh.Swyx [00:29:14]: The way you portrayed it with, like, RDMA versus TCP, like, very well laid out, but just the transfer speed change at scale for RL, like yeah, you have it, you have it built in. I'm sure someone found it. It's found it to be a lot more efficient before you made a thing out of it, right?Akshat [00:29:32]: Yeah. And not to split hairs, I guess the overlay network is the TCP overlay network.Akshat [00:29:39]: The reason we have that is you need that to do the key exchange for RDMA before you set up the RDMA network on top of that. but then people found the TCP part.Swyx [00:29:48]: Can I tell you, this is like a big aha moment for me becauseAkshat [00:29:51]: YeahSwyx [00:29:51]: So I review 2,200 submissions for the World's Fair.Akshat [00:29:56]: Yeah.Swyx [00:29:57]: And then I got this from John OsterhoutAkshat [00:29:58]: HuhSwyx [00:29:59]: Who I don't know if. Do John Osterhout by name?Akshat [00:30:01]: The name sounds familiar.Swyx [00:30:02]: He published a. He's a well-known professor, published a lot of interesting software design books, and this is the talk he chose to submit, is on RDMA at Inference. And I'm like, you wouldn't think that this guy, who is like operating systems guy, would care about RDMA.Akshat [00:30:20]: I, it makes sense to me because I,Swyx [00:30:24]: This is the cloud, right? YeahAkshat [00:30:25]: Like, the way you move around your KV cache and how efficiently you can do it, how efficiently you move, your weights from your training GPUs to your inference GPUs in RL is there's a lot of degrees of freedom, and it is a systems problemSwyx [00:30:41]: YeahAkshat [00:30:41]: Moving memory aroundSwyx [00:30:42]: YeahAkshat [00:30:43]: Scheduling.Swyx [00:30:44]: This shows you how primitive my understanding of networking stuff is.Swyx [00:30:46]: Is this like the domain of WireGuard as well?Akshat [00:30:50]: Not quite.Swyx [00:30:51]: It's adjacent?Swyx [00:30:53]: Explain everything.Akshat [00:30:54]: Sure.Swyx [00:30:56]: How do we move memory around GPUs?Akshat [00:30:58]: Well, so sorry. Yeah, that is memory. Sorry, I was talking more, and maybe I was talking like five minutes back, about the private IPv6, addressing that you've set up.Swyx [00:31:09]: Yeah.Akshat [00:31:09]: Is it like it's a VPN?Swyx [00:31:10]: Yeah, it is like a VPN, and yeah, WireGuard is, yeah, you're right. It is,Akshat [00:31:16]: Right. Yeah, you already moved on to new topicsSwyx [00:31:17]: A similarAkshat [00:31:18]: OkaySwyx [00:31:19]: In the same space, WireGuard is, encrypted and this is,Akshat [00:31:23]: And you don't need encryption.Swyx [00:31:23]: Yeah.Akshat [00:31:24]: Yeah.Swyx [00:31:24]: This is not encrypted. that's the main difference. This is TCP and we have eBPF programs that will reject or allow the TCP connection based on whether you're allowed to do it.Akshat [00:31:35]: Used to involve a full sidecar, but now you have eBPF in the Linux kernel.Swyx [00:31:39]: Yeah.Akshat [00:31:40]: Yeah. I don't know if this is a natural follow-on to the topic of like my skepticism on distributed training is that while, like, people spend a lot of money on, like, cables to hook up GPUs, and even that is not, like, fast enough, and that's the bottleneck, is your networking fast enough?Swyx [00:31:59]: Yeah. So I guess you're talking about fully distributed training like, Dialog or something which is like cross data centerAkshat [00:32:06]: That would be, yes.Swyx [00:32:07]: That's the extreme.Akshat [00:32:08]: Yeah.Swyx [00:32:08]: You're in the middle, and then other people would have like the Mellanox cables up in, like, their actual data center.Akshat [00:32:14]: When you run multi-node training on Modal, RDMA, I think Mellanox, is, or InfiniBand is like a, is all seen as RDMA. but it's a way to bypass the TCP networking stack and, transfer, stuff much faster, between one node, to the other. And we have I think like 3 terabit per second, internal networkingSwyx [00:32:40]: OkayAkshat [00:32:40]: Which is the standard that's needed.Swyx [00:32:42]: Okay. So I misunderstood whatAkshat [00:32:43]: 50Swyx [00:32:43]: What part of the stack you wereAkshat [00:32:44]: 50 gigs overSwyx [00:32:45]: YeahAkshat [00:32:45]: If you wentSwyx [00:32:45]: YeahAkshat [00:32:46]: RDMA.Swyx [00:32:46]: Okay.Swyx [00:32:48]: Yeah. I, very impressive work.Multi-Node Training, Post-Training, and Auto ResearchSwyx [00:32:52]: So effectively you're extending like the model philosophy to the training cluster, like, yeah.Akshat [00:32:59]: Yeah. And we're, we're not going for like large scale training runs. the thing that we've built multi-node training for is, we see a lot of, smaller scale post-training. like, people are post-training like medium sized fund models, so they can, get higher quality on inference. this is a perfect fit, for something like that.Swyx [00:33:21]: Yeah. That is my impression of how a lot of these labs explore branches in post-training and then eventually merge whatever they find in.Akshat [00:33:31]: Yeah. The other use case we've seen for multi-node training is even if you have a big cluster, your researchers are still doing small runsSwyx [00:33:38]: YesAkshat [00:33:39]: Having elasticity thereSwyx [00:33:40]: Right, sureAkshat [00:33:40]: Matters a lot more.Swyx [00:33:41]: Yeah. the, like, this is like the current limiting factor for auto research, which is like you need to give your model some GPUs in order for it to completely run.Akshat [00:33:51]: We have a blog post on auto resource and model is,Swyx [00:33:55]: YeahAkshat [00:33:56]: Yeah, like, turns out to be pretty good substrate for that.Swyx [00:33:59]: So my impression is auto research means many things, likeAkshat [00:34:01]: YeahSwyx [00:34:01]: Anything that Andrej coins. Right now it's still science fair, right? Like not like, I don't know how many people are doing this.Akshat [00:34:08]: We're having a golf.Swyx [00:34:08]: Yeah.Akshat [00:34:09]: I thought the same thing.Swyx [00:34:11]: Yeah, you would know.Akshat [00:34:12]: We, like, our internal both training and inference teams use this the general shape of this quite a bit. like we have this one internal repo called auto inference, which essentially we've automated our own forward-deployed engineering efforts using, this harness, which is, the agent will just spin up a sweep of different things. It'll even run like, NVIDIA inside profiler and it'll like tweak configs and it'll arrive the right thing. it'll change your GPUs both from H200 to B200, and works really well.Swyx [00:34:47]: Nice.Akshat [00:34:47]: So yeah.Swyx [00:34:48]: By the way, I enjoy that your forward-deployed engineering is so technical that you have to do these things.Swyx [00:34:52]: It's very different from forward-deployed engineering from other people.Akshat [00:34:54]: Yeah. For our forward-deployed engineering team is, essentially they're like applied inference researchers or applied training researchers.Swyx [00:35:02]: Someone told me like they have to be able to build, but they also have to be able to sell. do they have to sell or are they like they're good, they're just like post-sale type of thing?Akshat [00:35:09]: It does, being able to talk to a customer and engage effectively with themSwyx [00:35:13]: YeahAkshat [00:35:13]: Matters a lot.Swyx [00:35:14]: They want the same thing.Akshat [00:35:15]: Yeah.Swyx [00:35:15]: ?Akshat [00:35:15]: But it's it's not really a sales, thing. We pair them with-- We have solution architects as well that are more on the sales side.Swyx [00:35:23]: Okay. Let's spend a bit more time on auto research. This is a big focus for for this year. Where does this go? like, have people explored enough? Like, there's all these beautiful charts of like improve and then level off a bit and then you find the next thing. Is this one abstraction up from normal training? Is that how we think about it, or do you think about it differently? Like model level training versus high, like driven hyperparameter search.Auto Inference and Modal BenchAkshat [00:35:51]: Yeah, like,Swyx [00:35:51]: Someone, some people call it like neural architecture search or whatever, right? Like.Akshat [00:35:54]: Yeah, - So the stuff I've seen people do with it is nowhere on the architecture level. It's pretty much tweaking parameters, but it's it's a hyperparameter sweep that's guided by some model intuition, so it's like much more efficient than, whatever other, sweep you would have.Swyx [00:36:12]: Yeah, it's just, it's just a question of where you want to spend your compute?Akshat [00:36:16]: Right.Swyx [00:36:16]: ‘Cause yeah, you can just throw infinite amounts of money on this and somehow you'll bang out Shakespeare?Akshat [00:36:22]: Yeah, infinite monkey.Swyx [00:36:24]: Yeah, so like the very good for model. and I think it's also very important that agents can spin up other agents, can spin up their infrastructure. Like very good for you. how good is our LLMs at generating model code? Like the benefit of existing LLMs is that you are in the data.Akshat [00:36:42]: Yeah. They're, they're surprisingly good. I think like pre Cloud 4 they were not, and then now they're able to shot, stuff out of the box. But we're playing around with releasing like a Modal Bench for like the harderSwyx [00:36:55]: YeahAkshat [00:36:55]: Things, that the LLMs cannot do yet and maybeSwyx [00:36:59]: What's an example of that?Akshat [00:37:01]: I think the things that- Sometimes agents struggle with, without right guidance and a skill is, how to, use the rest of our observability. Like how to. Something is failing, like how do you look at the logs and then update the right thing? It's reasoning about that. But they're able to shot, likeSwyx [00:37:23]: Yeah. You can just add a skill to it?Compute Strategy and Capacity PlanningAkshat [00:37:26]: Yeah. So we have a Modal skill now that. Which is why we built this Modal Bench. It's to find things like that, so we can address them in our tool.Swyx [00:37:35]: Tune a skill. Yeah.Akshat [00:37:36]: Yeah.Swyx [00:37:36]: No. it's it's good. are you facing any shortages? like we talk a lot about GPU shortages, but also CPU, also memory.Swyx [00:37:44]: Yeah.Akshat [00:37:45]: We have had a lot of growth, which means that, there's - we've had to be much better aboutSwyx [00:37:53]: PlanningAkshat [00:37:54]: Proactive capacity planning.Swyx [00:37:55]: Yeah.Akshat [00:37:55]: So we have,Swyx [00:37:57]: Which by the way, like it's like a MBA's like dreamAkshat [00:38:00]: YesSwyx [00:38:00]: Is like just planning this stuff. I think last time you and I talked about something maybe about this.Akshat [00:38:03]: Yeah. we have a really competent team of people that we call, The role is called compute strategy. so yeah, if anyone listening here or wants to work on thatSwyx [00:38:13]: Compute strategy?Akshat [00:38:13]: Yeah.Swyx [00:38:14]: I think,Akshat [00:38:14]: I feel like,Swyx [00:38:15]: I think the normies call it FP&A or something.Akshat [00:38:18]: Well, it's more It's it's not FP&A. It's it's There's a lot of interesting financial questions of like what is the blend between one year and three-year reservations? how do we forecast our own capacity? how do we. especially since our capacity is very fungible across different GPU types and different regions, like you have to model a lot of it. and you also have to have an opinion on how the supply chain is gonna evolve, and then you have to like, take bets,Swyx [00:38:49]: YeahAkshat [00:38:49]: Based on that.Swyx [00:38:50]: Tokenomics.Akshat [00:38:50]: Yeah.Swyx [00:38:51]: This is like probably a not a real point, but, I was trying to think about like what other industries. I was trying to think about like, we cannot be first to like these kinds of problems.Akshat [00:38:59]: Yeah.Swyx [00:39:00]: And what other industries have had this? And I was like, airlines with fuel and like they have to hedge their fuel and like, I think for a long time Southwest because they made like a hero fuel bet, they like were like super low cost becauseAkshat [00:39:12]: OhSwyx [00:39:12]: Compared to everyone else.Akshat [00:39:14]: Yeah. I hadn't thought about that.Vibhu [00:39:16]: We're at a fun time too?Akshat [00:39:18]: Yeah. It's. A lot of the compute business in general, for us is also about being very good about capacity management. That is how you have great unit, economics. but also over time it's how you can unlock more value for customers. Like, one of the things we're building now is like a way for customers to get, If they don't care about latency, like get much cheaper pricing and they'll get results back in like next 24 hours or something, like a batch tier essentially.Batch Tiers and Latency-Insensitive WorkloadsSwyx [00:39:47]: Yeah.Akshat [00:39:47]: And those are levers we have because we control the whole stack and scheduling and whatnot to give people a sufficientSwyx [00:39:53]: Yeah. I feel like they're not as popular. Like those, like the Frontier Labs have all those APIs. They're not as popular as they should be.Akshat [00:40:00]: The demand that we see for something like that is not for LLMs. although sometimes people wanna run evals andSwyx [00:40:08]: OkayAkshat [00:40:08]: Synthetic data prep and there it makes sense.Swyx [00:40:10]: Okay.Akshat [00:40:11]: But it's from a lot of LLM companies, like people who are doing computational bio, like they have to run really big batch jobs and they don't care about when they get it back.Swyx [00:40:22]: Yeah. And like they have a reasonable. It's it's also like a cousin to the stopping problem of like, will this finish in time?Akshat [00:40:30]: Yeah. You can bound it.Swyx [00:40:33]: Yeah.Akshat [00:40:33]: Like you can give peopleSwyx [00:40:34]: YeahAkshat [00:40:34]: SLAs on it.Swyx [00:40:35]: Yeah. I think what's, what's interesting is like the next phase of model.Swyx [00:40:38]: Like what, do people expect from you, now that you're established and you're like well-known compute player among all these leading companies. You had an inference launch week, and we talked a little bit about the launches. like what else? Like what else should people know?What Modal Builds NextAkshat [00:40:55]: We are building primitives that make our users' lives much easier. So, I think for example, with LLM inference, thousands more companies are gonna post-train their own models and, deploy open source models for inference. so we're thinking a lot about what is the best product shape for that. And, that involves everything from our training gym to, then, endpoints that get frontier-level performance. again, but I haven't talked to anyone. It looks somewhat different on other verticals. Like, we're also seeing a lot of real-time, audio-video stuff in there, which is why like, we're working on things like regional routing, with fallbacks. So you can get GPUs that are as close to users as possible. so you get like low latency for video streaming and whatnot. And then on the agent side, it's,Akshat [00:41:52]: We're still working very closely with our customers because stuff is changing so fast in terms of what they need. And, I think beyond sandboxes and persistent file systems, there's a lot of other things people will need from this agent stack as they build production agents. So yeah, we're thinking about those other things that fit in there.Swyx [00:42:13]: I want to ask what the other things are.Akshat [00:42:15]: Yeah. I probably should share right now.Swyx [00:42:17]: I think-- I think, okay, so, I do think a lot about the principal components of cloud, and you do talk about compute storage networking.Akshat [00:42:25]: Yeah.Swyx [00:42:25]: Because so far for me, it's fine. so far for the. the first couple generations of cloud, it's fine. What's different, qualitatively different about agents that you need some new permission level? Like a lot of people, okay, and I'll just kinda spew tokens at you until it like hopefully sparks something.Akshat [00:42:43]: Yeah.Swyx [00:42:44]: Like the new level now is whatever Claude Code does, which is dangerously scope permissions or like allow list by command or like whatever, right? And sometimes they're like, “Well, okay, we have like this adaptive thinking mode where like, just trust me, bro. I will make the calls for you.” Is that it? like mediated permissions.Hard Guardrails vs. LLM-Mediated PermissionsVibhu [00:43:03]: Now you're looping it with a goal and letting it roll.Akshat [00:43:06]: Yeah, I'm, I'm skeptical of LLM media permission for stuff that is at the sandbox level because you do want hard boundaries.Swyx [00:43:16]: Yeah.Akshat [00:43:16]: Otherwise, someone can exfiltrate stuff.Swyx [00:43:20]: But likeAkshat [00:43:20]: YeahSwyx [00:43:20]: Maybe that's old school thinking. Maybe we're the dinosaurs.Swyx [00:43:23]: Maybe the AI OS or the LLM OS is really the kernel is a goddamn LLM.Swyx [00:43:30]: Like it makes you feel uncomfortable.Akshat [00:43:31]: Yeah, I'm, I'm toldSwyx [00:43:32]: But that's what trusting the LLM is. Like imagine a spherical cow perfect LLM.Akshat [00:43:36]: Right.Swyx [00:43:37]: That it.Akshat [00:43:39]: Maybe.Swyx [00:43:41]: I wanna test the boundaries, right?Akshat [00:43:42]: Yeah.Swyx [00:43:42]: Like, and I don't believe that, but I wanna see where I'm wrong ‘cause that's, that's the consensus.Akshat [00:43:49]: Yeah. I think you always need hard guardrails when you want, And you can pair those with softer guardrails, right? And that's gonna be a lot of mediated.Managed Agents and Specialized SandboxesSwyx [00:44:00]: There. I'll also get you a end with a couple of your commentary on like the ecosystem outside of Modal. Manage agents. Everyone has one. Gemini, OpenAI, Claude, very useful for you, but also like it is their way of starting to edge into your space.Akshat [00:44:17]: Yeah.Swyx [00:44:17]: What's going on?Akshat [00:44:19]: Yeah, we're, very excited to partner with Anthropic and some of the other foundation labs, will not name who we're also working with. the way we see it is the manage agent thing is a great place to start if you're starting out building an agent and, But then when you get to, building something more production grade, like you're a company that's like Ramp that's building their own, Ramp also runs their accounting agent on us, so their external-facing agent. You need a lot more control over, your compute primitive on things like, what sort - how do you persist different files that the agent has access to, and how do you snapshot and restore? How do you control the networking? maybe you want GPUs. When you get to that point, you kinda want, a specialized sandbox provider, that gives you those things, and that's the role that we are trying to play.Swyx [00:45:15]: YeahAkshat [00:45:16]: We don't really have an opinion on the harness, whether it runs - it's a cloud-managed agent, and you hook it up to Model Sandbox, or you run the harness in Model Sandbox. We'll see where people converge with that.Swyx [00:45:26]: Yeah. Do you any opinions on like the meta harnesses, or just another layer on top of these things?Akshat [00:45:31]: You mean like the OpenPipeSwyx [00:45:33]: OpenPipe is one. I think Vercel had one, which I can't remember the name of right now. Fredshot had one. and then, to me, most recently was Data Databricks that had Omnigen. All these are meta harness. Like it's kinda pseudo agent cloud type things.Akshat [00:45:50]: I personally have not played around with them.Swyx [00:45:53]: Yeah.Akshat [00:45:53]: Build agents with them.Swyx [00:45:54]: Everything's bullish Modal, as long as it consumes more infra.Akshat [00:45:57]: That's why we're focusing on the infra layer. It's somewhere where our, relative competence is and, also it's a hard problem to solve.Swyx [00:46:06]: Yeah. I will say like just generally reflecting on that, I don't know if - if there's other topics on Modal, but like just generally reflecting as an infra person, not as intense as you, but in that field, this has like been the most exciting time in infra. Like it was boring for a while, and you couldn't really get people excited about data infrastructure. Like Eric would get on Data Console, everyone just watched the video and like say, “Look at how many sandboxes I can spin up,” and no one gave a crap.Why Infrastructure Became Exciting AgainAkshat [00:46:39]: Yeah.Swyx [00:46:40]: And like now everyone gives a crap.Akshat [00:46:42]: That's true. It is a very exciting time, and I think a lot of that's driven by just the amount of scale all of this stuff needs.Swyx [00:46:50]: I think the, like a lot of your initiatives or a lot of your like product directions make sense in retrospect, which is like the best kind, but I wouldn't necessarily have thought about it myself, which.Akshat [00:47:00]: We need the predictions.Swyx [00:47:02]: I think there's a lot that you just don't even see, right? Like you have the batch, you have the voice, you have the multimodal, but what else?Akshat [00:47:10]: What else is coming up for usSwyx [00:47:11]: Yeah. Where do you see things going?Akshat [00:47:13]: Yeah. I, in generalBiotech, Robotics, and Non-LLM AI WorkloadsAkshat [00:47:15]: It's it's clear that there's there's a huge shift happening. I think one thing that's not as obvious to people because LLM inference gets talked about so much and is also we work a lot of companies that are, doing things like drug discovery and computational bio, like the Chai Discoveries of the world. Big things are probably gonna happen there. we work a lot of robotics companies that are putting robots in like active deployments and getting good results out of them.Swyx [00:47:45]: Is there Air Gap Modal? Is there a version that is like prem air gapped whatever?Akshat [00:47:50]: No. We,Swyx [00:47:51]: You should cloud only.Akshat [00:47:51]: Yeah.Swyx [00:47:52]: Yeah. Okay. But yeah, so what you're saying is like because you're focused on primitives and they're good primitives, you find use cases in all these kinds of things.Akshat [00:48:01]: Yeah.Swyx [00:48:01]: Probably diversifies you a little bit away from LMS all the time.Akshat [00:48:05]: Yeah, absolutely. We're, we'- our goal isn't to only serve the LLM inference market.Swyx [00:48:10]: There are a lot just on the website, the audio,Akshat [00:48:12]: Yeah. We said both onSwyx [00:48:14]: Computational bio images. Yeah, there's a lot here. There's QTA TTS, customizing. Oh, Chatterbox. there was customizing Whisper.Akshat [00:48:24]: Okay. Yeah.Swyx [00:48:25]: This screen reminds me of a fallen competitor, which Replicate.Model APIs vs. Differentiated AI ProductsSwyx [00:48:31]: What's your postmortem on what happened?Akshat [00:48:34]: This is one thing we've stayed away from is providing an API for models because I think providing model APIs is some of it ends up serving like a really hobbyist market, which is much less sticky.Swyx [00:48:50]: Yeah.Akshat [00:48:50]: And we've always wanted to build for companies that are building products and need more flexibility that's not just an API.Swyx [00:48:57]: Which you can build an API for a model and this is clearly what it is. But you - but what you're saying, you can wrap it into a more fully functioning back end that you run.Akshat [00:49:06]: Yeah. So all of our examples, it's not that spin up this model, here's an API token, use it. They're all code.Swyx [00:49:13]: Okay.Akshat [00:49:13]: And so the point is that this is just an example.Swyx [00:49:16]: Starter code.Akshat [00:49:17]: Yeah. But you can tweak it however you want.Swyx [00:49:20]: Yeah.Akshat [00:49:21]: And if you're like a company building a product, like, computational bio whatnot, yeah.Swyx [00:49:26]: I guess I'm trying to tease out for listenersAkshat [00:49:28]: YeahSwyx [00:49:28]: When does it stop becoming, oh, you're just an API call and you're just a wrapper on API to becoming what you call a product, right?Swyx [00:49:36]: Like, what is that layer? Like what-- Like, more lines of code, but like beyond that, what is the substance that people add that qualifies it to be something more?Akshat [00:49:46]: I think there's a little bit of like a selection effect of like a lot of the companies who do wanna get deeper into that level are probably building something that's more differentiated. And, I think, an example is like - with LLM inference, originally we, worked with companies that were building their own post-training frameworks or they were, - Ramp early in the day was training their own tokenizer and like swapping out the tokenizer in Llama and whatnot. I'm not saying that's, that successful, in that case. But a better example is like, let's say Suno. because Suno, does not use Modal for training.Swyx [00:50:26]: Mikey on the pod. Yeah.Akshat [00:50:27]: But they use Modal for all their inference and that's because they have like a custom-- They have completely custom model architecture and that means that they have to be at the code level and tweak things that are not, just an API.Swyx [00:50:41]: It's interesting as well, like we had, Ethan, most recently on the xAI Groq team make a prediction that like the next tier in video gen is not a better video model, it's a better model or agent that orchestrates video models.Video Agents and Production WorkflowsAkshat [00:50:56]: Oh, interesting.Vibhu [00:50:56]: Language model backbone that can use toolsAkshat [00:50:58]: RightVibhu [00:50:59]: And write code.Akshat [00:51:00]: Like, yes, I can make my second video or my second video from Groq, but I want my minute video.Akshat [00:51:06]: And I'm not going there through normal video gen.Swyx [00:51:10]: Yeah, that's interesting. I - So we have GPU sandboxes and recently have seen a few companies doing agents that do video manipulation or,Akshat [00:51:22]: Yeah. Give it FFmpeg and just do it.Swyx [00:51:23]: Run FFmpeg. But likeAkshat [00:51:25]: That's not enough.Swyx [00:51:25]: Yeah.Akshat [00:51:26]: You need to give it Adobe.Swyx [00:51:27]: Yeah, I hadn't put it together with like it would be a video production thing. in my mind these things were going more towards editingAkshat [00:51:36]: Yeah.Vibhu [00:51:36]: Well, shout out Mantis.Akshat [00:51:37]: I think about this a lot.Swyx [00:51:38]: .Akshat [00:51:41]: Yeah. Sorry.Vibhu [00:51:41]: Luma. Luma Agent is a version of this for video production, but it's a off.Swyx [00:51:46]: I was gonna get your quick takes, on some other stuff that happensGitpod/Ona, CI, and Runtime SandboxesSwyx [00:51:50]: In recent news and just-just see if you have anything interesting. Gitpod, very li

On répète qu'il suffit d'acheter l'outil, de créer son produit et de vendre. Franck Berthelot a passé près de dix ans à réparer les projets qui ont cru ça. Fondateur de Darix, il ouvre le capot de Salesforce et du CPQ pour montrer ce qui se cache derrière une simple case à cocher : 40 objets qui s'ajoutent d'un coup, toute une chaîne métier à réaligner, et un projet vendu six semaines qui finit par en durer six mois. Un épisode pour comprendre pourquoi la vraie difficulté d'un projet Lead to cash n'est presque jamais technique, mais humaine.Franck évolue dans l'écosystème Salesforce depuis 2018. Passé par le e-commerce chez Yves Rocher, la finance de marché, puis l'AMOA et les cabinets d'intégration (4C, Levana), il a été référent Business System chez Aircall pendant quatre ans avant de fonder Darix, un cabinet de conseil Lead to cash / Quote to cash, avec son associé Yacine Zelmat, architecte technique.Ce que tu vas apprendre dans cet épisode :Ce qu'est vraiment le CPQ (Configure, Price, Quote), et pourquoi cette brique devient le cœur du réacteur dès qu'on veut automatiser sa facturationPourquoi le fameux « on a déjà tout spécifié, il n'y a plus qu'à implémenter » est le signal d'un projet qui va déraper, et quelles questions reposer avant de se lancerComment une refonte Quote to cash touche toute l'entreprise (sales, customer success, ADV, finance), et pas seulement les commerciauxJusqu'où va le no-code Salesforce, et à quel moment le code devient inévitable (calcul d'ARR multi-SKU, renewals à gros volume)Pourquoi le RevOps qui brille est un profil hybride, aussi à l'aise avec le métier qu'avec le système, et jamais un pur expert d'outilComment la dimension métier protège les Ops face à l'IA, quand une knowledge base suffit déjà à remplacer un prestataire purement techniqueNotes complètes, ressources et captures de l'épisode : [LIEN]Chapitres :[00:02:54] Le parcours de Franck, d'Yves Rocher à l'écosystème Salesforce[00:10:09] Le CPQ expliqué : catalogue, pricing, devis[00:16:06] Le piège du « j'ai juste à implémenter »[00:21:08] La boîte de Pandore : pourquoi six semaines deviennent six mois[00:25:12] Le RevOps hybride et l'équipe système[00:41:03] Les déclencheurs d'un projet Quote to cash[00:57:04] Facturer des jours ou de la valeur : les Ops face à l'IA[01:28:22] Le facteur humain : reconvertir l'ADV, aligner avant de systémiserPour aller plus loin :L'épisode d'Engrenages avec Raphaël Brousse, qui définit le Quote to cash comme une porte d'entrée sur la compréhension businessDarix, le cabinet de Franck et YacineLes concepts cités : CPQ, ETL, multi-SKU, Data CloudLes outils évoqués : Salesforce, NetSuite, HubSpot, AttioHébergé par Ausha. Visitez ausha.co/politique-de-confidentialite pour plus d'informations.

In this episode of the Global Trade Gal Deep Dive Podcast, we explore why UL and ETL certifications are crucial for businesses aiming to sell electrical products in the United States. These certifications are more than just legal requirements—they're essential for ensuring product safety, meeting compliance standards, and building a trusted brand.We break down the key difference between having certified components and obtaining a finished-product safety mark, which evaluates the entire product's ability to handle heat, mechanical stress, and more. Without these certifications, companies risk massive liability lawsuits and may face barriers to working with major retailers and wholesalers in the U.S. market.Additionally, we highlight the importance of recurring factory audits that come with these certifications, which help ensure long-term manufacturing quality and consistency. By investing in UL or ETL certifications, companies not only meet compliance mandates but also lay the groundwork for accessing premium global distribution channels and establishing a strong reputation in the marketplace.Tune in to learn why these safety marks are non-negotiable for success in the competitive U.S. electrical product market!Read our blog https://mondoro.com/why-ul-or-etl-certifications-matters-for-selling-in-the-us/

In this talk, Ivan, Senior Engineering Manager at Personio, shares his deep expertise in the data and software space from his early days building traditional NLP systems and massive ETL pipelines to his current leadership role in Identity and Access Management (IAM). We explore the rapid evolution of Generative AI, the reality of managing AI agents in production, and the emerging field of context engineering to optimize developer workflows.You'll learn about:- The buy vs. build dilemma for AI infrastructure and local LLMs.- How AI agents are shifting workloads and evolving code reviews.- Why AI is currently better at fixing tech debt than building from scratch.- Measuring the ROI of AI integration using DORA metrics and cycle times.- Strategies to manage vendor lock-in and minimize AI provider dependency.- Using "context engineering" and specification-driven development to maximize LLM quality.- Why hiring junior engineers is still essential and how AI accelerates their onboarding.TIMECODES:00:00 Career Journey in Data Science and NLP07:37 Industry Adoption of Generative AI and Agents11:45 Buy vs Build Dilemma for AI Infrastructure15:46 AI Capability Limits in Fixing Tech Debt19:32 Developer Workloads and AI Code Contributions24:49 Experimentation with Open Source AI Agent Architectures30:06 Measuring ROI and Business Value of AI Integration35:10 Tracking AI Impact Using DORA Metrics39:51 Impact of AI Code Generation on CI/CD System Reliability43:00 Best Practices for Team AI Tool Adoption48:20 Managing Vendor Lock-In Risks with AI Providers51:27 Importance of Hiring Junior Software Engineers56:28 Accelerated Junior Developer Onboarding with AI Assistants01:00:12 Specification-Driven Development and Context EngineeringThis talk is perfect for software engineers, engineering managers, and technical leaders looking to practically integrate AI tools into their teams without sacrificing code quality or system reliability. It is especially valuable for tech professionals navigating the complexities of AI adoption, CI/CD pipeline management, and organizational scaling in the GenAI era.Connect with DataTalks.Club:- Join the community - https://datatalks.club/slack.html- Subscribe to our Google calendar to have all our events in your calendar - https://calendar.google.com/calendar/r?cid=ZjhxaWRqbnEwamhzY3A4ODA5azFlZ2hzNjBAZ3JvdXAuY2FsZW5kYXIuZ29vZ2xlLmNvbQ- Check other upcoming events - https://lu.ma/dtc-events- GitHub: https://github.com/DataTalksClub- LinkedIn - https://www.linkedin.com/company/datatalks-club/ - Twitter - https://twitter.com/DataTalksClub - Website - https://datatalks.club/ Connect with Ivan:- LinkedIn - https://www.linkedin.com/in/ivan-bilan/ - Twitter - https://x.com/demiourgosua - Github - https://github.com/ivan-bilan - Website - https://github.com/ivan-bilan

What if the data engineering skills you have today become obsolete in five years? In this episode, host Benjamin Wagner sits down with Pranav Motarwar, a data engineer who's witnessed the industry's transformation from traditional ETL to AI-powered pipelines, to explore how AI is fundamentally reshaping data engineering roles, why you need to master both "AI for data" and "data for AI" to stay relevant, and the emerging infrastructure required to handle multimodal data at scale. Whether you're a data engineer wondering about your career longevity or a builder curious about next-gen data stacks, this conversation unpacks the skills you'll need, the tools defining 2026, and why data engineers aren't disappearing - they're just evolving faster than ever.

Réécoutez FG Chic invite One Benodet avec Belle Boutique du lundi 15 juin 2026L'équipe passionnée de CGH events est heureuse de vous proposer à l'occasion de l'arrivée de l'été et de la fête de la Musique, une soirée exceptionnelle dans les jardins de la résidence LeOne, située à l'entrée de Bénodet. Le dimanche 21 juin 2026, l'équipe et son Dj résident Nicolas Nucci, recevrons Arnaud Godefroy programmateur et DJ de la radio Fg Chic et co fondateur du label “Belle Boutique” pour une fête en open air sur la grande terrasse et sous la tente aux inspirations berbères, dans le jardin. Au programme, deux Dj sets exclusifs au son des reprises et remixes branchés ! Nous proposerons notre carte habituelle Bar, Champagne cocktails et snack, planches, tapas.L'entrée sera libre mais surveillée.Rendez vous donc le 21 juin dès 16h pour une fête de la Musique exceptionnelle à la Résidence le One, 1 route de Bénodet, 29950 Clohars-Fouesnant.www.cgh-events.frFb & Ig: le_one_benodetTracklist : 1 - Marine Neuilly Confetti (Kraak & Smaak Remix)2 - Avener (The) - Fade Out Lines (Belle Boutique Re-Edit)3 - Patti Smith Group - Frederick (Belle Boutique Re-Edit)4 - Belle Boutique - Fly - Extended Version5 - Caroline Legrand - J'aurais Voulu Te Dire (Belle Boutique Re-Edit)6 - Flirts (The) - Passion (Belle Boutique Re-Edit)7 - HUGEL & SOLTO (FR) Jamaican (Bam Bam) [Extended Mix]8 - Michael Zager Band - Let's All Chant (Belle Boutique Re-Edit)9 - Abba - Money, Money, Money (Belle Boutique Re-Edit)10 - Anoraak & Julia Jean-Baptiste - Magnifique (Italo Edit)11 - Lifelike - The Chase12 - Disciples - They Don't Know (Original Mix)13 - Marie Myriam - L'oiseau Et L'enfant (Belle Boutique Re-Edit)14 - Zz Top - Gimme All Your Lovin' (Belle Boutique Re-Edit)15 - My Mine - Hypnotic Tango (Alt1) (Belle Boutique Re-Edit)Sélection et mix par Arnaud Godefroy pour Belle Boutique ParisInstagram : www.instagram.com/belleboutiquemusicSoundcloud : @belleboutiqueFacebook : www.facebook.com/BelleBoutique.frYoutube : www.youtube.com/@BELLEBOUTIQUEPARISLinktree : https://linktr.ee/belleboutiqueparisManagement : Belleboutiquemusic@gmail.com

In this episode of Alexa's Input (AI), I sit down with David Aronchick, co-founder and CEO of Expanso and former product lead for Kubernetes at Google.Data is growing everywhere outside your data center. Solar panels in remote across a country. Security cameras at retail stores. IoT sensors across factory floors. And moving that data to the cloud for processing? It's expensive, slow, and often restricted by compliance.David is an expert when it comes to solving distribution problems. He led Kubernetes product at Google, co-founded Kubeflow to bring ML to production, and now he's building Expanso to tackle a difficult constraint: when your data can't move, how do you process it where it lives?We discuss:- The need for distributed data orchestration-Upstream data control: filtering and transforming at the source- Three forces making edge computing inevitable (physics, regulations, economics)- How to build successful open source infrastructure projects- Customer discovery and finding real pain points- His transition from Protocol Labs to founding Expanso- ETL pipelines: moving the first four steps closer to the data- Context loss and lineage in distributed systems- Processing 400,000 signals per second with 150MB agents- AI observability: attaching source metadata to training data- Running ML pipelines at the edge- Real-world deployment challenges (bandwidth, regulations, cost)Expanso is rethinking how we process data in an AI-native world—moving compute to data instead of data to compute. If you want to understand where distributed systems and edge computing are heading, this is a deep dive into the infrastructure layer beneath modern AI applications.General Podcast LinksWatch: https://www.youtube.com/@alexa_griffith Read: https://alexasinput.substack.com/ Listen: https://creators.spotify.com/pod/profile/alexagriffith/ More: https://linktr.ee/alexagriffithLearn more about the host atWebsite: https://alexagriffith.com/ LinkedIn: https://www.linkedin.com/in/alexa-griffith/Find out more about the guest atLinkedIn: https://www.linkedin.com/in/aronchick/ Twitter/X: https://x.com/aronchick GitHub: https://github.com/aronchick Expanso Website: https://expanso.io/ResourcesExpanso Website: https://expanso.io/ Kubernetes: https://kubernetes.io/ Kubeflow: https://www.kubeflow.org/ CNCF (Cloud Native Computing Foundation): https://www.cncf.io/ Protocol Labs: https://protocol.ai/KeywordsDavid Aronchick, Expanso, Kubernetes, Kubeflow, distributed systems, edge computing, data pipelines, ETL, upstream data control, Google Kubernetes Engine, open source, CNCF, observability, log processing, data lineage, provenance, schema enforcement, IoT, edge AI, distributed data, machine learning infrastructure, Protocol Labs, IPFS, Filecoin, data governance, compliance, GDPR, bandwidth optimization, data aggregation, AI infrastructure, multi-cloud, hybrid cloud, real-time processing

Send us Fan MailModernizing SAP environments requires far more than executing a software upgrade or signing a new licensing agreement. Organizations migrating to SAP S/4HANA or consolidating multiple regional ERP systems into SAP often face significant risks tied to fragmented legacy data models, inconsistent master data, and undocumented transformation logic that can undermine production cutover readiness. In this session, SNP CTO Steele Arbeeny explains how Kyano Crossway supports legacy-to-SAP conversion programs beyond traditional ETL approaches by governing the entire data conversion lifecycle. Rather than treating migration as isolated data loads, Crossway structures and manages mapping logic, transformation rules, validation workflows, and traceability controls to ensure transparency and audit confidence throughout the process. As a result, organizations gain visibility into what changed, why it changed, and whether the transformed data is fully prepared for production deployment within a governed SAP-ready architecture.Video: https://www.elevatiq.com/events-and-webinars/legacy-sap-workloads-readiness-turning-legacy-data-into-sap-ready-data/Questions for Panelists?

This week Noah and Steve talk about why they'll be at Southeast Linuxfest. Noah introduces everyone to the Ai embedded in his rental car, and the boys finally find a Z-Wave thermostat! -- During The Show -- 00:50 Intro When AI can be good Conversing with AI Varying AI model quality Creativity and AI Prompt engineering Write into the show about AI Accelerator for people already have a base skill 10:00 SELF and Picking Conferences Self is this Weekend Please come to SELF Really use help Wednesday New Venue Why go to SELF Conference size Watch for the ANS meetup 15:55 Door bells - Ziggy Steve would run Ethernet Why WiFi cameras are bad ReoLink Support 22:15 Thermostats - Tim RadioThermostat CT50 Honeywell T6 Pro Z-Wave Programmable Thermostat (Model TH6320ZW2007) 25:30 UL Listed UL, ETL, CE certification European way of doing things UL508A Gosund Wo1 Tuya smart outlet 30:00 AI in Cars AI built into rental Noah's AI experience AI Audio clips AI engages with you Safety concerns Possibility of latent functionality 40:20 Age Verification California's Digital Age Assurance Act Open Source getting a pass in CA and CO Privacy trade offs Pattern of freedom and contraction ItsFoss 48:00 OpenAI IPO Large evaluations Fundraising with out fundraising May be a cash out move Going public "opens the books" CNBC -- The Extra Credit Section -- For links to the articles and material referenced in this week's episode check out this week's page from our podcast dashboard! This Episode's Podcast Dashboard Phone Systems for Ask Noah provided by Voxtelesys Join us in our dedicated chatroom #GeekLab:linuxdelta.com on Matrix -- Stay In Touch -- Find all the resources for this show on the Ask Noah Dashboard Ask Noah Dashboard Need more help than a radio show can offer? Altispeed provides commercial IT services and they're excited to offer you a great deal for listening to the Ask Noah Show. Call today and ask about the discount for listeners of the Ask Noah Show! Altispeed Technologies Contact Noah live [at] asknoahshow.com -- Twitter -- Noah - Kernellinux Ask Noah Show Altispeed Technologies

Podcast: Tech TransformedGuest: Mihir Nanavati, GM and Product Executive in MarTech and AdTechHost: Doug Laney, Research & Advisory Fellow at BARC and Author of Infonomics & Data JuiceAI might have overtaken the industry with processing data, automating workflows, and creating content. The next big thing could be a major one, says Mihir Nanavati, GM and Product Executive in MarTech and AdTech, “AI is moving from managing data to making decisions with it.”In the recent episode of the Tech Transformed podcast, host Doug Laney, Research & Advisory Fellow at BARC and Author of Infonomics & Data Juice, sat down with Nanavati to talk about a larger transformation in data and decision-making systems driven by AI.They particularly focus on the integration of agentic AI in marketing and customer data platforms. They explore the challenges of fragmentation in ad tech, the importance of connecting customer data to revenue outcomes, and the transformative role of AI in decision-making processes. Mihir shares insights on how companies can leverage AI to enhance their marketing strategies and the future of first-party data."This is not a cost exercise, it's about how much more you can get done and how many more ideas you can execute," said Nanavati.For years, enterprises went through waves of technological change, including cloud infrastructure, mobile platforms, and customer data platforms (CDPs). Each development helped enterprises collect, store, and manage larger amounts of data. However, Nanavati asserts that humans making most decisions will never change. Now, AI agents are introducing a new model.How AI has Moved from Data Navigation to Making DecisionsIn the past, customer data initiatives aimed to create a unified view of customers. Enterprises built warehouses, ETL pipelines, and data platforms that were designed to be reliable. However, Nanavati suggests that AI agents are changing these expectations. "Machines can reason, and that is fundamentally different."Rather than simply serving as another analytical feature in existing systems, AI agents are increasingly acting as decision-makers. They weigh trade-offs, learn from results, and execute plans based on specific goals.This change has significant implications for customer data platforms. CDPs are not just repositories for customer information now. Instead, they are becoming layers that enable intelligent actions."The role of customer data platforms is evolving into ‘how do you make meaning of this?'" While, decisions about which customer segment to target, which message to send, or which offer to present may increasingly be guided by AI-driven systems.What's the Fragmentation Problem in Modern AdTechWhile AI agents create new opportunities, Nanavati pointed out a persistent issue in the AdTech and MarTech ecosystem – fragmentation. Brands today tend to lean towards deploying multiple advertising and customer engagement platforms. These include social platforms, retail media networks, email tools, and specialised ad technologies. Each system may optimise effectively within its own space, but often fails to connect at the customer level.Nanavati calls it a "paradox of choice." "Each system is optimising locally for its own clicks and conversions, but none of that is coordinated at the consumer level."The result is a customer experience that many consumers notice, alluding to repeated retargeting for products they have already bought, irrelevant recommendations, or disconnected interactions across channels.As enterprises adopt AI agents, fragmented data environments may become an even bigger problem. AI systems can process information quickly, but they still rely heavily on context. "AI doesn't need perfect data in many cases, but it needs context."What's Next for Enterprise Tech?As AI adoption continues, Nanavati believes that successful enterprises will be recognised not by how many experiments they run, but by how fast they learn and use the results."Learn very rapidly. Then scale what you've learned." For leaders, this may require a stronger commitment than just isolated pilot programs or limited rollouts. It may also need organisational changes that place AI decision-making and customer context at the centre of growth strategies.For companies navigating the intersection of AI agents, CDPs, and customer data, the question may no longer be whether AI can automate processes. The ultimate question is about who is calling the shots.Key TakeawaysAI is fundamentally changing how decisions are made in marketing.The shift from third-party to first-party data is crucial for businesses.Fragmentation in ad tech leads to a paradox of choice for brands.Connecting customer data to revenue outcomes is essential for success.AI can help marketers make better decisions without needing perfect data.Customer data platforms are evolving to support real-time decision-making.Companies can run significantly more marketing experiments with AI.Leaders must personally drive change in their Enterprises.Successful AI implementation requires a focus on revenue outcomes.First-party data collection is becoming more sophisticated and essential.Chapters00:00 Navigating the Shift in Data and AI03:03 The Evolution of Decision-Making in Marketing05:55 Challenges of Fragmentation in Ad Tech09:00 Connecting Customer Data to Revenue Outcomes11:56 The Role of AI in Customer Data Platforms14:55 Real-World Applications of Agentic AI18:05 Blueconic's Approach to Customer Growth21:14 The Future of First-Party Data24:02 Building Habits for Successful AI ImplementationListen to the full episode of Tech Transformed for a deeper discussion on AI agents, customer data platforms (CDPs), first-party data strategies and the future of AdTech. Subscribe for upcoming episodes and join the conversation across our social channels.BlueConic LinkedIn: @BlueConicEM360Tech YouTube: @enterprisemanagement360EM360Tech LinkedIn: @EM360TechEM360Tech X: @EM360TechFor more information, please visit em360tech.com and blueconic.com.

Mike & Tommy dive into Building Agentic Tools for Microsoft Fabric With Alex Powers, exploring Microsoft's Alex Powers' new tool 'Taskflow Assistant' for Microsoft Fabric and how AI is transforming the way professionals work with Fabric workflows, discussing whether we're helping users master the platform or avoid learning it altogether.https://github.com/microsoft/fabric-task-flowshttps://community.fabric.microsoft.com/t5/Fabric-Updates-Blogs/Discover-items-across-workspaces-with-the-OneLake-Catalog-Search/ba-p/5176768https://community.fabric.microsoft.com/t5/Fabric-Updates-Blogs/Pipelines-are-evolving-beyond-ETL/ba-p/5177527Get in touch:Send in your questions or topics you want us to discuss by tweeting to @PowerBITips with the hashtag #empMailbag or submit on the PowerBI.tips Podcast Page.Visit PowerBI.tips: https://powerbi.tips/Watch the episodes live every Tuesday and Thursday morning at 730am CST on YouTube: https://www.youtube.com/powerbitipsSubscribe on Spotify: https://open.spotify.com/show/230fp78XmHHRXTiYICRLVvSubscribe on Apple: https://podcasts.apple.com/us/podcast/explicit-measures-podcast/id1568944083‎Check Out Community Jam: https://jam.powerbi.tipsFollow Mike: https://www.linkedin.com/in/michaelcarlo/Follow Tommy: https://www.linkedin.com/in/tommypuglia/

Shinji Kim, founder of SelectStar (acquired by Snowflake in December), joins the show to discuss the deal, the integration into Snowflake's Horizon catalog, and where data cataloging is actually headed.We get into the weeds on a claim Shinji makes early: in a few years, we may stop calling these things "data catalogs" at all. The category is evolving into an AI context layer, a living surface that combines metadata, semantic models, business glossaries, and ontologies, continuously updated by both humans and agents. Shinji walks through how SelectStar built toward this with semantic model management, MCP server support, and an AI agent that started serving data analysts and eventually answered business users' questions directly.We also dig into where data catalog implementations go wrong (spoiler: it's almost always adoption, not tooling), why marketing teams are an underrated ETL persona, and what it actually took to get acquired by Snowflake after three years as a premier partner.Plus: if Shinji were starting SelectStar today, what would she do differently? We talk about distribution in the AI era and how the startup playbook is mutating.Connect with Shinji:LinkedIn — https://www.linkedin.com/in/shinjikim/

Ryan MacDonald, country leader for SAS Canada Recorded on site at SAS Innovate 2026 in Grapevine, Texas, today’s In The Channel features Ryan MacDonald, country leader at SAS Canada, in a wide-ranging conversation about what the week’s major announcements mean for Canadian organizations – and where SAS sees its channel and partner opportunity growing. The conversation opens on the energy at SAS Innovate, which marks the company’s fiftieth anniversary, and what the announcement lineup – including the new SAS AI Navigator for AI governance and the expansion of agentic AI capabilities across the Viya platform – means for the Canadian market specifically. MacDonald describes Canadian enterprise AI maturity as strong in intellectual capital but still building toward consistent economic output, with the governance and trust framework a necessary foundation before organizations can scale. He draws a direct line between Canada’s regulatory environment – OSFI E-21 in particular – and the practical operational pressure organizations are feeling as model validation volumes have grown from two a week to multiple per day. On the competitive landscape, MacDonald addresses the challenge from Microsoft Fabric and Databricks with an argument about SAS’s existing footprint in business-critical decisioning layers – often invisible infrastructure organizations don’t always realize they’re sitting on, and an upgrade path through Viya designed to deliver incremental value rather than a rip-and-replace. The conversation also covers the evolution of SAS’s channel strategy, the managed services opportunity in a data sovereignty environment, and the MCP-based openness that is letting external AI agents call SAS analytics directly. Read Full Transcript Robert Dutt: Hello, and welcome to In The Channel from ChannelBuzz.ca, bringing news and information to the Canadian IT channel for the last 16 years. I’m Robert Dutt, editor of ChannelBuzz.ca, and your host for the show. This week, I’m coming to you from Grapevine, Texas, where I’ve been on the ground at SAS Innovate 2026. It’s a significant week for SAS Institute on a couple of fronts. The company is marking its 50th anniversary this year, and the announcement lineup has been one of the more substantive in recent memory, with major moves in AI governance, agentic AI across the Viya platform, and a meaningful shift in how the platform opens up to external AI agents and frameworks. My guest today is Ryan Macdonald, country manager [CHECK: title recorded as “country manager” – should be “managing director” if you want to punch in] for SAS Canada. Ryan’s been with SAS Canada for about a decade, and has just stepped into a role leading the country this year. He has a front row seat to some significant strategic changes – the move to Viya, the expansion of the partner and channel program, and now what I think is a genuinely important moment as AI governance moves from theoretical concern to practical operational requirement, particularly in Canada’s regulated industries. We cover a lot of ground – what this week’s announcements mean for Canadian organizations, where Canadian enterprise stands on AI maturity right now, the OSFI E-21 story, how SAS is thinking about its channel ecosystem and the mid-market opportunity, and a candid conversation about managed services and data sovereignty. Let’s get right into it. My chat with Ryan Macdonald. [MUSIC] Robert Dutt: Ryan, thanks for taking the time, and what I’m sure is a busy week for you. Ryan MacDonald: Yes, of course. Thanks for having me, Robert. Robert Dutt: You guys turned 50 this year, and it feels like one of the bigger product lineup announcements at Innovate in a while. Curious what you felt from the room. What’s the energy, what’s the vibe that you’re getting from this year at Innovate, especially given that 50 years of SAS framing? Ryan MacDonald: I agree with the energy you’re feeling. Certainly a ton of energy around our 50th and just what we’re seeing in terms of AI tooling and where we fit into that ecosystem. So lots of conversations about the data estate, how that’s evolving, and then just really looking for the reality check on where practical value lives in the new AI ecosystem that’s being framed around, especially for enterprise technology stacks. Robert Dutt: Look at the announcement stack this week. You’ve got Navigator for AI governance. You’ve got the agentic AI expansion in Viya, the various industry solutions. Curious – and I’m sure you’ve seen some of these before they were announced to the public and been following their development – what is kind of activating your Spidey senses in terms of, “ooh, that’s going to play well at home right now.” What are we seeing as sort of the big early day opportunities out of those innovations? Ryan MacDonald: Certainly in Canada, the regulatory domain around model risk management and model management and lineage and explainability is front of mind for everybody. I think that’s the major limiting factor in terms of proliferating cost of AI, in terms of actually calculating a per unit cost of running a model or introducing intelligence to something that was maybe traditionally rules-based. And so I think not only is there a regulatory driver, but people are seeing that as a practical constraint. So a lot in the governance and trust domain is certainly a hot topic. Robert Dutt: And that kind of speaks to where I wanted to go next, actually, which is you guys have been in Canada across verticals for a long time, obviously. Curious how you would describe the overall kind of AI maturity of the Canadian market right now. Are we kind of leading, lagging? Or is there something distinctly Canadian to it? Ryan MacDonald: Yeah, great question. This is close to home. We have the benefit of working with thought leaders in AI, folks like Ajay Agrawal. And just knowing the pedigree of intellectual property around this conversation in Canada, we have so much there. Of course, Geoffrey Hinton and Ilya Sutskever and the folks at U of T have just delivered so much to this community. I think that said, enterprise adoption and converting this into economic output is still something that we’re figuring out. So I think our investments generally, relative to peer groups around the world, we’re still a little behind. I think we’re doing some advanced things. There are some exceptions to this, where use cases are at the forefront of what’s being delivered globally. But generally, I think the data estate and this trust dynamic and the need for establishing a scalable framework for trust and governance – it’s a responsible thing to do. But relative to other geographies, it’s setting a foundation before we really run away with some use cases and deliver. Robert Dutt: One thing we’re tracking – I’m sure a lot of people are – is the idea of AI initiatives that get a start and a lot of fanfare and then fizzle out before hitting production or certainly proving their worth. I’ve heard a lot of the framing of the idea of trust and governance as kind of the growth driver, rather than the compliance tax. How is that hitting in Canada? And is that any different than what you’ve seen in terms of reactions and feeling and overall motion in the states or elsewhere? Ryan MacDonald: I think there are certainly differences in the tone of this conversation. For me, the purview is mostly north and south of the border – the US and Canada. But I think in Canada, we have a regulatory domain that is really prioritizing these things. So it’s not optional for a lot of – especially in a regulated market, this isn’t really a luxury you’d have to say, do I comply with this or not? But I think it’s also putting a per unit cost parameter on this for folks that is important. We’re seeing a huge proliferation of AI. Everything – your microwave, your lawnmower, everything has some sort of AI enablement component to it. Is it necessary? Are you getting the appropriate uplift? And these teams that are validating and pushing these models through the organization – what we’re hearing from them – this went from two a week, to a month, to two a day, five a day, ten a day. And so the systems – it’s not just a luxury or a question really of the ethics. Are we doing the right thing? Is this responsible? It’s a framework that’s required for the validation process, even just table stakes, to really scale through the organization. Robert Dutt: To that point, in Canada we’ve got financial services, and particularly we’ve got OSFI E-21 coming up. That’s pretty scary – things attached to it if you’re not hitting the bar. Are you seeing that create urgency? Or are customers still in a wait and see kind of space around that? Ryan MacDonald: I think the regulatory conversations there are interesting. There’s a lot of assessment of what peers are doing. And I think OSFI, to their credit, really listens to the community. Rather than setting a standard blind lead, just based on their intellectual property and what they see as being a requirement, they really listen to the community and measure from where everybody is, taking stock of that. So I don’t believe there’s a lot of fear and panic. I think organizations – as we did a lot of work around E-21 [CHECK: transcript rendered as “E23” – confirm on playback] specifically in this space – they were really well prepared. They had some ideas on how to make this more efficient, really focus on the materiality of where the risk lives and develop a framework that’s consistent with the risk posture in other domains. And I think that’s really – nobody was suggesting, “hey, this isn’t a good idea. This is too much pressure. This is putting a cost burden on us.” That wasn’t really the dialogue. Robert Dutt: Beyond financial services and other regulated industries especially, what are you seeing in terms of how customers are wrestling with AI governance right now? Ryan MacDonald: I think the scale of maturity across industries just varies so greatly. You have some organizations that are really just getting started, and they’re acknowledging that. In some of the roundtables we’ve had the benefit of participating in, some folks are trying to find their first step in AI. What does this even mean? They’re trying to find the right resources that can guide them. They’re still building their technology estate. And then, conversely, you have folks that are, as we spoke about earlier, leading the world – the global community – in terms of things like automated decisioning frameworks and integrating what were previously siloed processes. We see this in risk and fraud domains merging together. So I think we’re seeing both ends of that spectrum in Canada, certainly. Robert Dutt: Analytics has become a crowded space lately – with Databricks, with Snowflake, with Microsoft Fabric getting in there, all in territory that you guys have been in for a long time. How do you make the case to Canadian organizations that have been told, especially by Microsoft, “hey, you can just have analytics as part of what you already have?” What’s the competitive message there? Ryan MacDonald: Yeah, that’s a regular conversation for us, of course. I think what we really offer institutions, especially given the scale of the organizations we support – and we work in almost every major industry, every major enterprise in Canada – we offer a very different risk posture in moving through this process. So they may have what were traditional analytics with SAS. Maybe we had dabbled in what was previously BI, something like that. But for a lot of institutions, we support business-critical payload. There is a core application to their business that’s being delivered with a component of SAS. And oftentimes, as our relationships diversify across the organization, maybe we have a specific technology sponsor that helped build this alongside their business counterpart. Maybe they’ve moved on. And that decisioning layer is sort of obfuscated. So we spend a lot of time identifying – hey, is this what looks like ETL work potentially, in a report or an assessment that’s performed? Is this really a decisioning layer in your organization? And that’s what we’re really finding is there. And what folks are really interested in is taking that framework – what was previously identified as legacy SAS – and seeing what we offer in terms of Viya. It’s scaling far beyond what the competition can offer in terms of decisioning frameworks and automating process and delivering core value. A lot of the AI discussion is focused now on where are you seeing ROI? How long do we have to wait? What is the roadmap to finally get something out of this? And I think that’s really the core difference. Yes, there’s a lot of tools. It’s a crowded space. The competition is fierce and they can do some very exciting things. I think what we offer organizations is really the opportunity to do those same things and more, and to take your current investments, your current intellectual property, through that framework – which delivers value incrementally rather than a build within a complete new paradigm. Robert Dutt: One of the announcements that really caught my eye this week was the addition of the MCP – in that essentially you guys are opening up the analytics engine to external AI agents like Claude to call it directly. It seems like a pretty significant shift in terms of thinking about openness, thinking about consuming SAS from wherever folks want to consume it. What does that motion mean for the Canadian organization and for your Canadian customers? Ryan MacDonald: I think this is an extrapolation of what we spoke about earlier, in the sense of we are providing these deterministic decision frameworks to these organizations today. And so we talk about this almost in the sense of the Apple/Android paradigm. This was a previously closed ecosystem. The SAS code base was proprietary. The compute infrastructure was proprietary. And the open source motion was the first move here – running Python and R and other code frameworks natively within SAS is something that we’ve supported now for years within Viya. And it’s an extrapolation of this – meeting our customers where they are. SAS did not endeavor to compete directly with the frontier labs and build LLM models. But we certainly see the benefit – this is providing the market the productivity increase, the creativity of use cases, and what this adds to decisioning frameworks. I think the shortcoming is still the deterministic component, where something can be built in a hard and trusted capacity, presented to a regulator with the appropriate lineage. That’s really where we see these worlds coming together. So I don’t think it’s a great strategic decision if SAS were to impose, “we have one specific framework, one partner in this space.” We’re seeing, in addition to the frontier labs, a lot of custom work in this space as well – enterprises that are building more small language models around their data sets. So imposing this integration framework, I think, allows us to really meet customers where they are. Robert Dutt: A few years ago there was a flurry of things going on on the channel side for you guys. You brought on TD SYNNEX as a distributor. I believe it was a worldwide, not Canadian-specific figure that you were going for – 30% of contribution through partners. Where’s the channel scene at for you today? How would you characterize where you’re at against those goals and others? Ryan MacDonald: I think we’re still making progress in that domain. The channel business is still growing very aggressively. It’s a big shift to turn, frankly, in terms of getting the allotment of customers we had when we segmented what work was going to the channel, how that was going to be developed. And we compare ourselves to our peers in the industry – they’ve been at this for a lot longer. So just the maturity continues to develop. I think we’re seeing great progress, great feedback from customers in terms of the way that the channel is able to support them. And we see proliferation of niche players here that have come out of the woodwork that are very industry-specific. So I think that’s really the opportunity – where we had a general technology-based approach for certain industry segments, what we’re seeing is these channel partners can really tie together these business outcome-driven discussions in a way that was much more expensive and difficult for SAS to scale to. Robert Dutt: What does the community look like today in terms of scale, profile of partners, what they’re doing, and where do you see that evolving over the near future? Ryan MacDonald: I think we’re seeing this change very quickly with the advent of AI in terms of what use cases are being prioritized. I think in Canada, a lot of organizations have hit a wall in terms of understanding their data foundations – they’re not necessarily ready to scale them towards all the outcomes they’re seeking to deliver. And so channel partners are that domain. What are our peers doing? And this is GSIs and niche consulting firms and everybody in between. So we’re really seeing those conversations take shape of almost a reset of the roadmap, a reprioritization of how they’re building out their target state ecosystem. And that industry expertise is, I believe, the real differentiator. There’s a lot of competition. It’s a crowded space in that sense. So having an outcomes-focused point of view, whether that’s from SAS directly or a channel partner, is really important. Robert Dutt: Is the changing nature of what you guys are focused on in terms of AI governance and all those kinds of things that we’ve been talking about changing the definition of who you’re working with as a partner? Or is that something that’s likely to happen in the near future? Ryan MacDonald: I don’t think it’ll necessarily change. We might add some things to it, but they’re really part of the same conversation. I don’t think you can have a conversation about scaling AI without a discussion about the governance framework. And in a lot of cases, model inventory work, and just being the core platform of delivering models in this decisioning layer, is something that SAS had a lot of experience and an existing footprint within. So I think it’s really germane to the way we’ve been working with these customers today. Robert Dutt: How does the service mix – how they actually bring this all to market as partners – change as kind of what you’re going after changes? Ryan MacDonald: I think there’s a lot more consultative work right now around these outcome-focused and prioritization discussions. So I think it certainly is changing. And if you’re seeing this sort of increased competition in the technology domain and more commoditization of certain tool sets, it just puts more weight on – how do I really navigate? It crowds the pathway and creates more obstacles in terms of delivering outcomes. And so I think just refocusing on outcome-oriented discussion – and a lot of times these are deep partnerships between a niche consulting vendor, or somebody that now is a channel partner to SAS, and these firms in sectors across Canada. So it’s not necessarily changing the way we’re working with them. It’s changing the prioritization of the discussion, putting consulting maybe ahead of technology. Robert Dutt: Before we sat down to record, just as we were getting to know each other, you mentioned that part of your path through SAS Canada was you had managed services, at least for a while – and I believe that to be internally. How has that shaped, and how does this moment shape, how you think about working with partners who are in that managed services kind of motion? Ryan MacDonald: Yeah, that conversation is changing everywhere in the world. The political landscape, of course, is relevant here – in terms of we’re seeing some location dictate where customers are willing to send or host data. We’re seeing geo-repatriation in that sense. We’re seeing movement to the cloud change the dynamics of the cost model, what folks are seeing in terms of stable applications that don’t necessarily need the scalability or proximity to data. We’re seeing them pull some things back on premises and build clouds internally with OpenShift and other technologies. So I think it’s a cycle like most things in technology, where we’ve had the gold rush of moving everything to the cloud. And I think especially enterprise customers are now deciding not only how do they divide that workload amongst hyperscaler partners, but what is appropriate for internal clouds, which are now growing in popularity. And I think in Canada, we’re not seeing a huge disruption in this space, but we’re seeing a lot more of our business grow in terms of managed services. And as we talk about more outcome-driven engagements – less just providing raw access to the technology – the managed service really bridges the gap in terms of the various integration points that need to be managed along the way. And so it’s not just simply providing the infrastructure and application support. We’re seeing the managed service domain, especially around SAS – where this is not a one-size-fits-all approach – really extrapolate into “can we help you really derive your outcome” with expertise in either transformations of data, or we’re providing models now in terms of a service offering, in addition to consulting work of building models custom to each application. So that’s really evolving quickly. Robert Dutt: One of the trends that we follow a lot is this move across the industry to look at partners less as a direct, straight-through channel and more as an ecosystem – a lot more multi-partner engagements, especially given where you guys sit in the complexity and custom nature of a lot of what customers are asking of you. How are you guys thinking about that ecosystem, multi-partner play? Ryan MacDonald: I think the list of partners is generally growing as we talk about extrapolating into channel and SAS’s ambition to have, as you stated, 30% of our revenue flowing through the channel in Canada. I think the customer really dictates the specific mix. And so customers in large enterprise have a preference of GSI and specific domains. And what we’re seeing more is the introduction of niche players alongside GSIs, where typically that was binary previously. They would typically – let’s say they work with Deloitte or EY, for example – that would be their preference to continue in that direction. And now we’re seeing them want to leverage the scale those organizations offer, but really like the thought leadership and expertise delivered by a niche partner, and want to bring us all together. So we’re seeing a lot more partners enter the conversation, which I think is very healthy for the competitive domain and just in terms of getting to specific outcomes very quickly. Robert Dutt: The traditional sweet spot for SAS has been clearly enterprise, and Canada’s a very SMB-heavy nation, obviously. But a lot of the stuff that’s going on right now between the Viya SaaS model and the stuff going up on GitHub and the move towards managed services suggests that there might be even more of a mid-market play than before. I’m curious what you see in terms of what a Canadian reseller can realistically and credibly pursue right now. Ryan MacDonald: That has been the way the economy has been structured in Canada for decades, of course, and something that I think our channel strategy really celebrates and prioritizes. SAS – it’s hard to work both ends of the spectrum. And so our legacy of working with enterprise customers, to explore some of the topics we’ve covered in the regulatory domain and how that takes shape, the reach to SMB customers has been something that we’ve candidly struggled with at times. The channel is really the resolution to that. So we’re seeing, as we talk about more entities in this space, the mix of consulting partners or partners in general proliferating – that’s really where we’re seeing it, down more towards the SMB segments, less on the enterprise side. Robert Dutt: Acknowledging that there’s going to be a wide range of things here, and it may even depend partner to partner, but looking at the channel as an aggregate – what do you need more of from your partners right now in terms of areas of focus, in terms of opportunities to be going at, in terms of skillsets? Ryan MacDonald: I think because we are trying to aggressively pursue this market in Canada and service this customer base – which, again, the channel is just better suited for, all around – to me, it’s the feedback loop. That’s something that we challenge, of course, our frontline in an enterprise setting. You have a consistent flow of communication that’s bidirectional. We’re getting feedback on what’s important to them, what they are doing with the platform at times in our tool sets. And having that flow through an additional intermediary is an additional step in the process in the channel segment. But I think that’s really important – just to make sure we’re collecting feedback not just from channel partners, but direct from customers – their experience with SAS, how our channel partners feel in terms of support and enablement, pricing and mechanics and the rest of it as well. Robert Dutt: Curious what you see success at SAS Canada looking like over the next 12 to 18 months. What are the conversations you want to be having that you aren’t yet? What are the measurements that you’re looking at? Ryan MacDonald: We have been growing the business – in terms of revenue, of course, is always important to us – but influence in the market, I think, is something else. SAS, having such a – as we celebrate 50 years – our legacy is something we’re incredibly proud of. It’s afforded us the opportunity to build these great partnerships in Canada, all across the country, various enterprises. I think at times the double-edged sword there is they may equate us to the way they had built with SAS previously and don’t necessarily take stock of some of the things you’re seeing us bring to market today and announcing here at Innovate. So I think that is really what we look for – not just in terms of revenue growth and are we delivering more outcomes and scaling the progress with these customers. Are we really – are they delivering within the new framework? Are we changing the narrative in terms of what they see from SAS and who we are to them? Robert Dutt: My last and definitely most important question – how many dinners did you have last night? Ryan MacDonald: I had one dinner. Robert Dutt: One? One dinner. Oh, that’s an accomplishment. I appreciate you taking the time, Ryan. Thanks. Ryan MacDonald: Thank you, Robert. Really, really nice to meet you here today. Thank you, I appreciate your time. Robert Dutt: There you have it – Ryan Macdonald from SAS Canada. I’d like to thank Ryan for his time. This was our first in-person recording with the new setup, and I think you can hear the difference. And thank you for listening. A few things I’m taking away from this one. First – the AI governance story in Canada is moving faster than it might look from the outside. Ryan’s framing stuck with me: the volume of models organizations are pushing through validation has gone from two a week to five to ten a day. The governance framework isn’t a compliance tax – it’s the operational infrastructure that makes any of this scalable. And for Canadian financial services firms, OSFI E-21 isn’t on the horizon anymore – it’s here. Second – SAS’s competitive argument is more interesting than the standard “we’ve been around longer” play. The pitch is that there’s already a business-critical decisioning layer in your organization that’s been built on SAS. And the real question is whether you’re going to upgrade and grow from that investment, or build something new from scratch alongside it. For a lot of Canadian enterprises, that’s a conversation worth having. And third – Ryan was candid that the direct sales model doesn’t reach the SMB, and the channel is the answer. What’s interesting is where the growth is coming from – niche, industry-specific partners alongside the big GSIs, with customers already wanting both in the room. If you’re a Canadian reseller or systems integrator with deep vertical expertise, SAS is worth a conversation. We’ll be back tomorrow with more from on the ground here at SAS Innovate 2026, as we chat with the global channel chief at SAS Institute, John Carey [CHECK: transcript rendered as “John Kerry” – confirm on playback before publishing]. If you found this one useful, follow or subscribe to In The Channel from ChannelBuzz.ca. We’re on Apple Podcasts, Spotify, YouTube, and most of the major directories. Ratings and reviews are always appreciated and genuinely help other people in the channel find the show. Until next time, I’m Robert Dutt for ChannelBuzz.ca, and I’ll see you in the channel.

Your biggest advantage in marketing right now isn't better ads. It's understanding what actually makes people buy — and it's probably not what your feed is telling you.Rage bait is everywhere. It gets views. It gets engagement. But it doesn't build trust — and it definitely doesn't drive real revenue in B2B.In this episode, we sit down with Jason Levin, co-founder of Memelord.com, to break down why meme marketing is quietly outperforming rage bait, how humor builds trust with high-value buyers, and the exact systems top marketers are using to scale meme-driven acquisition.The deeper insight: the best marketers aren't chasing attention — they're engineering relatability at scale.You'll learn how to operationalize memes across multiple accounts, why “remixing” is the real creative advantage, and how to turn humor into a repeatable growth engine.If you're thinking about distribution in 2026, this is a playbook most companies still aren't using.GuestJason Levin — co-founder of Memelord.com, an AI-powered meme marketing platform helping companies scale humor, distribution, and content velocity through AI-generated memes and multi-account social strategies.LinkedIn: https://www.linkedin.com/in/iamjasonlevinX: https://x.com/iamjasonlevinWhat You'll LearnWhy rage bait drives views… but fails to convert high-value customersThe difference between attention farming and buyer-driven attentionHow meme marketing builds trust faster than traditional contentWhy humor is a lever — not a strategy replacementHow to run multiple niche meme accounts for different ICPsWhy remixing content beats originality in modern distributionHow AI is enabling meme velocity at scaleWhy relationships still outperform automation in closing dealsTimestamps00:00 - Introduction to Meme Marketing00:21 - Guest Introduction: Jason Levin from Memelord.com00:41 - Memes vs Rage Bait Marketing01:13 - Tactical Meme Marketing Strategies02:24 - The Importance of Branding and Trust03:27 - Rage Bait vs Smart Bait Philosophy05:01 - Why Meme Marketing Drives Revenue06:43 - Building Trust in B2B Through Humor08:13 - Niche Meme Accounts and High-LTV Distribution10:19 - The Problem with Rage Bait Culture in Silicon Valley15:00 - Inside Memelord.com: Product, Demo & AI Tools30:43 - Scaling Distribution, Verified Orgs & MeasurementKey Topics & Insights1. Rage Bait Gets Attention — But Not RevenueThere's a growing belief that anger = growth.But here's the reality:Rage bait attracts the wrong audience.It pulls in:Low-intent usersPeople looking to argueLow purchasing-power audiencesThe problem: High-value buyers don't respond to manipulation — they recognize it.And when trust is broken, conversion dies.The takeaway: Views are not revenue.2. Meme Marketing = Relatability at ScaleMemes work because they create instant recognition.Instead of forcing attention, they generate:Emotional alignmentShared pain pointsFast trust-building through humorWhen people feel understood, they convert faster.3. Humor Is a Lever, Not a StrategyMemes don't replace strategy — they amplify it.Smart marketing stacks multiple levers:Educational contentLong-form trust buildingPaid acquisitionHumor as distribution acceleration4. Remixing Is the Real Growth EngineModern content velocity comes from remixing, not originality.Instead of creating from scratch:Take what's already trendingApply your ICP's pain pointAdd context and distributionThis is how meme engines scale.5. Multi-Account Distribution StrategyScaling meme marketing requires fragmentation:Multiple niche accountsEach targeting a specific personaEach speaking in a tailored voiceThis creates parallel distribution channels instead of relying on one brand feed.6. Verified Org Arbitrage on XA key growth hack discussed:$1,000/month for verified orgAbility to spin affiliate meme accountsNetworked distribution across accountsThis creates ubiquity and compounding reach.7. Relationships Still Close RevenueEven in a world of automation:ConversationsTrustLong-term relationshipsstill outperform pure distribution hacks.8. Measurement Shift: Branded SearchInstead of tracking vanity metrics:Focus on branded search growthUse Google Search ConsoleMeasure demand creation, not just clicksThis becomes the true signal of market pull.9. The Meme Stack Is Becoming a SystemMemelord.com represents a shift:Trend detectionAI generationMulti-account publishingRapid iteration loopsMemes are no longer content — they are infrastructure.SponsorToday's episode is brought to you by Graphed – an AI data analyst & BI platform.With Graphed you can:Connect data like GA4, Facebook Ads, HubSpot, Google Ads, Search Console, AmplitudeBuild interactive dashboards just by chatting (no Looker Studio/Tableau learning curve)Use it as your ETL + data warehouse + BI layer in one placeAsk:“Build me a stacked bar chart of new users vs. all users over time from GA4”…and Graphed just builds it for you.

In this episode, host Sandy Vance and Andy Fanning, the CEO & Co-founder of Optura.ai, sit down to talk about why so many healthcare organizations are making AI headlines without actually transforming anything. Andy breaks down how Optura helps payers, providers, and life sciences companies move from a scattered list of AI ideas to a prioritized, production-ready roadmap with measurable return on investment. From crowdsourcing use cases across an entire organization to aligning AI investments with executive strategy, this episode is packed with practical insight for any healthcare leader who wants AI to actually move the needle. In this episode, they talk about: Most healthcare leaders feel they are not moving fast enough with AI, despite the headlines Shadow AI is just an unmet need, and governance is the answer Crowdsourcing AI ideas from the bottom up reveals hotspots that leadership often cannot see Aligning AI use cases to existing strategic initiatives makes adoption dramatically easier Point solutions do not share context, and that missing context is where the real value lives Return on AI investment requires defining what value actually means for your specific organization Agentic AI is the next big wave, and organizations need to decide where they sit on the risk spectrum Trust at the frontline is built by showing workers how AI follows their own standard operating procedures If finance cannot see the ROI, they will conclude that AI does not work A little about Andy: Andy is the Co-Founder and CEO of Optura.ai, where he's on a mission to help healthcare organizations stop dabbling in AI and start seeing real returns from it. His team built an AI Orchestration Platform designed from the ground up for healthcare, giving organizations the infrastructure, trust, and clarity to turn AI ambition into measurable ROI. The platform does it all in one place: spotting high-value AI opportunities, building and deploying custom agents, unlocking data without the ETL headache, auto-generating workflows from existing SOPs, and tracking return on AI investment in real time. No black boxes, no guesswork. Just AI that actually proves its worth.

Send a textIn this episode of the Sidecar Sync, Mallory and Amith dive deep into two seismic shifts rocking the AI landscape for associations: the rise of “agent tollgates” in SaaS platforms and the growing security concerns around Model Context Protocol (MCP). Sparked by comments from HubSpot's CEO about monitoring and monetizing agent access to customer data, the conversation explores what happens when vendors start charging for AI agents to access “your” data—and why this may signal a broader shift in the software business model. Amith unpacks why true data ownership matters more than ever and explains how AI data platforms eliminate traditional ETL bottlenecks while preserving control. Then, the duo pivots to MCP security risks, breaking down real-world attack vectors—from prompt injection to supply chain compromises—and offering practical guardrails for safe experimentation. The message is clear: embrace AI boldly, but build with governance, ownership, and security top of mind. 

Oil and gas companies generate enormous volumes of operational, geological, and production data. Despite this abundance, much of that data remains fragmented, inconsistent, and difficult to trust. Teams often spend a significant portion of their time preparing datasets rather than analyzing them. The result is delayed decision-making, inflated costs, and reduced operational agility. The core complication lies in data quality, data governance, and data readiness. Duplicate records, null values, drift, and structural inconsistencies make it difficult to move quickly from raw data to actionable insight. Asset teams frequently work semi-independently, each rebuilding transformation processes from scratch. Without reliable data foundations, scaling analytics, automation, or advanced modelling becomes difficult and costly.  In this episode, I'm in conversation with Shravan Gunda, CEO of Kaarvi, to discuss how a structured approach to data ingestion, anomaly detection, ETL transformation, and data lineage can reduce time-to-insight from weeks to hours. He outlines how upstream teams can standardize workflows, support governance requirements such as SOC 2, and deploy platforms either on-premises or via SaaS. Clean, trusted data is a prerequisite for accelerating analytics and enabling more advanced digital capabilities.

If you've ever wondered how Oracle Database really works inside AWS, this episode will finally turn the lights on.   Join Senior Principal OCI Instructor Susan Jang as she explains the two database services available (Exadata Database Service and Autonomous Database), how Oracle and AWS share responsibilities behind the scenes, and which essential tasks still land on your plate after deployment.   You'll discover how automation, scaling, and security actually work, and which model best fits your needs, whether you want hands-off simplicity or deeper control.   Oracle Database@AWS Architect Professional: https://mylearn.oracle.com/ou/course/oracle-databaseaws-architect-professional/155574 Oracle University Learning Community: https://education.oracle.com/ou-community LinkedIn: https://www.linkedin.com/showcase/oracle-university/ X: https://x.com/Oracle_Edu   Special thanks to Arijit Ghosh, Anna Hulkower, Kris-Ann Nansen, Radhika Banka, and the OU Studio Team for helping us create this episode.   ------------------------------------------------------------   Episode Transcript:   00:00 Welcome to the Oracle University Podcast, the first stop on your cloud journey. During this series of informative podcasts, we'll bring you foundational training on the most popular Oracle technologies. Let's get started! 00:26   Lois: Hello and welcome to the Oracle University Podcast! I'm Lois Houston, Director of Communications and Adoption with Customer Success Services, and with me is Nikita Abraham, Team Lead: Editorial Services with Oracle University.  Nikita: Hi everyone! In our last episode, we began the discussion on Oracle Database@AWS. Today, we're diving deeper into the database services that are available in this environment. Susan Jang, our Senior Principal OCI Instructor, joins us once again.  00:56 Lois: Hi Susan! Thanks for being here today. In our last conversation, we compared Oracle Autonomous Database and Exadata Database Service. Can you elaborate on the fundamental differences between these two services?     Susan: Now, the primary difference is between the service is really the management model. The Autonomous is fully-managed by Oracle, while the Exadata provides flexibility for you to have the ability to customize your database environment while still having the infrastructure be managed by Oracle.   01:30 Nikita: When it comes to running Oracle Database@AWS, how do Oracle and AWS each chip in? Could you break down what each provider is responsible for in this setup?  Susan: Oracle Database@AWS is a collaboration between Oracle, as well as AWS. It allows the customer to deploy and run Oracle Database services, including the Oracle Autonomous Database and the Oracle Exadata Database Service directly in AWS data centers.   Oracle provides the ability of having the Oracle Exadata Database Service on a dedicated infrastructure. This service delivers full capabilities of Oracle Exadata Database on the Oracle Exadata hardware. It offers high performance and high security for demanding workloads. It has cloud automation, resource scaling, and performance optimization to simplify the management of the service.  Oracle Autonomous Database on the dedicated Exadata infrastructure provides a fully Autonomous Database on this dedicated infrastructure within AWS. It automates the database management tasks, including patching, backups, as well as tuning, and have built-in AI capabilities for developing AI-powered applications and interacting with data using natural language. The Oracle Database@AWS integrates those core database services with various AWS services for a comprehensive unified experience.  AWS provides the ability of having a cloud-based object storage, and that would be the Amazon S3. You also have the ability to have other services, such as the Amazon CloudWatch. It monitors the database metrics, as well as performance. You also have Amazon Bedrock. It provides a development environment for a generative AI application.   And last but not the least, amongst the many other services, you also have the SageMaker. This is a cloud-based platform for development of machine learning models, a wonderful integration with our AI application development needs.  03:54 Lois: How has the work involved in setting up and managing databases changed over time?  Susan: When we take a look at the evolution of how things have changed through the years in our systems, we realize that transfer responsibility has now been migrated more from customer or human interaction to services. As the database technology evolves from the traditional on-premise system to the Exadata engineered system, and finally to the Autonomous Database, certain services previously requiring significant manual intervention has become increasingly automated, as well as optimized.  04:34 Lois: How so?  Susan: When we take a look at the more traditional database environment, it requires manual configuration of hardware, operating system, as well as the software of the database, along with initial database creation. As we evolve into the Exadata environment, the Exadata Database, specifically the Exadata cloud service, simplifies provisioning through web-based wizard, making it faster and easier to deploy the Oracle Database in an optimized hardware.     But when we move it to an Autonomous environment, it automates the entire provisioning process, allowing users to rapidly deploy mission-critical databases without manual intervention, or DBA involvement. So as customers move toward Autonomous Database through Exadata, we have fewer components that the customer needs to manage in the database stack, which gives them more time to focus more on important parts of the business.  With the Exadata Database, it provides a co-management of backup, restore, patches and upgrade, monitoring, and tuning. And it allows the administrator the ability to customize the configuration to meet their very specific business needs. With Autonomous Database, it's now fully automated and it's a greater responsibility is shift toward the service. With Autonomous Database on dedicated infrastructure, it provides that fine-grained tuning more for Oracle to help you perform that task.  06:15 Nikita: If we narrow it down just to Oracle and AWS for a moment, which parts of the infrastructure or day-to-day ops are handled by each company behind the scenes?  Susan: When we take a look at Oracle Database@AWS, it operates under a shared responsibility model, dividing the service responsibilities between AWS, as well as Oracle, as well as you, the customer.   The AWS has the data center. Remember, this is where everything is running. The Oracle Database@AWS, the Oracle Database infrastructure may be managed by Oracle and run in OCI, but is physically located within the AWS regions, as well as the availability zones and the AWS data centers.  The AWS infrastructure, in this case, is AWS's responsibility to secure the environment, including the physical security of the data center, the network infrastructure, and the foundational services like the compute, the storage, and the networking, all within AWS.  The next thing of who's responsible for the shared responsibility, it's Oracle. And that would be the hardware. We provide the hardware. While the hardware may physically reside in the AWS data center, Oracle's Cloud Infrastructure operational team will be the one managing this infrastructure, including software patching, infrastructure update, and other operations through a connection to OCI. This means Oracle handles the provisioning, as well as the maintenance of any of the underlying Exadata infrastructure hardware.  When we take a look at the next thing that it manages, it is also responsible besides the infrastructure of the Exadata. It is also the ability to manage the hardware, the environment of that hardware through the database control plane. So Oracle manages the administration and the operational for the Oracle Database@AWS service, which resides in OCI. So this includes the capabilities for management, upgrade, and operational features.  08:37 Nikita: And what are the key things that still remain on the customer's plate?   Susan: If you are in an Exadata environment or in an Autonomous environment, it is you, the customer, who is responsible for most of the database administration operation, as well as managing the users and the privileges of the user to access the database. No one knows the database and who should be accessing the data better than you.  You will be responsible for securing the applications, the data of the database, which now allows you to define who has access to it, control the data encryption, and securing the application that interacts with the Oracle Database@AWS.  09:29 Lois: Susan, we've talked about both Autonomous Database and Exadata Database Service being available on Oracle Database@AWS, but what's different about how each works in this environment, and why might someone pick one over the other?  Susan: Both databases, even though they run on the same Exadata Cloud Infrastructure, both can be deployed on both public cloud, as well as the customer data center, which is Oracle Cloud@Customer.  The Autonomous Database is a fully managed, completely automated environment. And this provides a capability of having a fully Autonomous Database Service running on a dedicated Oracle Exadata Infrastructure within your AWS data center.  The Exadata is a service that is provided and managed by Oracle and is physically running in the AWS data center, but is designed for mission critical workload and includes RAC environment, Real Application Cluster, offering a high performance availability and full feature capability that is similar to other Exadata environment, such as those running in our customers' data center.  The primary difference is really between the two services. When you take a look at the Exadata, the customer only pays for the compute resources that is used. Autoscaling can be used for a variety or variable resources, the workload, to automatically scale to the compute resources up or down when required.  The Autonomous Database also has automatic optimization for data warehousing, transaction processing, as well as JSON workload. The Exadata service, the customer again, also pays for the compute resources that they allocate. But that's the key thing. The customer can initiate the scaling because it's very specific to the workload that is needed.  So when you take a look at the two database services, one gives the ability to let Oracle fully manage it, including the scaling capability. The other, the Exadata, provides you the capability of having the environment that it's running on the infrastructure be managed by Oracle that adds a database administrator. You may wish to have a little bit more granular control of how you want the database to not only be scaling, but how you wish to customize how the database will be running.  12:10 Nikita: Focusing on Autonomous Database for a moment, what should teams know about how it actually runs within AWS?   Susan: The Autonomous Database on the Oracle Database@AWS brings the power of the Oracle's self-managing, self-securing, and self-repairing database into your AWS environment.  It provides the capability of the database automatically, automates many of the traditional, complex, and time-consuming database management tasks, such as the provisioning of the database, the patching, the backing up, and the scaling, and the performance tuning, reducing the need for any manual intervention by the database administrator.  Running the Autonomous Database in your AWS region enables low latency access for your AWS applications and services that is deployed within AWS, thus improving performance and response time. With the Autonomous Database, it automates many of the traditional things that is now automatically done by Oracle. It also supports integration with various AWS services, such as the ability of the not in addition to AIM, but the cloud formation, the CloudWatch for monitoring and the S3 for the storage.  You can easily migrate existing Exadata workload, including those running on Oracle RAC to AWS with minimum or no change to any of your databases or applications. In addition, there's a really powerful capability and feature of the database is called zero ETL, and that's zero extract, transformation, and load.  It's an integration capability with services like your Amazon Redshift, enabling near real time analytics and machine learning on your transactional database that is stored within the Autonomous Database on in your AWS environment. So with the Autonomous Database, it checks off many of the boxes for automatic capability, securing, tuning, as well as scaling the database.  With the Autonomous Database in the Dedicated Exadata Infrastructure, the Exadata Cloud Infrastructure resource represents the physical system, which can be expanded with storage, as well as compute services, the compute host. This now provides the ability to have an isolated zone for the highest protection from other tenants. The data is stored on a dedicated server only for one customer. That would be you.  14:56 Lois: Could you explain the role of Autonomous VM? What are its primary benefits?  Susan: The virtual machine or as we refer to them as the cluster, includes the grid infrastructure and provides a private network isolation. This provides you the capability of having custom memory, core, and storage allocation.  The Oracle Grid Infrastructure includes the Oracle Clusterware, which manages the cluster, as well as the servers, and ensure that the database can failover to another server in case of any failure.  15:34 Be a part of something big by joining the Oracle University Learning Community! Connect with over 3 million members, including Oracle experts and fellow learners. Engage in topical forums, share your knowledge, and celebrate your achievements together. Discover the community today at mylearn.oracle.com.  15:55 Nikita: Welcome back! Susan, what is the Autonomous Container Database?  Susan: With the Autonomous Container Database, and you need that if you're going to create an Autonomous Database, you need to provision that within your Autonomous Exadata VM Cluster. It serves as a container to hold or to house one or more Autonomous Databases.  This allows multiple Autonomous Databases to coexist in the same infrastructure while still being logically separated. And this allows for the separation of databases based on their intended use. Think of a database for production. Think of a database for development. Think of a database for testing. You may have different database versions within the same infrastructure.  This isolation makes it easier for you to be able to meet your SLA, your Service Level Agreement, any long-term backups you may have, very specific encryption key needs to prevent issues from one database impacting another. So, the ability to have everything be isolated and secure is still grouping it in a manner that will meet your business needs.  17:08 Lois: Looking at Exadata Database Service specifically, what are some standout advantages for customers who deploy it on Oracle Database@AWS? Is there anything in particular they should get excited about in terms of performance or integration with AWS?  Susan: The Exadata Database Service is running on a dedicated Exadata Infrastructure that's deployed within your AWS data center. It delivers the same Exadata service experience in cloud control planes as the Oracle Cloud Infrastructure, allowing you to leverage existing skills and processing across your multi-cloud environment.  It addresses the data resiliency, or residency rather. And that's the scenario where many of our customers has the need. You have a need because of your security compliance to have the data local to you. By having the Exadata Database in your Oracle Database@AWS, it is running in your data center. So, this addresses that very important need, data residency, to have it close to you.  It also allows for seamless integration with other AWS services and applications. So now you have a capability of a hybrid cloud architecture leveraging the benefit of both Oracle Exadata and your AWS system. It has built-in high availability, the RAC application cluster, as well as Data Guard, a capability of addressing disaster recovery capability.  This also provides the ability for you to scale your compute, as well as your storage and your I/O resources independently. So as mentioned with Exadata, you have flexibility of how you want your database to be running individually. So just like the Autonomous, the Exadata Database checks off many of the boxes for running a mission-critical with high availability, highly redundant hardware and software features, along with extreme performance, scalability, and reliability.  This now allows you to run your AI environment, your online transaction processing, your analytic workload on any scale on the Exadata Infrastructure running in the Oracle Cloud. And in this case, running in your data center.  19:45 Nikita: If a business suddenly needs more capacity, how does scaling work with Exadata Database Service versus Autonomous Database on Oracle Database@AWS?   Susan: So with the Exadata scaling, you now can scale to meet expected demands so you know at certain point I will need more. I will then ask it to scale at that point when I will assign it-- and I'm using an example, I will assign it three computer cores all the time. But there may be demands. Think of your end of the quarter, end of the year processing that you may need more. So, you are enabling the compute cores to scale at the time you need it.  And what's cool is it will then, when it's no longer needed, it will then scale back down to the original three cores that you assign. So, you only pay for the enabled cores. But what's very cool about the Autonomous is that it is real-time scaling. So, with Autonomous, now you have the capability using Autonomous Database since it is self-tuning, self-monitoring, the Autonomous Database actually monitors the workload requirement and scales to match the workload demand.  Once the minimum level of the compute is defined and enabled, the automatic scaling is set. Autonomous Database will adjust to the consumption when it's needed, and it will scale back down when it's not. So though the Exadata is pretty cool, it will scale up and down on the workload demand.  This is with the Autonomous is even more powerful. It is real-time scaling based on that usage at that moment. Built-in automatic increase to meet the workload demands when it spikes and it automatically scales back when it's not needed.  A very powerful capability with all of our Oracle databases, the ability, even with traditional, to allow you to define what you may need with Exadata scaling for peak demands, as well as Autonomous scaling for real-time consumption and scaling when needed.  When you look at all of our options, one of the key things to bear in mind is a phrase that we use: performance scale as more servers are added. And what this is really saying is Oracle's automated scaling ability for the database, it basically has the ability to maintain or improve its performance under increased workload by automatically adding computational resources when needed.  This process is also known as horizontal scaling. It involves adding more servers, compute instances, to a cluster to share the processing load. And it has that capability automatically.  22:53 Nikita: There's so much more we can discuss about Oracle Database@AWS, but let's pause here for today! Thank you so much Susan for joining us.  Lois: Yeah, it's been really great to have you, Susan. If you want to dive deeper into the topics we covered today, go to mylearn.oracle.com and search for the Oracle Database@AWS Architect Professional course. Until next time, this is Lois Houston…  Nikita: And Nikita Abraham, signing off!  23:23 That's all for this episode of the Oracle University Podcast. If you enjoyed listening, please click Subscribe to get all the latest episodes. We'd also love it if you would take a moment to rate and review us on your podcast app. See you again on the next episode of the Oracle University Podcast.

Send us a textIn this episode of Exchanges with Hitachi Solutions, host Ginny Lebeck joins Matt Volke, Chris Satterly, and Mark Shoesmith to preview FabCon 2026 in Atlanta. The conversation explores why FabCon has become a must-attend event for data and AI professionals, how customer expectations around data have evolved, and what organizations need to succeed on their AI journey. Drawing from real-world client conversations and healthcare use cases, the panel discusses Microsoft Fabric as a unifying data platform, the importance of strong data foundations and governance, and how emerging capabilities—from low-code tools to advanced analytics—are enabling teams to move faster and unlock more value from their data. You can find Hitachi Solutions at Fabcoon Booth #545 March 18-20.Highlights:FabCon brings together customers, partners, and Microsoft experts to share practical insights on data management, analytics, and AI, with a strong emphasis on real-world business challenges.Organizations are increasingly treating data as a strategic asset rather than a byproduct, focusing on real-time insights, advanced analytics, and AI readiness.Microsoft Fabric is emerging as a key enabler, simplifying data ingestion, modeling, and curation while supporting both self-service analytics and advanced data science workloads.Healthcare organizations are adopting Fabric to modernize legacy, on-premises data platforms, reduce manual ETL effort, and better handle diverse and high-volume data sources.Tools like pipelines, copy jobs, mirroring, and Copilot-assisted development help small data teams work more efficiently and respond faster to analytics demands.Strong data models, governance, and curated information are essential prerequisites for successful AI and advanced analytics initiatives.Hitachi Solutions' presence at FabCon focuses on connecting customers with experienced data engineers, data scientists, and industry experts to help translate technology into outcomes.global.hitachi-solutions.com

La France est inquiète des conséquences de la politique prédatrice américaine dans la Caraïbe. Et la question se pose de savoir si Paris peut compter sur sa présence en Guyane pour peser sur le continent. La Commission des Affaires étrangères de la Défense et des Forces armées du Sénat rendait publique au début du mois de janvier 2026 un rapport de mission. La visée américaine sur le Groenland et l'enlèvement de Nicolas Maduro au Venezuela sont autant de sources d'inquiétudes. Les effets de la doctrine Monroe, invoquée pour justifier la domination des États-Unis dans cet hémisphère et l'émergence de nouveaux États pétroliers aux immenses réserves que sont le Guyana et le Suriname font craindre des risques de déstabilisation, auxquels s'ajoutent le narcotrafic et l'orpaillage illégal. Les territoires français de la Caraïbe -Guadeloupe, Martinique, Saint-Martin, Saint-Barthélémy et Guyane sont aujourd'hui au cœur d'un enjeu géopolitique et sécuritaire d'ampleur. La souveraineté française s'y négocie au quotidien. Invités : Fred Constant, professeur des Universités en Science politique à l'Université des Antilles. Auteur de « Géopolitique des Outre-mer », aux éditions le Cavalier bleu et « Atlas des Outre-mer », aux éditions Autrement Yannick Chenevard, officier supérieur de réserve. Député du Var, rapporteur du budget de la Marine et de l'exécution de la loi de programmation militaire. Chercheur associé au Lab'HOMERe Patrick Roger, ancien journaliste au quotidien Le Monde, auteur de « Nouvelle-Calédonie, la Tragédie », récompensé par le Prix des Députés 2025. Et « L'archipel de la discorde. Paris-Nouméa. Demain le Pacifique », aux éditions du Cerf. 

Les Outre-mer du globe traversent une période difficile, attentifs qu'ils sont aux menaces qui pèsent sur le Groenland, pays constitutif du royaume du Danemark et territoire associé à l'Union européenne.  De fait, ils sont nombreux les Outre-mer contemporains, produit résiduel de l'expansionnisme colonial qui a poussé en son temps des nations rivales à étendre leurs frontières au-delà des mers et des océans, ces territoires offshore se situant souvent dans des zones névralgiques pour les échanges mondiaux. Engagé à partir du XVIIè siècle, ce mouvement avait pour but de satisfaire les intérêts économiques de ces nations et promouvoir leurs idéaux politiques et religieux. Et voilà qu'il refait surface sur fond de velléités de Donald Trump de mettre la main sur le Groenland. Une démarche qu'on peut imaginer être observée de près et dans un certain silence, par le grand rival des États-Unis qu'est la Chine et qui n'en est pas moins active elle aussi dans d'autres zones et de manière moins ouverte….quoique. Géopolitique se saisit de cette occasion pour jeter un regard sur les Outre-mer français -dans les Caraïbes, le Pacifique, l'océan Indien et l'Atlantique Sud- qui fondent une large part du rayonnement mondial de la France et qui se trouvent confrontés à deux situations qui, parfois, ne sont pas sans lien : entre mouvements de contestation au sein même de ces territoires et stratégies de déstabilisation portées par divers acteurs internationaux. Regard sur les vulnérabilités de l'Outre-mer français ou comment, hier, marginal cet espace est devenu stratégique. Invités : Fred Constant, professeur des Universités en Science politique à l'Université des Antilles. Auteur de « Géopolitique des Outre-mer », aux éditions le Cavalier bleu et « Atlas des Outre-mer », aux éditions Autrement Yannick Chenevard, officier supérieur de réserve. Député du Var, rapporteur du budget de la Marine et de l'exécution de la loi de programmation militaire. Chercheur associé au Lab'HOMERe Patrick Roger, ancien journaliste au quotidien LE MONDE, auteur de « Nouvelle-Calédonie, la Tragédie » récompensé par le Prix des Députés 2025. Et « L'archipel de la discorde. Paris-Nouméa. Demain le Pacifique », aux éditions du Cerf.

Your “source of truth” for customer acquisition isn't GA4. It's what people tell you when they sign up — and right now, that story is changing fast.In this episode, we unpack a simple but brutally effective tactic: adding a required “How did you hear about us?” field to your signup form — and using that data to understand where real discovery is happening. The surprise? More and more B2B customers are saying social media, even when analytics tools claim otherwise.But here's the deeper shift: organic social is hard to measure… unless you track the right trailing indicator. That indicator is branded search.You'll learn how to use Google Search Console to track brand-name impressions over time, why it's becoming the only KPI that matters for modern founder-led marketing, and how branded search creates a defensible moat competitors can't easily steal.If you're planning your marketing strategy for 2026, this is the measurement system you need.What You'll LearnWhy signup form attribution is often more reliable than your analytics dashboardsThe biggest B2B acquisition shift happening right now: from search → socialWhy organic social is nearly impossible to ROI… and how to measure it anywayThe “branded search” metric that acts as a trailing indicator for social discoveryWhy branded search is a marketing moat your competitors can't take from youHow to build a branded-search chart using Google Search Console in minutesThe exact prompt to pull branded impressions by query and track them over timeTimestamps00:00:00 - Customer Discovery Starts at Signup00:00:10 - The Shift: Search → Social00:00:31 - Why Organic Social Now Matters Most00:00:52 - The Measurement Problem (and the Fix)00:01:12 - Branded Search = Your Trailing Indicator00:01:33 - Why Branded Search Is a Moat00:01:54 - Where to Invest Time, Money, and Energy00:02:04 - The 2026 Strategy: Grow Brand Searches00:02:15 - How to Track Branded Search in GSC00:02:25 - Building the Branded Impressions Chart00:02:46 - Live Demo: Google Search Console Setup00:03:07 - Final ThoughtsKey Topics & Insights1. Signup Attribution Beats Analytics (Almost Every Time)One of the fastest ways to understand how customers actually found you is simple: add a required “How did you hear about us?” field in your signup form.Why it works:It captures customer intent in their wordsIt reveals channels analytics often misattributesIt shows the real discovery story (not the last-click story)And the punchline: it often contradicts what GA4 says.2. The B2B Discovery Shift: Search → SocialIf you've been paying attention to the data, something big is happening:People aren't discovering new software products through search anymore. They're discovering them on social — then Googling them afterward.This shift has accelerated over the past 12–18 months. Even in B2B, where trends typically lag behind DTC.What this means:SEO is no longer the first touchpointSocial is becoming the top-of-funnel discovery engineSearch is evolving into a validation channel3. Organic Social Has a Measurement ProblemThe hardest part about investing in organic social is that it's difficult to tie to ROI.Whether you're doing:Founder-led contentCreator sponsorshipsCommunity distributionOrganic growth loops…it doesn't fit neatly into traditional attribution.So instead of forcing bad ROI models, track the trailing indicator that proves social discovery is working.4. Branded Search Is the Trailing Indicator That MattersHere's the key idea:When someone discovers your product on social, they don't click your link. They Google your name.That branded search becomes the measurable proof:A discovery event happenedPeople care enough to look you upYour brand is entering the market's memoryThis is why branded search growth is one of the strongest indicators of momentum.If branded search is increasing month-over-month, your brand is winning.5. Branded Search Creates a Defensible MoatThis is where it becomes more than measurement — it becomes strategy.Branded search is difficult for competitors to steal. Once people are searching your name, you own that demand.The only way competitors can interfere:They bid on your brand in Google AdsThey try to outspend youOr they attempt to confuse the marketBut that's expensive, obvious, and usually temporary.So branded search is not only a KPI — it's defensibility.6. How to Track Branded Search in Google Search ConsoleThis is the tactical part.To track branded search over time, you want a chart that shows:Impressions over timeFor queries containing your brand nameCaptured in every format your audience might type itAnd this is surprisingly easy to pull from Google Search Console.7. The Exact Chart & Prompt to Build ItThe goal is to extract Search Console impressions where queries include your brand name.Example prompt:“Build a chart showing total impressions over time for queries containing ‘YOURBRAND'.”Then your job becomes simple:Increase branded impressions month-over-month through:social contentdistributioncreator partnershipspodcast mentionsrepeated brand exposureconsistent visibilityThis becomes the clearest signal that marketing is compounding.Action Steps (Do This Today)Add a required “How did you hear about us?” field on signupReview responses weekly (and compare against analytics)Use Google Search Console to track branded query impressionsCreate a monthly KPI: branded impressions growthUse branded search growth as the scoreboard for your organic social effortsSponsorToday's episode is brought to you by Graphed – an AI data analyst & BI platform.With Graphed you can:Connect data like GA4, Facebook Ads, HubSpot, Google Ads, Search Console, AmplitudeBuild interactive dashboards just by chatting (no Looker Studio/Tableau learning curve)Use it as your ETL + data warehouse + BI layer in one placeAsk:“Build me a stacked bar chart of new users vs. all users over time from GA4”…and Graphed just builds it for you.

If you're not getting cited by ChatGPT, your “AI SEO” strategy isn't working, no matter what your dashboards say. Most of it is observability theater: dashboards, charts, synthetic prompts — and zero actual placement.In this episode, we chat with Shawn Schneider, founder of Eldil AI, about what actually determines whether your company shows up in ChatGPT answers. The short answer: LLMs don't reward more content, clever prompts, or prettier dashboards. They reward a small set of trusted third-party sources — and most brands aren't mentioned in any of them.Shawn breaks down why observability alone creates a false sense of progress, how to identify the specific citations that dominate your category, and how to turn that insight into real placements through outreach and negotiation. We also unpack why Google Search Console is still the best signal we have for AI-driven queries, how to prioritize the one citation that actually matters, and what the first 30–90 days can look like when you do this correctly.GuestShawn Schneider — founder of Eldil AI, a GEO / AI SEO platform focused on identifying and securing the citations LLMs rely on most; helps brands and agencies win visibility in ChatGPT by targeting the power-law sources that shape AI answers.Guest LinksLinkedIn: https://www.linkedin.com/in/shawn-schneider-61b2b5207/ Company Website: https://www.eldil.ai/What You'll LearnWhy most GEO / AI SEO observability tools are meaningless without actual placements The only thing that reliably improves AI search visibility: citation placementsHow to use Google Search Console to surface AI fan-out queriesWhy synthetic prompt data is still unreliable (and what to trust instead)The power law of citations: why only 1–3 sources actually matterHow Eldil turns citation discovery into outreach and negotiated placementsWhat 30–90 days can look like when you secure the right citationWhich industries should invest heavily — and which should ignore this for nowWhy ChatGPT dominates referral traffic compared to other LLMsWhat happens when ads arrive inside AI search resultsTimestamps00:00 — GEO, AI SEO, AEO: noise vs. reality00:21 — Why observability tools don't move the needle03:55 — Where GEO tools get their data (and why it's messy)07:16 — Using Google Search Console as a prompt proxy09:40 — The three pillars: technical, content, authority12:07 — Citations as the dominant ranking lever13:07 — The power law: thousands of citations, one winner19:07 — How fast results actually show up20:39 — When building your own citation content makes sense30:41 — Which business models win with GEO37:11 — ChatGPT ads and the future of AI search41:32 — Where to find Shawn and closing thoughts Key Topics & Ideas1. Why dashboards feel good but don't create outcomes.Most tools are essentially “Google Analytics for LLMs”ChatGPT referrals rise naturally as usage increasesCharts go up even if you do nothingWithout placements, observability is just vanity2. The three common approaches in the market today:Guessing prompts with LLMsClickstream data sourced from Chrome extensions and brokersSynthetic prompts without transparencyEldil uses Google Search Console + Analytics as the best available proxy for real intent.3. How to spot AI-generated fan-out queries:50+ character queriesHigh impressionsLow or zero clicksThese often represent LLMs expanding short prompts into long-form searches.4. The three pillars: Technical, Content, AuthorityTechnical — can an LLM crawl and understand your site?Content — does useful information exist?Authority — does anyone credible back it up?Authority is the multiplier most teams ignore.5. What actually shapes AI answers:Citations are not backlinks, they are semantic explanationsLLMs repeatedly return to the same trusted sourcesThird-party listicles and niche blogs dominate citation share6. The Power Law of Citations10k–15k citations may exist200–300 matter1–3 actually move the needleIf you're not in those, content volume won't save you.7. The real workflow:Identify high-value customer questionsExtract dominant citationsRank them by weightContact site ownersNegotiate placementMonitor AI visibility and referral trafficThis is where most tools stop — and where Eldil focuses.8. How many placements do you need?Surprisingly few.You don't need 100 placementsYou need the right oneThen expand into adjacent verticalsThis is concentrated betting, not spray-and-pray SEO.9. Why GEO feels different from traditional SEO:You are inserting into sources that already rankChanges can show up in weeks, not yearsMeaningful referral growth often appears within ~60–90 days10. Who Should (and Shouldn't) Do ThisBest fit:High-ACV B2B SaaSLong buying cyclesHigh-LTV e-commerce (supplements, skincare)ICPs that already live in ChatGPTIf your customers do not use LLMs yet, start elsewhere.11. Why ChatGPT is the main eventBased on Eldil's data:ChatGPT referrals dwarf Perplexity and othersFor most companies, this is where focus belongsSmaller channels still matter for high-ticket sales12. What's coming nextPaid placements inside LLMsOrganic plus paid becoming a one-two punchCitation inventory getting expensive fastThe window for cheap dominance will not last.SponsorToday's episode is brought to you by Graphed – an AI data analyst & BI platform.With Graphed you can:Connect data like GA4, Facebook Ads, HubSpot, Google Ads, Search Console, AmplitudeBuild interactive dashboards just by chatting (no Looker Studio/Tableau learning curve)Use it as your ETL + data warehouse + BI layer in one placeAsk:“Build me a stacked bar chart of new users vs. all users over time from GA4”…and Graphed just builds it for you.

Benedikt turns a year older. Benedicte moves forward despite the curveballs.Benedikt took a week off work to celebrate his 40th birthday. He spent his birthday week with a few parties with family and friends, and seeing two concerts. On the work front, he and the team built a Snowflake integration. And with the ETL infrastructure now in place, this potentially opens doors for other integrations.Despite another major extended family upheaval, Benedicte carries on by going on morning walks and focusing on her projects. She recently shipped the new Framer plugin version, made demos, and is planning to get the documentation for the plugin.Benedikt and Benedicte talk about books, how fast the internet is breaking nowadays, and more.Mentioned on the show:Surrounded by Idiots – a book by Thomas EriksonAntifragile – a book by Nassim Nicholas TalebNonviolent Communication – a book by B. Marshall Rosenberg

Your brand doesn't exist until it ranks on page one—and most founders have no idea how to make that happen.In this episode, we break down the exact playbook for getting a brand-new domain to show up in Google for your company name. After going through this process firsthand with Graphed.com, you'll learn how to choose a rankable name, build the right backlinks, trigger branded search behavior, and use Google Ads to accelerate the whole process.If you're launching anything new, this is the tactical blueprint you wish you had earlier.What You'll LearnWhy ranking for your brand name is the first real trust signal for any startupHow to pick a name and domain you can actually rank forThe “first 100 links” strategy that trains Google to recognize your brandSimple ways to generate branded search behavior across social and contentHow Google quietly tests your domain—and how to know when it's happeningHow to use Google Search Ads to accelerate ranking and protect your brandWhy .com still matters more than any other TLDTimestamps00:00 — Why your new domain must rank for your own brand name00:31 — Why ranking for your brand name is a critical early trust signal01:03 — The rookie mistake founders make when picking a brand name01:13 — What ideal, non-competitive SERPs should look like01:35 — Graphed.com's journey to finally ranking in position one01:45 — Overview of the process to teach Google your brand01:55 — Step 1: Build backlinks to your homepage03:29 — Step 2: Drive branded search with social posts & content04:21 — Step 3: Run Google Search Ads on your exact brand name05:45 — Why you should always buy the .com for your brand06:16 — Final thoughts + Graphed free trialKey Topics & Insights1. Ranking for Your Brand Name = Early-Stage TrustIf someone Googles your company and doesn't find you, credibility collapses. Ranking for your brand name is one of the first—and easiest—trust signals to secure. Graphed.com took ~2 months to rank, but with this framework, it can happen in as little as 24–48 hours.2. How to Choose a Rankable NameAvoid names already used by active companiesLook for search results filled with noise, not competitorsIdeal: two words, few syllables, easy to spellAnd always, always buy the .com3. Build the First 100 Backlinks (Brand-Name Anchors Only)Your #1 job early is to teach Google what your company is.Do this by:Building backlinks to your homepageUsing your brand name as the anchor text (not keywords)These are foundational “identity” links that help Google map brand → domain.How to build them:Submit to software directoriesUse link submission servicesCold email companies for guest post swapsLayer PR on top later4. Trigger Branded Search BehaviorOnce Google sees your backlinks, you need humans to reinforce the signal:Search your brand nameClick your domainSpend time on the pageGoogle then learns:“When people search this name, this is the site they want.”You create this behavior through:Social postsNewslettersPodcast mentionsRepeated use of the brand everywhere5. How Google Tests Your DomainGoogle will quietly experiment by showing your domain for branded queries.You'll see this in Search Console via:Rising impressionsIncreasing CTRSudden jumps in average positionThis is the moment Google “decides” you belong on page one.6. Accelerate Everything With Google Search AdsRun a brand campaign:Exact-match brand keywordMinimum bid: around $5Send traffic to homepageThis forces the association between brand name → your site, and accelerates your rise in organic search.Brand protection tips:Raise bids to block competitorsAdd sitelinks to take more SERP real estateOptional: multiple ad accounts (with caution)7. Why .com Still Beats Every Other DomainConsumers inherently trust .com more than .io, .co, .xyz, etc.It drives higher CTR and reduces friction in word-of-mouth.If the .com isn't available, pick a new name—don't settle.SponsorToday's episode is brought to you by Graphed – an AI data analyst & BI platform.With Graphed you can:Connect data like GA4, Facebook Ads, HubSpot, Google Ads, Search Console, AmplitudeBuild interactive dashboards just by chatting (no Looker Studio/Tableau learning curve)Use it as your ETL + data warehouse + BI layer in one placeAsk:“Build me a stacked bar chart of new users vs. all users over time from GA4”…and Graphed just builds it for you.

An airhacks.fm conversation with Stanislav Bashkyrtsev (@sbashkirtsev) about: scientific software for chemists and drug discovery, peaksel flagship software for analyzing mass spectrometer data, parsing binary instrument formats up to gigabytes in size, mass spectrometry measuring molecular weights using electric fields and detectors, daltons as mass units, isotope patterns for molecule identification, storing experimental data in PostgreSQL with potential big data challenges, S3 storage solutions, drug discovery process from hit identification to molecule modifications, molecular libraries and combinatorial chemistry, enumeration of molecular structures in computers, synthesis reactions mixing reactants with solvents and various conditions, liquid handlers and laboratory automation challenges, return on investment issues in early drug discovery automation, lab of the future concepts, Molbrett product combining excalidraw with chemical structure drawing capabilities, SMILES format for representing molecular structures as strings, graph-based molecular formats storing atom connections and bond types, 2D vs 3D molecular visualization preferences, Meve centralized event system for tracking molecular experiments across different software systems, ETL processes for data integration, Crystalline software for documenting protein crystallography experiments, protein structure determination using X-ray crystallography, Synchrotron facilities for high-energy X-ray generation, crystal growing conditions and documentation, fishing crystals with microscope and lasso wands, liquid nitrogen cooling for crystal preservation, Java backend, JavaScript frontend, minimal dependencies approach, six-person team structure, sponsorship business model for open source scientific software development, free updates for sponsors, subscription model for non-sponsors, checkout: https://elsci.io Stanislav Bashkyrtsev on twitter: @sbashkirtsev

AWS Morning Brief for the week of December 1st, with Corey Quinn. Links:Protect sensitive data with dynamic data masking for Amazon Aurora PostgreSQLAmazon CloudFront announces support for mutual TLS authenticationAmazon EC2 announces interruptible Capacity ReservationsIntroducing guidelines for network scanningPractical implementation considerations to close the AI value gapEverything you don't need to know about Amazon Aurora DSQL: Part 4 – DSQL componentsSimplify data integration using zero-ETL from self-managed databases to Amazon RedshiftAutomatic quota management is now AWS Service Quotas adds support for automatic quota managementAnnouncing Amazon Route 53 Accelerated Recovery for managing public DNS recordsAnnouncing Unused NAT Gateway Recommendations in AWS Compute OptimizerAmazon EKS introduces Provisioned Control PlaneAWS Finally Lets You Find Your Idle NAT Gateways

Flavia Saldanha, a consulting data engineer, joins host Kanchan Shringi to discuss the evolution of data engineering from ETL (extract, transform, load) and data lakes to modern lakehouse architectures enriched with vector databases and embeddings. Flavia explains the industry's shift from treating data as a service to treating it as a product, emphasizing ownership, trust, and business context as critical for AI-readiness. She describes how unified pipelines now serve both business intelligence and AI use cases, combining structured and unstructured data while ensuring semantic enrichment and a single source of truth. She outlines key components of a modern data stack, including data marketplaces, observability tools, data quality checks, orchestration, and embedded governance with lineage tracking. This episode highlights strategies for abstracting tooling, future-proofing architectures, enforcing data privacy, and controlling AI-serving layers to prevent hallucinations. Saldanha concludes that data engineers must move beyond pure ETL thinking, embrace product and NLP skills, and work closely with MLOps, using AI as a co-pilot rather than a replacement. Brought to you by IEEE Computer Society and IEEE Software magazine.

There's a whole narrative right now that “vibe coding is a bubble” and all the MRR from AI-built apps isn't real.In this episode, we chat with Jacob Klug, founder of the agency Creme, which specializes in building lovable MVPs on top of tools like Lovable and AI coding assistants. Jacob makes the case that most of the “AI apps are trash” discourse is really a skill issue, not a tool issue—and he breaks down the exact process his team uses to ship full platform-level apps in two-week sprints.We dig into how to scope and design software that doesn't look AI-generated, how to think about personal operating systems vs. SaaS, why ideas are getting worse even as tools get better, and how creators and agencies can turn niche domain expertise into real products.If you're an operator, marketer, or founder trying to figure out how to actually use AI coding tools (instead of just tweeting about them), this one's for you.GuestJacob Klug — founder of Creme, an agency building “lovable MVPs” and full-stack products with Lovable + AI tools; helps founders, startups & enterprises ship production apps in weeks without sacrificing UX.Guest LinksWebsite: https://www.creme.digital/LinkedIn: https://www.linkedin.com/in/jacob-klug-37b254156/X (Twitter): https://x.com/JacobsklugWhat You'll LearnWhy the “vibe coding is a bubble” take is mostly a skill and discipline problemHow Jacob's agency ships full startup-grade products using Lovable and AIThe PRD-first formula they use before ever opening a builderHow to decide when to build vs. when to buy software in 2025Why we're entering a wave of personal OSes and custom internal toolsHow to avoid shipping janky AI UI and make your app look intentionally designedThe mindset shift from “I could build anything” → “I will build this one specific thing”Why specializing in one AI tool (Lovable, Cursor, n8n, etc.) beats being “the AI guy”Tactical content and lead-gen plays for agencies on LinkedIn and YouTubeHow to learn AI tooling without getting paralyzed by the infinite possibilitiesTimestamps00:00 — Vibe coding: bubble or breakthrough?02:23 — Effective use of no-code tools05:23 — Stack and scoping for MVP development07:08 — Trends in personal software development10:33 — Personal projects: blood work analysis tool13:00 — Steps to start building custom software17:49 — Successful and unsuccessful product categories21:01 — Learning and adopting AI tools27:45 — Creator collaboration in software development32:14 — Lead generation strategies for AI-powered agenciesKey Topics & Ideas1. Bubble or Skill Issue?Why early no-code/AI apps looked jankyHow tools like Lovable increased automation from ~50% → ~85%The remaining 10–15% where real engineering still mattersMany failures come from non-devs skipping fundamentals2. How Creme Builds Lovable MVPsEvery project starts with a clear PRD (often drafted with ChatGPT)AI is used to tighten scope before buildingWhen Creme stays fully in Lovable vs. moving code to CursorUsing Lovable Cloud for hosting, database, and analytics3. Personal Operating Systems & Internal ToolsPeople replacing SaaS subscriptions with their own custom toolsIn a 20-person cohort, nearly everyone built workflow appsRise of the Personal OS: one system for life + workExample builds:Bloodwork tracker from PDF uploadsUnified messaging CRM (WhatsApp, Telegram, SMS, email)Automated 30-second sales briefings4. How to Learn AI Coding ToolsHalf the cohort hadn't built anything before startingMain blocker: overwhelm, not skillLearn core concepts: frontend vs. backend, auth, roles, securityBuild daily reps, focus on the next thing you need—not “all of AI”5. Designing Apps That Don't Look AI-GeneratedGood design is still the hardest and biggest edgeCreme process: build a /components library, define buttons/cards/inputs, assign stable IDsTools: Mobbin, Figma Community kits, 21st.devBest prompt: “Here's a screenshot → copy this.”6. What Works in Product IdeasMost of Creme's builds are full startup platforms, not micro-toolsAI makes shipping easier, but ideas are getting worse without depthReal advantage = domain expertise + niche problem + AI speed7. Creators x SoftwareCreators can now ship products without capitalJacob prefers retainers over equityAnalogy: Like creator brands—most fail, a few go huge8. Career Strategy: SpecializeFuture = verticalized expertise, not “AI generalists”Specialist lanes: Lovable, Cursor, n8n, automationBe the person for one tool + one market9. Content & Lead GenJacob's two rules for content: people are selfish and people are boredBuild content that teaches, sparks emotion, and creates curiosityPost ~5x/week, prioritize visual postsLong-term: YouTube deep dives for high-intent inboundSponsorToday's episode is brought to you by Graphed – an AI data analyst & BI platform.With Graphed you can:Connect data like GA4, Facebook Ads, HubSpot, Google Ads, Search Console, AmplitudeBuild interactive dashboards just by chatting (no Looker Studio/Tableau learning curve)Use it as your ETL + data warehouse + BI layer in one placeAsk:“Build me a stacked bar chart of new users vs. all users over time from GA4”…and Graphed just builds it for you.

Andy Pernsteiner is the Field CTO at VAST Data, working on large-scale AI infrastructure, serverless compute near data, and the rollout of VAST's AI Operating System.The GPU Uptime Battle // MLOps Podcast #346 with Andy Pernsteiner, Field CTO of VAST Data.Huge thanks to VAST Data for supporting this episode!Join the Community: https://go.mlops.community/YTJoinInGet the newsletter: https://go.mlops.community/YTNewsletter// AbstractMost AI projects don't fail because of bad models; they fail because of bad data plumbing. Andy Pernsteiner joins the podcast to talk about what it actually takes to build production-grade AI systems that aren't held together by brittle ETL scripts and data copies. He unpacks why unifying data - rather than moving it - is key to real-time, secure inference, and how event-driven, Kubernetes-native pipelines are reshaping the way developers build AI applications. It's a conversation about cutting out the complexity, keeping data live, and building systems smart enough to keep up with your models. // BioAndy is the Field Chief Technology Officer at VAST, helping customers build, deploy, and scale some of the world's largest and most demanding computing environments.Andy has spent the past 15 years focused on supporting and building large-scale, high-performance data platform solutions. From humble beginnings as an escalations engineer at pre-IPO Isilon, to leading a team of technical Ninjas at MapR, he's consistently been in the frontlines solving some of the toughest challenges that customers face when implementing Big Data Analytics and next-generation AI solutions.// Related LinksWebsite: www.vastdata.comhttps://www.youtube.com/watch?v=HYIEgFyHaxkhttps://www.youtube.com/watch?v=RyDHIMniLro The Mom Test by Rob Fitzpatrick: https://www.momtestbook.com/~~~~~~~~ ✌️Connect With Us ✌️ ~~~~~~~Catch all episodes, blogs, newsletters, and more: https://go.mlops.community/TYExploreJoin our Slack community [https://go.mlops.community/slack]Follow us on X/Twitter [@mlopscommunity](https://x.com/mlopscommunity) or [LinkedIn](https://go.mlops.community/linkedin)] Sign up for the next meetup: [https://go.mlops.community/register]MLOps Swag/Merch: [https://shop.mlops.community/]Connect with Demetrios on LinkedIn: /dpbrinkmConnect with Andy on LinkedIn: /andypernsteinerTimestamps:[00:00] Prototype to production gap[00:21] AI expectations vs reality[03:00] Prototype vs production costs[07:47] Technical debt awareness[10:13] The Mom Test[15:40] Chaos engineering[22:25] Data messiness reflection[26:50] Small data value[30:53] Platform engineer mindset shift[34:26] Gradient description comparison[38:12] Empathy in MLOps[45:48] Empathy in Engineering[51:04] GPU clusters rolling updates[1:03:14] Checkpointing strategy comparison[1:09:44] Predictive vs Generative AI[1:17:51] On Growth, Community, and New Directions[1:24:21] UX of agents[1:32:05] Wrap up

In this episode of Crazy Wisdom, host Stewart Alsop talks with Jessica Talisman, founder of Contextually and creator of the Ontology Pipeline, about the deep connections between knowledge management, library science, and the emerging world of AI systems. Together they explore how controlled vocabularies, ontologies, and metadata shape meaning for both humans and machines, why librarianship has lessons for modern tech, and how cultural context influences what we call “knowledge.” Jessica also discusses the rise of AI librarians, the problem of “AI slop,” and the need for collaborative, human-centered knowledge ecosystems. You can learn more about her work at Ontology Pipeline and find her writing and talks on LinkedIn.Check out this GPT we trained on the conversationTimestamps00:00 Stewart Alsop welcomes Jessica Talisman to discuss Contextually, ontologies, and how controlled vocabularies ground scalable systems.05:00 They compare philosophy's ontology with information science, linking meaning, categorization, and sense-making for humans and machines.10:00 Jessica explains why SQL and Postgres can't capture knowledge complexity and how neuro-symbolic systems add context and interoperability.15:00 The talk turns to library science's split from big data in the 1990s, metadata schemas, and the FAIR principles of findability and reuse.20:00 They discuss neutrality, bias in corporate vocabularies, and why “touching grass” matters for reconciling internal and external meanings.25:00 Conversation shifts to interpretability, cultural context, and how Western categorical thinking differs from China's contextual knowledge.30:00 Jessica introduces process knowledge, documentation habits, and the danger of outsourcing how-to understanding.35:00 They explore knowledge as habit, the tension between break-things culture and library design thinking, and early AI experiments.40:00 Libraries' strategic use of AI, metadata precision, and the emerging role of AI librarians take focus.45:00 Stewart connects data labeling, Surge AI, and the economics of good data with Jessica's call for better knowledge architectures.50:00 They unpack content lifecycle, provenance, and user context as the backbone of knowledge ecosystems.55:00 The talk closes on automation limits, human-in-the-loop design, and Jessica's vision for collaborative consulting through Contextually.Key InsightsOntology is about meaning, not just data structure. Jessica Talisman reframes ontology from a philosophical abstraction into a practical tool for knowledge management—defining how things relate and what they mean within systems. She explains that without clear categories and shared definitions, organizations can't scale or communicate effectively, either with people or with machines.Controlled vocabularies are the foundation of AI literacy. Jessica emphasizes that building a controlled vocabulary is the simplest and most powerful way to disambiguate meaning for AI. Machines, like people, need context to interpret language, and consistent terminology prevents the “hallucinations” that occur when systems lack semantic grounding.Library science predicted today's knowledge crisis. Stewart and Jessica trace how, in the 1990s, tech went down the path of “big data” while librarians quietly built systems of metadata, ontologies, and standards like schema.org. Today's AI challenges—interoperability, reliability, and information overload—mirror problems library science has been solving for decades.Knowledge is culturally shaped. Drawing from Patrick Lambe's work, Jessica notes that Western knowledge systems are category-driven, while Chinese systems emphasize context. This cultural distinction explains why global AI models often miss nuance or moral voice when trained on limited datasets.Process knowledge is disappearing. The West has outsourced its “how-to” knowledge—what Jessica calls process knowledge—to other countries. Without documentation habits, we risk losing the embodied know-how that underpins manufacturing, engineering, and even creative work.Automation cannot replace critical thinking. Jessica warns against treating AI as “room service.” Automation can support, but not substitute, human judgment. Her own experience with a contract error generated by an AI tool underscores the importance of review, reflection, and accountability in human–machine collaboration.Collaborative consulting builds knowledge resilience. Through her consultancy, Contextually, Jessica advocates for “teaching through doing”—helping teams build their own ontologies and vocabularies rather than outsourcing them. Sustainable knowledge systems, she argues, depend on shared understanding, not just good technology.

Marketing analytics stacks struggle with outdated, siloed data that delays critical business decisions. Noha Rizk, CMO of Incorta, explains how live data integration transforms enterprise analytics capabilities. She demonstrates how questioning "why" behind data patterns unlocks actionable insights and discusses eliminating complex ETL processes through real-time analysis across all business systems. The conversation covers practical frameworks for moving from raw data collection to immediate business intelligence that drives customer behavior understanding.See Privacy Policy at https://art19.com/privacy and California Privacy Notice at https://art19.com/privacy#do-not-sell-my-info.

Marketing analytics stacks struggle with outdated, siloed data that delays critical business decisions. Noha Rizk, CMO of Incorta, explains how live data integration transforms enterprise analytics capabilities. She demonstrates how questioning "why" behind data patterns unlocks actionable insights and discusses eliminating complex ETL processes through real-time analysis across all business systems. The conversation covers practical frameworks for moving from raw data collection to immediate business intelligence that drives customer behavior understanding.See Privacy Policy at https://art19.com/privacy and California Privacy Notice at https://art19.com/privacy#do-not-sell-my-info.

Incorta is the first and only open data delivery platform that enables real-time analysis of live, detailed data across all systems of record—without the need for complex ETL processes.See Privacy Policy at https://art19.com/privacy and California Privacy Notice at https://art19.com/privacy#do-not-sell-my-info.

Incorta is the first and only open data delivery platform that enables real-time analysis of live, detailed data across all systems of record—without the need for complex ETL processes.See Privacy Policy at https://art19.com/privacy and California Privacy Notice at https://art19.com/privacy#do-not-sell-my-info.

Incorta is the first and only open data delivery platform that enables real-time analysis of live, detailed data across all systems of record—without the need for complex ETL processes.See Privacy Policy at https://art19.com/privacy and California Privacy Notice at https://art19.com/privacy#do-not-sell-my-info.

Most companies rely on stale dashboards while AI demands live data for real-time decisions. Noha Rizk, CMO of Incorta, explains how enterprises can transition from legacy data systems to real-time analytics infrastructure. She covers identifying high-ROI use cases like retail waste optimization and supply chain management, implementing live data without complex ETL processes, and enabling business users to query data instantly for creative problem-solving.See Privacy Policy at https://art19.com/privacy and California Privacy Notice at https://art19.com/privacy#do-not-sell-my-info.

In this episode, I interview Josh Cook - the founder of ETL - Enter the Lion. Josh joined the Baltimore Police Department as a police officer, eventually transferring to the Anne Arundel County Police Department. Feeling called to minister to law enforcement directly, he separated from his department, relocated to Tennessee and began Enter the Lion  (ETL).  ETL is a Christian ministry that provides a completely free retreat to first responders and law enforcement who are desiring rest and time in nature. ETL provides biblical counseling, mentorship, and discipleship to those seeking a connection with other believers. To contact Josh or inquire about attending a retreat, contact him through his website:www.enterthelion.coYou can access The Tactical Debrief on Apple, Spotify, or Audible Podcasts.

Points of Interest00:00 – 01:30 – Introduction: Marcel welcomes Parakeeto's Kristen Kelly back to discuss a recurring misconception in agency operations—the belief that a better project management or PSA tool can solve profit management challenges.01:30 – 03:25 – The PM Tool “Silver Bullet” Myth: Kristen explains how leaders and PMs often adopt new tools to tame chaos, believing marketing promises that they'll also solve utilization, capacity, and profitability issues.03:25 – 06:00 – Why Agencies Fall for It: Marcel and Kristen note that while PM tools are valuable, they're often oversold as full profit-management systems. Agencies end up frustrated by missing fields, tool quirks, and data limitations.06:00 – 08:45 – Hitting the Wall: Many teams find themselves with tools that improve delivery workflows but still leave them unable to make key financial or operational decisions because the data remains fragmented across systems.08:45 – 11:43 – Introducing the Framework → Data → Process Model: Marcel outlines Parakeeto's three-part sequence for solving profit management: define the framework (metrics and formulas), structure the data, and establish ongoing processes for hygiene and cadence.11:43 – 12:46 – Why Sequencing Matters: Without first defining what needs to be measured, agencies make poor configuration choices in PM tools—creating rework, confusion, and endless tool migrations.12:46 – 15:19 – Defining the Framework: Agencies must precisely define how metrics like utilization, delivery margin, and project profitability are calculated, and understand the relationships between those measures before configuring tools.15:19 – 19:54 – The Role of Process and Data Hygiene: Marcel explains that real-time reporting fails if data quality is poor. Clean, reliable reporting requires an ETL (Extract, Transform, Load) process, not direct reporting from source data.19:54 – 22:55 – The Precision Trap: Kristen and Marcel explore the conflict between PMs needing granular precision and executives needing simple, high-level rollups. Forcing perfect data consistency across teams destroys usability and compliance.22:55 – 26:28 – Practical Limits of In-Tool Reporting: Marcel describes how building detailed profitability reporting directly in PM tools creates unsustainable complexity, unrealistic data maintenance, and unreliable results.26:28 – 34:38 – Building a Sustainable Data Architecture: They outline how Parakeeto's ETL pipeline works—extracting time data (person, project, hours), joining it with payroll and project grids, normalizing fields, and applying ongoing QA to ensure accuracy.34:38 – 42:37 – The Big Takeaway: Kristen and Marcel conclude that PM tools are essential for delivery but not the whole profit solution. Agencies should use them for managing work while relying on a clear framework and data pipeline for accurate reporting.Show NotesConnect with Kristen via LinkedInFree Agency ToolkitParakeeto Foundations CourseFree access to our Model PlatformLove this Episode?Leave us a review here. Hosted by Simplecast, an AdsWizz company. See pcm.adswizz.com for information about our collection and use of personal data for advertising.

Roger Baudet - "Compliainte a Deux" - Musique Électronique Pour La Scène Et L'image 1976 - 1992 https://www.wfmu.org/playlists/shows/156794

MMPs give you a strong foundation for measuring mobile campaigns. But what if that's not enough? What if the data you're missing could unlock faster growth, smarter user acquisition, and better ROI?That's where Extract comes in. In this episode of Growth Masterminds, John Koetsier talks with Maayan Schor about why mobile marketers need a next-gen ELT and reverse ETL platform to move raw data in and out of their systems. From app stores to social to ad networks, Extract helps you pull it all together with your MMP data ... and make smarter decisions with more confidence.Leading, of course, to more growth.We cover:- Why there's more data than MMPs provide alone- How to access raw app store, organic social, and granular ad network data- Real-world use cases from top mobile marketers- How BI teams and marketers collaborate to make Extract work- Why flexibility and context are key to growth in mobile marketingIf you're working in mobile marketing, user acquisition, data engineering, or growth analytics, this conversation is packed with insights you can use today.

Would you like access to our advanced agency training for FREE? https://www.agencymastery360.com/training Are you still thinking of AI as just “ChatGPT with a better prompt”? Or maybe you've played around with Zapier automations and thought, yeah, that's good enough. Today's featured guest knows that the agencies pulling ahead right now are building full-on AI agent networks that replace routine tasks, streamline data pipelines, and give their teams superpowers. She's re-engineering her agency around AI and will talk about where she finds top-tier talent and why you don't need to code to lead your agency into the future. Jennifer Bagley is the CEO and founder of CI Web Group, a fully virtual digital marketing agency registered in 22 U.S. states with clients across the United States and Canada. A former corporate operator turned entrepreneur, Jennifer started in real estate and mortgage brokerage before leaning into the marketing work she built to support those businesses. Today she runs a modern, tech-forward agency that's rebuilt its stack around AI, centralized data, and agentic networks, all while carrying the scars and lessons of scaling, pivoting, and re-founding a business from the ground up. In this episode, we'll discuss: Feeling trapped by the business. Hiring, firing, and the people reset AI, reskilling, and the end of “middle” roles What does this talent cost? Subscribe Apple | Spotify | iHeart Radio Sponsors and Resources E2M Solutions: Today's episode of the Smart Agency Masterclass is sponsored by E2M Solutions, a web design, and development agency that has provided white-label services for the past 10 years to agencies all over the world. Check out e2msolutions.com/smartagency and get 10% off for the first three months of service. From Corporate Ladder to Accidental Agency Founder Jennifer came from an operations background, a self-proclaimed black belt in Six Sigma and certified project manager. Having built that corporate background, she had made a promise to herself (“by 30 I'll be an entrepreneur”), and started to build the side hustle that became the main event. She started in real estate and mortgage brokering where she had to learn marketing the hard way; not because she wanted to be a marketer, but because the survival of her businesses depended on it. Initially, Jennifer didn't set out to build a scalable agency; she built a team to support her broker network. When the market collapsed in 2008, the same team that did marketing for agents suddenly had a market outside real estate. That “we'll just help this painter or HVAC company” phase is where the web group was born: small, service-focused, and useful to people in her network. That accidental turn became a business by solving real, pressing problems for paying clients, then leaned into that. Trading Time for Freedom: The Hard Pivot For the first five years, Jennifer describes the business as a “lifestyle” operation, profitable maybe, but trapping her time. She was trading billable hours for income and was reaching her limit when she hired a coach that forced a reckoning: if entrepreneurship isn't buying you time, money, and freedom, what's the point? So she made the brutal choice of cutting consulting contracts and burning the bridge to the “safety” of hourly work, and effectively gave herself a mulligan. This is the classic founder pivot: you have to choose between growth that keeps you doing the work and growth that scales the business without you. Jennifer's reset wasn't pretty, for a while she lost everything and she and her son lived in an office for a while, but it bought her the permission to build something salable, not just sustainable. Agency owners who feel trapped in delivery need to remember that sometimes you have to give up short-term revenue to create long-term value. Feeling Trapped by the Agency and Becoming a CEO Those first five years, Jennifer continued to run a business that started as a supply chain consulting and eventually turned into a sales supply chain consulting. This change meant the business was now a good lead generator for the agency but it also meant Jennifer was essentially selling her image and her time. Until she ran out of time. Once she felt trapped by the business, Jennifer actually hired a business coach that helped her change the model from “selling Jennifer with marketing on the side” to an actual sustainable business. She had to go back to the basics and remember she, like every entrepreneur, started the business with the idea of having more time, money, and freedom. It took losing everything, but Jennifer knew she didn't want a lifestyle business, she wanted a sellable business. The antidote was delegation plus systems. If you want growth and a future exit, you need to own those CEO responsibilities and be comfortable with letting go of the day-to-day. Hiring, Firing, and Resetting the Team Jennifer's talent strategy has evolved with each stage of growth. Her early hires were the classic “friends, family, fools” bootstrap crew; later she invested in developers, content teams, project managers, and over time, more strategic hires like CFOs, chief of staff, BI teams, and AI engineers. Each five-year arc brought a new set of needs and a new level of sophistication in hiring. Now, she divides her time between promoting her agency's work in podcasts and content and thinking of ways to navigate her business in these volatile and exciting times. Her most recent addition to the team was a technology and transformation team that is revisiting all of the agency's processes, investments, and infrastructure. As a result, she has downsized her team from over 300 W2 employees and refocus the team. The takeaway for agency owners: be honest about whether your people are builders or maintainers, and hire accordingly. The workforce you need for growth is not the same as the workforce you need for stable operations. Building AI Agent Networks with Centralized Data Jennifer's agency shifted from WordPress to Webflow and built agentic networks: hundreds of AI agents that crawl competitors, do strategy homework, and automate tasks that humans used to do. More importantly, they rebuilt infrastructure into a hub-and-spoke model with a centralized min.io data layer and ETL pipelines feeding analytics and BI. Two big lessons here. One: invest in your tech stack deliberately so you're not a Frankenstein of five different platforms that don't talk to each other. Two: design your data architecture so your people (and your AI agents) have a single source of truth. That's how you get from fire-fighting in six dashboards to proactive, predictive signals that tell you when a client engagement needs attention. AI, Reskilling, and Shrinking Middle Roles Jennifer draws a hard line: the agency now tends to hire either very seasoned client-facing leaders or AI engineers; the middle is shrinking. With agentic networks giving junior staff “superpowers,” the agency can afford fewer mid-level “lever pullers.” At this level there's no room for slow execution or elementary work. That's a cultural and ethical challenge, both for hiring and for workforce development. For agency owners, this raises practical HR questions: do you reskill your people, or replace them? Jennifer suggests building agent-driven systems that augment humans, and being brutally honest about who can grow into that future. It's also a call to action for how we prepare the next generation: schools won't teach this; companies will need to. Playing with AI Platforms: Why Leaders Need to Just Know Enough to Be Dangerous Jennifer started like a lot of agency owners dipping into AI, playing around on tools like n8n, Make.com, Relevance, and Longchain. Her dev team laughed, calling her an “elementary school kid on a tricycle,” but here's the point: she didn't need to master the tech. She needed to know enough to point her team in the right direction. Instead of obsessing over code, she framed the problem differently: “Here's what I don't want a human doing anymore. Can you make that happen?” That mindset shift is key for agency owners. You don't need to be a full-stack AI engineer to lead an agency into the future; you just need to clearly define outcomes and invest in people who can deliver them. Find Real AI Talent in Unlikely Places This is where most agencies get stuck. You're not going to find your next AI architect on Upwork. Jennifer leaned on her network, starting with her cousin Chris, a hardcore developer who initially thought AI platforms were “rookie business.” Once Chris realized the power of agentic networks to scale his expertise, he became the backbone of CI Web Group's transformation. Now, she hunts talent in unconventional places: hackathons, LinkedIn, and especially YouTube. Forget the flashy “10x growth hack” videos — she looks for nerds with four views, geeking out about orchestrators and ETL pipelines. Those are the builders who care about solving real problems, not just building hype. Her tip: if you find one, reach out immediately. They don't want sales, they just want to build. Designing AI Agents Like an Agency Org Chart Jennifer compares AI agents to a company org chart. You don't hire one person to do everything, that's a recipe for burnout. Same thing with AI. Each agent should tightly focus on a single task, with checks, auditors, and orchestrators overseeing the system. The payoff was massive efficiency gains. Instead of six different platforms that don't talk, her agency built a centralized hub with min.io, ClickHouse, and AI layers on top. That's how you go from patchwork automation to true predictive intelligence. The Real Cost of AI Talent If you're wondering how much this all costs, the answer is… a lot. On the high end, seasoned AI engineers can run you a quarter million in salary. On the low end, Jennifer tests new hires on project-based sprints, maybe $6K for a 10-hour challenge. The point isn't to cut costs; it's to prove quickly who can deliver and who can't. Her recruiting process is brutal but effective: give candidates a project, a tight deadline, and see how they perform. If they stall, they're out. If they screen-share fast and solve problems live, they're in. No fluff, no endless interviews. Do You Want to Transform Your Agency from a Liability to an Asset? Looking to dig deeper into your agency's potential? Check out our Agency Blueprint. Designed for agency owners like you, our Agency Blueprint helps you uncover growth opportunities, tackle obstacles, and craft a customized blueprint for your agency's success.

Send us a textWhat if AI could tap into live operational data — without ETL or RAG? In this episode, Deepti Srivastava, founder of Snow Leopard, reveals how her company is transforming enterprise data access with intelligent data retrieval, semantic intelligence, and a governance-first approach. Tune in for a fresh perspective on the future of AI and the startup journey behind it.We explore how companies are revolutionizing their data access and AI strategies. Deepti Srivastava, founder of Snow Leopard, shares her insights on bridging the gap between live operational data and generative AI — and how it's changing the game for enterprises worldwide.We dive into Snow Leopard's innovative approach to data retrieval, semantic intelligence, and governance-first architecture.04:54 Meeting Deepti Srivastava 14:06 AI with No ETL, no RAG 17:11 Snow Leopard's Intelligent Data Fetching 19:00 Live Query Challenges 21:01 Snow Leopard's Secret Sauce 22:14 Latency 23:48 Schema Changes 25:02 Use Cases 26:06 Snow Leopard's Roadmap 29:16 Getting Started 33:30 The Startup Journey 34:12 A Woman in Technology 36:03 The Contrarian View

Send us a textWhat if AI could tap into live operational data — without ETL or RAG? In this episode, Deepti Srivastava, founder of Snow Leopard, reveals how her company is transforming enterprise data access with intelligent data retrieval, semantic intelligence, and a governance-first approach. Tune in for a fresh perspective on the future of AI and the startup journey behind it.We explore how companies are revolutionizing their data access and AI strategies. Deepti Srivastava, founder of Snow Leopard, shares her insights on bridging the gap between live operational data and generative AI — and how it's changing the game for enterprises worldwide.We dive into Snow Leopard's innovative approach to data retrieval, semantic intelligence, and governance-first architecture.04:54 Meeting Deepti Srivastava 14:06 AI with No ETL, no RAG 17:11 Snow Leopard's Intelligent Data Fetching 19:00 Live Query Challenges 21:01 Snow Leopard's Secret Sauce 22:14 Latency 23:48 Schema Changes 25:02 Use Cases 26:06 Snow Leopard's Roadmap 29:16 Getting Started 33:30 The Startup Journey 34:12 A Woman in Technology 36:03 The Contrarian View

I invited Atalia Horenshtien to unpack a topic many leaders are wrestling with right now. Everyone is talking about AI agents, yet most teams are still living with rule based bots, brittle scripts, and a fair bit of anxiety about handing decisions to software. Atalia has lived through the full arc, from early machine learning and automated pipelines to today's agent frameworks inside large enterprises. She is an AI and data strategist, a former data scientist and software engineer, and has just joined Hakoda, an IBM company, to help global brands move from experiments to outcomes. The timing matters. She starts on the 18th, and this conversation captures how she thinks about responsible progress at exactly the moment she steps into that new role. Here's the thing. Words like autonomy sound glamorous until an agent faces a messy real world task. Atalia draws a clear line between scripted bots and agents with goals, memory, and the ability to learn from feedback. Her advice is refreshingly grounded. Start internal where you can observe behavior. Put human in the loop review where it counts. Use role based access rather than feeding an LLM everything you own. Build an observability layer so you can see what the model did, why it did it, and what it cost. We also get into measurements that matter. Time saved, cycle time reduction, adoption, before and after comparisons, and a sober look at LLM costs against any reduction in FTE hours. She shares how custom cost tracking for agents prevents surprises, and why version one should ship even if it is imperfect. Culture shows up as a recurring theme. Leaders need to talk openly about reskilling, coach managers through change, and invite teams to be co creators. Her story about Hakoda's internal AI Lab is a good example. What began as an engineer's idea for ETL schema matching grew into agent powered tools that won a CIO 100 award and now help deliver faster, better outcomes for clients. There are lighter moments too. Atalia explains how she taught an ex NFL player the basics of time series forecasting using football tactics. Then she takes us behind the scenes with McLaren Racing, where data and strategy collide on the F1 circuit, and admits she has become a committed fan because of that work. If you want a practical playbook for moving from shiny demos to dependable agents, this episode will help you think clearly about scope, safeguards, and speed. Connect with Atalia on LinkedIn, explore Hakoda's work at hakoda.io, and then tell me how you plan to measure your first agent's value. ********* Visit the Sponsor of Tech Talks Network: Land your first job  in tech in 6 months as a Software QA Engineering Bootcamp with Careerist https://crst.co/OGCLA  

During the IT Press Tour, I had the pleasure of speaking with Weimo Liu, CEO and co-founder of PuppyGraph, and hearing firsthand how his team is rethinking graph technology for the enterprise. In this episode of Tech Talks Daily, Weimo joins me to share the story behind PuppyGraph's “zero ETL” approach, which lets organizations query their existing data as a graph without ever moving or duplicating it. We discuss why graph databases, despite their promise, have struggled with mainstream adoption, often because of complex pipelines and heavy infrastructure requirements. Weimo explains how PuppyGraph borrows from his time at TigerGraph and Google's F1 engine to build something new: a distributed query engine that maps tables into a logical graph and delivers subsecond performance on massive datasets. That shift opens the door for use cases in cybersecurity, fraud detection, and AI-driven applications where latency and accuracy matter most. We also unpack the developer experience. Instead of rewriting schemas or reloading data every time requirements change, PuppyGraph allows teams to define nodes and edges directly from existing tables. That design lowers the barrier for SQL-focused teams and accelerates time to value. Weimo even touches on the role of graph in reducing AI hallucinations, showing how structured relationships can make enterprise AI systems more reliable. What struck me most in our conversation is how PuppyGraph's playful branding belies its serious engineering depth. Behind the “puppy” name lies a distributed engine built to scale with today's data volumes, backed by strong early adoption and a team that listens closely to customer needs. Whether you're exploring graph for cybersecurity, AI chatbots, or supply chain analytics, this discussion offers a glimpse of how the next generation of graph tech might finally break free from its niche and go mainstream. ********* Visit the Sponsor of Tech Talks Network: Land your first job  in tech in 6 months as a Software QA Engineering Bootcamp with Careerist https://crst.co/OGCLA

What does it really mean to be data-driven? Mark Gergess, VP of Data and BI at DoubleVerify, joins the show to unpack how data teams can go beyond dashboards to drive meaningful business action. From building an internal consulting lens to evaluating the latest AI tools, Mark shares how his team translates complex data flows into measurable revenue impact. If you've ever wrestled with the gap between insights and outcomes, this conversation will hit home.Key Takeaways• Being data-driven is about driving action, not just reporting numbers• Stakeholders don't care about your data problems—they care about business outcomes• The biggest challenge with AI adoption isn't the model, it's the use cases• Efficiency gains from AI should shift focus from ETL tasks to solving real business problems• Data culture health is measured by how naturally teams rely on data day-to-dayTimestamped Highlights01:17 How DoubleVerify helps advertisers build safer, more effective digital campaigns04:55 Why the definition of “data-driven” still varies and why it matters09:25 Measuring whether data efforts are moving the needle on revenue13:15 How to separate hype from value when evaluating AI and GenAI tools17:10 Lessons from the data science boom and why companies must go “all in” with AI25:31 Can AI act as your junior analyst? Where efficiency gains really show up27:01 How freeing up time changes the structure of data teams and boosts business impactA thought worth holding onto“It's not about dashboards. It's not about reporting. It's about doing something with the information.”Pro TipsMark recommends treating AI as a “junior analyst”—let it handle quick, lower-priority questions so your team can focus on bigger business challenges.Call to ActionEnjoyed the conversation? Share this episode with a colleague who talks about being “data-driven.” Subscribe on your favorite podcast platform and connect with me on LinkedIn for more insights from leaders shaping the future of data and technology.

In this episode, Dave and Jamison answer these questions: I'm the CTO of a small startup. We're 3 devs including me and one of them is a junior developer. My current policy is to discourage the use of AI tools for the junior dev to make sure they build actual skills and don't just prompt their way through tasks. However I'm more and more questioning my stance as AI skills will be in demand for jobs to come and I want to prepare this junior dev for a life after my startup. How would you do this? What's the AI coding assistant policy in your companies. Is it the same for all seniority levels? Hi everyone! Long-time listener here, and I really appreciate all the insights you share. Greetings from Brazil! I recently joined a large company (5,000 employees) that hired around 500 developers in a short time. It seems like they didn't have enough projects aligned with everyone's expertise, so many of us, myself included, were placed in roles that don't match our skill sets. I'm a web developer with experience in Java and TypeScript, but I was assigned to a data-focused project involving Python and ETL pipelines, which is far from my area of interest or strength. I've already mentioned to my manager that I don't have experience in this stack, but the response was that the priority is to place people in projects. He told me to “keep [him] in the loop if you don't feel comfortable”, but I'm not sure that should I do. The company culture is chill, and I don't want to come across as unwilling to work or ungrateful. But I also want to grow in the right direction for my career. How can I ask for a project change, ideally one that aligns with my web development background, without sounding negative or uncooperative? Maybe wait for like 3 months inside of this project and then ask for a change? Thanks so much for your thoughts!

Joey DeVilla of Tampa Tech fame and accordion playing glory joins Mike to discuss the Tampa Tech scene, some Python goodness, a little Rust and much more. Try Mailtrap for free (https://l.rw.rw/coder_radio_6) Joey's Blog (https://www.joeydevilla.com/) Mike on X (https://x.com/dominucco) Mike on BlueSky (https://bsky.app/profile/dominucco.bsky.social) Coder on X (https://x.com/coderradioshow) Coder on BlueSky (https://bsky.app/profile/coderradio.bsky.social) Show Discord (https://discord.gg/k8e7gKUpEp) Alice (https://alice.dev)

Unlock the power of Alteryx for tax professionals in this insightful episode of Alter Everything! Join us in an interview with Adrian Steller, Director of Tax Technology at Ryan, to explore how Alteryx revolutionizes tax processes, automates data workflows, and enhances efficiency for tax teams. Discover real-world Alteryx use cases in VAT compliance, transfer pricing, and automation, and learn practical tips for transitioning from Excel to Alteryx. Whether you're a tax analyst, data professional, or business leader, this episode provides actionable insights on leveraging Alteryx for tax data transformation, reporting, and analytics.Panelists: Adrian Steller, Director @ International Tax Technology - LinkedInMegan Bowers, Sr. Content Manager @ Alteryx - @MeganBowers, LinkedInShow notes: Ryan (Company)Ryan Tax Lab (Podcast)Alteryx Community BlogsAlteryx Help Docs Interested in sharing your feedback with the Alter Everything team? Take our feedback survey here!This episode was produced by Megan Bowers, Mike Cusic, and Matt Rotundo. Special thanks to Andy Uttley for the theme music.