Podcasts about streamlit

Try Keboola

Can a combo of Large Language Models (LLMs) and Domain-Specific Languages (DSLs) streamline development by automating repetitive patterns across teams? In this Mob Mentality Show episode, we dive deep into the intersection of AI-driven automation, code generation, and lean software development. Join us as we explore: ✅ The "Generator for the Generator" Concept – How AI-powered tools and Mob Programming can create DSLs that automate code generation at scale. ✅ Handling Cross-Domain Development with DSLs – How DSL arguments can be leveraged to generate applications across multiple domains while maintaining usability. ✅ Serverless Infrastructure as Code (IaC) & Auto-Generated Apps – How to use DSLs to automate cloud deployment with Angular or Streamlit apps. ✅ The Challenge of UI/UX in Generated Code – When UI is too generic, does it hurt usability? Can a DSL strike the right balance between automation and user experience? ✅ Regeneration vs. Continuous Development – Should teams work exclusively in the DSL, or also refine the code it generates? How to handle sync issues when regenerating applications. ✅ Turning Docs into Code with a DSL Converter – Automating workflows by transforming team documentation into executable code. ✅ Mob Automationist Role Inception – Is the next evolution of Mob Programming automating the automation? ✅ ZOMBIE Test Generation & Nested Python Dictionaries – How automated testing fits into the DSL-driven workflow and whether a DSL can be as simple as a structured Python dictionary.

Applications for the NYC AI Engineer Summit, focused on Agents at Work, are open!When we first started Latent Space, in the lightning round we'd always ask guests: “What's your favorite AI product?”. The majority would say Midjourney. The simple UI of prompt → very aesthetic image turned it into a $300M+ ARR bootstrapped business as it rode the first wave of AI image generation.In open source land, StableDiffusion was congregating around AUTOMATIC1111 as the de-facto web UI. Unlike Midjourney, which offered some flags but was mostly prompt-driven, A1111 let users play with a lot more parameters, supported additional modalities like img2img, and allowed users to load in custom models. If you're interested in some of the SD history, you can look at our episodes with Lexica, Replicate, and Playground.One of the people involved with that community was comfyanonymous, who was also part of the Stability team in 2023, decided to build an alternative called ComfyUI, now one of the fastest growing open source projects in generative images, and is now the preferred partner for folks like Black Forest Labs's Flux Tools on Day 1. The idea behind it was simple: “Everyone is trying to make easy to use interfaces. Let me try to make a powerful interface that's not easy to use.”Unlike its predecessors, ComfyUI does not have an input text box. Everything is based around the idea of a node: there's a text input node, a CLIP node, a checkpoint loader node, a KSampler node, a VAE node, etc. While daunting for simple image generation, the tool is amazing for more complex workflows since you can break down every step of the process, and then chain many of them together rather than manually switching between tools. You can also re-start execution halfway instead of from the beginning, which can save a lot of time when using larger models.To give you an idea of some of the new use cases that this type of UI enables:* Sketch something → Generate an image with SD from sketch → feed it into SD Video to animate* Generate an image of an object → Turn into a 3D asset → Feed into interactive experiences* Input audio → Generate audio-reactive videosTheir Examples page also includes some of the more common use cases like AnimateDiff, etc. They recently launched the Comfy Registry, an online library of different nodes that users can pull from rather than having to build everything from scratch. The project has >60,000 Github stars, and as the community grows, some of the projects that people build have gotten quite complex:The most interesting thing about Comfy is that it's not a UI, it's a runtime. You can build full applications on top of image models simply by using Comfy. You can expose Comfy workflows as an endpoint and chain them together just like you chain a single node. We're seeing the rise of AI Engineering applied to art.Major Tom's ComfyUI Resources from the Latent Space DiscordMajor shoutouts to Major Tom on the LS Discord who is a image generation expert, who offered these pointers:* “best thing about comfy is the fact it supports almost immediately every new thing that comes out - unlike A1111 or forge, which still don't support flux cnet for instance. It will be perfect tool when conflicting nodes will be resolved”* AP Workflows from Alessandro Perili are a nice example of an all-in-one train-evaluate-generate system built atop Comfy* ComfyUI YouTubers to learn from:* @sebastiankamph* @NerdyRodent* @OlivioSarikas* @sedetweiler* @pixaroma* ComfyUI Nodes to check out:* https://github.com/kijai/ComfyUI-IC-Light* https://github.com/MrForExample/ComfyUI-3D-Pack* https://github.com/PowerHouseMan/ComfyUI-AdvancedLivePortrait* https://github.com/pydn/ComfyUI-to-Python-Extension* https://github.com/THtianhao/ComfyUI-Portrait-Maker* https://github.com/ssitu/ComfyUI_NestedNodeBuilder* https://github.com/longgui0318/comfyui-magic-clothing* https://github.com/atmaranto/ComfyUI-SaveAsScript* https://github.com/ZHO-ZHO-ZHO/ComfyUI-InstantID* https://github.com/AIFSH/ComfyUI-FishSpeech* https://github.com/coolzilj/ComfyUI-Photopea* https://github.com/lks-ai/anynode* Sarav: https://www.youtube.com/@mickmumpitz/videos ( applied stuff )* Sarav: https://www.youtube.com/@latentvision (technical, but infrequent)* look for comfyui node for https://github.com/magic-quill/MagicQuill* “Comfy for Video” resources* Kijai (https://github.com/kijai) pushing out support for Mochi, CogVideoX, AnimateDif, LivePortrait etc* Comfyui node support like LTX https://github.com/Lightricks/ComfyUI-LTXVideo , and HunyuanVideo* FloraFauna AI* Communities: https://www.reddit.com/r/StableDiffusion/, https://www.reddit.com/r/comfyui/Full YouTube EpisodeAs usual, you can find the full video episode on our YouTube (and don't forget to like and subscribe!)Timestamps* 00:00:04 Introduction of hosts and anonymous guest* 00:00:35 Origins of Comfy UI and early Stable Diffusion landscape* 00:02:58 Comfy's background and development of high-res fix* 00:05:37 Area conditioning and compositing in image generation* 00:07:20 Discussion on different AI image models (SD, Flux, etc.)* 00:11:10 Closed source model APIs and community discussions on SD versions* 00:14:41 LoRAs and textual inversion in image generation* 00:18:43 Evaluation methods in the Comfy community* 00:20:05 CLIP models and text encoders in image generation* 00:23:05 Prompt weighting and negative prompting* 00:26:22 Comfy UI's unique features and design choices* 00:31:00 Memory management in Comfy UI* 00:33:50 GPU market share and compatibility issues* 00:35:40 Node design and parameter settings in Comfy UI* 00:38:44 Custom nodes and community contributions* 00:41:40 Video generation models and capabilities* 00:44:47 Comfy UI's development timeline and rise to popularity* 00:48:13 Current state of Comfy UI team and future plans* 00:50:11 Discussion on other Comfy startups and potential text generation supportTranscriptAlessio [00:00:04]: Hey everyone, welcome to the Latent Space podcast. This is Alessio, partner and CTO at Decibel Partners, and I'm joined by my co-host Swyx, founder of Small AI.swyx [00:00:12]: Hey everyone, we are in the Chroma Studio again, but with our first ever anonymous guest, Comfy Anonymous, welcome.Comfy [00:00:19]: Hello.swyx [00:00:21]: I feel like that's your full name, you just go by Comfy, right?Comfy [00:00:24]: Yeah, well, a lot of people just call me Comfy, even when they know my real name. Hey, Comfy.Alessio [00:00:32]: Swyx is the same. You know, not a lot of people call you Shawn.swyx [00:00:35]: Yeah, you have a professional name, right, that people know you by, and then you have a legal name. Yeah, it's fine. How do I phrase this? I think people who are in the know, know that Comfy is like the tool for image generation and now other multimodality stuff. I would say that when I first got started with Stable Diffusion, the star of the show was Automatic 111, right? And I actually looked back at my notes from 2022-ish, like Comfy was already getting started back then, but it was kind of like the up and comer, and your main feature was the flowchart. Can you just kind of rewind to that moment, that year and like, you know, how you looked at the landscape there and decided to start Comfy?Comfy [00:01:10]: Yeah, I discovered Stable Diffusion in 2022, in October 2022. And, well, I kind of started playing around with it. Yes, I, and back then I was using Automatic, which was what everyone was using back then. And so I started with that because I had, it was when I started, I had no idea like how Diffusion works. I didn't know how Diffusion models work, how any of this works, so.swyx [00:01:36]: Oh, yeah. What was your prior background as an engineer?Comfy [00:01:39]: Just a software engineer. Yeah. Boring software engineer.swyx [00:01:44]: But like any, any image stuff, any orchestration, distributed systems, GPUs?Comfy [00:01:49]: No, I was doing basically nothing interesting. Crud, web development? Yeah, a lot of web development, just, yeah, some basic, maybe some basic like automation stuff. Okay. Just. Yeah, no, like, no big companies or anything.swyx [00:02:08]: Yeah, but like already some interest in automations, probably a lot of Python.Comfy [00:02:12]: Yeah, yeah, of course, Python. But I wasn't actually used to like the Node graph interface before I started Comfy UI. It was just, I just thought it was like, oh, like, what's the best way to represent the Diffusion process in the user interface? And then like, oh, well. Well, like, naturally, oh, this is the best way I've found. And this was like with the Node interface. So how I got started was, yeah, so basic October 2022, just like I hadn't written a line of PyTorch before that. So it's completely new. What happened was I kind of got addicted to generating images.Alessio [00:02:58]: As we all did. Yeah.Comfy [00:03:00]: And then I started. I started experimenting with like the high-res fixed in auto, which was for those that don't know, the high-res fix is just since the Diffusion models back then could only generate that low-resolution. So what you would do, you would generate low-resolution image, then upscale, then refine it again. And that was kind of the hack to generate high-resolution images. I really liked generating. Like higher resolution images. So I was experimenting with that. And so I modified the code a bit. Okay. What happens if I, if I use different samplers on the second pass, I was edited the code of auto. So what happens if I use a different sampler? What happens if I use a different, like a different settings, different number of steps? And because back then the. The high-res fix was very basic, just, so. Yeah.swyx [00:04:05]: Now there's a whole library of just, uh, the upsamplers.Comfy [00:04:08]: I think, I think they added a bunch of, uh, of options to the high-res fix since, uh, since, since then. But before that was just so basic. So I wanted to go further. I wanted to try it. What happens if I use a different model for the second, the second pass? And then, well, then the auto code base was, wasn't good enough for. Like, it would have been, uh, harder to implement that in the auto interface than to create my own interface. So that's when I decided to create my own. And you were doing that mostly on your own when you started, or did you already have kind of like a subgroup of people? No, I was, uh, on my own because, because it was just me experimenting with stuff. So yeah, that was it. Then, so I started writing the code January one. 2023, and then I released the first version on GitHub, January 16th, 2023. That's how things got started.Alessio [00:05:11]: And what's, what's the name? Comfy UI right away or? Yeah.Comfy [00:05:14]: Comfy UI. The reason the name, my name is Comfy is people thought my pictures were comfy, so I just, uh, just named it, uh, uh, it's my Comfy UI. So yeah, that's, uh,swyx [00:05:27]: Is there a particular segment of the community that you targeted as users? Like more intensive workflow artists, you know, compared to the automatic crowd or, you know,Comfy [00:05:37]: This was my way of like experimenting with, uh, with new things, like the high risk fixed thing I mentioned, which was like in Comfy, the first thing you could easily do was just chain different models together. And then one of the first things, I think the first times it got a bit of popularity was when I started experimenting with the different, like applying. Prompts to different areas of the image. Yeah. I called it area conditioning, posted it on Reddit and it got a bunch of upvotes. So I think that's when, like, when people first learned of Comfy UI.swyx [00:06:17]: Is that mostly like fixing hands?Comfy [00:06:19]: Uh, no, no, no. That was just, uh, like, let's say, well, it was very, well, it still is kind of difficult to like, let's say you want a mountain, you have an image and then, okay. I'm like, okay. I want the mountain here and I want the, like a, a Fox here.swyx [00:06:37]: Yeah. So compositing the image. Yeah.Comfy [00:06:40]: My way was very easy. It was just like, oh, when you run the diffusion process, you kind of generate, okay. You do pass one pass through the diffusion, every step you do one pass. Okay. This place of the image with this brand, this space, place of the image with the other prop. And then. The entire image with another prop and then just average everything together, every step, and that was, uh, area composition, which I call it. And then, then a month later, there was a paper that came out called multi diffusion, which was the same thing, but yeah, that's, uh,Alessio [00:07:20]: could you do area composition with different models or because you're averaging out, you kind of need the same model.Comfy [00:07:26]: Could do it with, but yeah, I hadn't implemented it. For different models, but, uh, you, you can do it with, uh, with different models if you want, as long as the models share the same latent space, like we, we're supposed to ring a bell every time someone says, yeah, like, for example, you couldn't use like Excel and SD 1.5, because those have a different latent space, but like, uh, yeah, like SD 1.5 models, different ones. You could, you could do that.swyx [00:07:59]: There's some models that try to work in pixel space, right?Comfy [00:08:03]: Yeah. They're very slow. Of course. That's the problem. That that's the, the reason why stable diffusion actually became like popular, like, cause was because of the latent space.swyx [00:08:14]: Small and yeah. Because it used to be latent diffusion models and then they trained it up.Comfy [00:08:19]: Yeah. Cause a pixel pixel diffusion models are just too slow. So. Yeah.swyx [00:08:25]: Have you ever tried to talk to like, like stability, the latent diffusion guys, like, you know, Robin Rombach, that, that crew. Yeah.Comfy [00:08:32]: Well, I used to work at stability.swyx [00:08:34]: Oh, I actually didn't know. Yeah.Comfy [00:08:35]: I used to work at stability. I got, uh, I got hired, uh, in June, 2023.swyx [00:08:42]: Ah, that's the part of the story I didn't know about. Okay. Yeah.Comfy [00:08:46]: So the, the reason I was hired is because they were doing, uh, SDXL at the time and they were basically SDXL. I don't know if you remember it was a base model and then a refiner model. Basically they wanted to experiment, like chaining them together. And then, uh, they saw, oh, right. Oh, this, we can use this to do that. Well, let's hire that guy.swyx [00:09:10]: But they didn't, they didn't pursue it for like SD3. What do you mean? Like the SDXL approach. Yeah.Comfy [00:09:16]: The reason for that approach was because basically they had two models and then they wanted to publish both of them. So they, they trained one on. Lower time steps, which was the refiner model. And then they, the first one was trained normally. And then they went during their test, they realized, oh, like if we string these models together are like quality increases. So let's publish that. It worked. Yeah. But like right now, I don't think many people actually use the refiner anymore, even though it is actually a full diffusion model. Like you can use it on its own. And it's going to generate images. I don't think anyone, people have mostly forgotten about it. But, uh.Alessio [00:10:05]: Can we talk about models a little bit? So stable diffusion, obviously is the most known. I know flux has gotten a lot of traction. Are there any underrated models that people should use more or what's the state of the union?Comfy [00:10:17]: Well, the, the latest, uh, state of the art, at least, yeah, for images there's, uh, yeah, there's flux. There's also SD3.5. SD3.5 is two models. There's a, there's a small one, 2.5B and there's the bigger one, 8B. So it's, it's smaller than flux. So, and it's more, uh, creative in a way, but flux, yeah, flux is the best. People should give SD3.5 a try cause it's, uh, it's different. I won't say it's better. Well, it's better for some like specific use cases. Right. If you want some to make something more like creative, maybe SD3.5. If you want to make something more consistent and flux is probably better.swyx [00:11:06]: Do you ever consider supporting the closed source model APIs?Comfy [00:11:10]: Uh, well, they, we do support them as custom nodes. We actually have some, uh, official custom nodes from, uh, different. Ideogram.swyx [00:11:20]: Yeah. I guess DALI would have one. Yeah.Comfy [00:11:23]: That's, uh, it's just not, I'm not the person that handles that. Sure.swyx [00:11:28]: Sure. Quick question on, on SD. There's a lot of community discussion about the transition from SD1.5 to SD2 and then SD2 to SD3. People still like, you know, very loyal to the previous generations of SDs?Comfy [00:11:41]: Uh, yeah. SD1.5 then still has a lot of, a lot of users.swyx [00:11:46]: The last based model.Comfy [00:11:49]: Yeah. Then SD2 was mostly ignored. It wasn't, uh, it wasn't a big enough improvement over the previous one. Okay.swyx [00:11:58]: So SD1.5, SD3, flux and whatever else. SDXL. SDXL.Comfy [00:12:03]: That's the main one. Stable cascade. Stable cascade. That was a good model. But, uh, that's, uh, the problem with that one is, uh, it got, uh, like SD3 was announced one week after. Yeah.swyx [00:12:16]: It was like a weird release. Uh, what was it like inside of stability actually? I mean, statute of limitations. Yeah. The statute of limitations expired. You know, management has moved. So it's easier to talk about now. Yeah.Comfy [00:12:27]: And inside stability, actually that model was ready, uh, like three months before, but it got, uh, stuck in, uh, red teaming. So basically the product, if that model had released or was supposed to be released by the authors, then it would probably have gotten very popular since it's a, it's a step up from SDXL. But it got all of its momentum stolen. It got stolen by the SD3 announcement. So people kind of didn't develop anything on top of it, even though it's, uh, yeah. It was a good model, at least, uh, completely mostly ignored for some reason. Likeswyx [00:13:07]: I think the naming as well matters. It seemed like a branch off of the main, main tree of development. Yeah.Comfy [00:13:15]: Well, it was different researchers that did it. Yeah. Yeah. Very like, uh, good model. Like it's the Worcestershire authors. I don't know if I'm pronouncing it correctly. Yeah. Yeah. Yeah.swyx [00:13:28]: I actually met them in Vienna. Yeah.Comfy [00:13:30]: They worked at stability for a bit and they left right after the Cascade release.swyx [00:13:35]: This is Dustin, right? No. Uh, Dustin's SD3. Yeah.Comfy [00:13:38]: Dustin is a SD3 SDXL. That's, uh, Pablo and Dome. I think I'm pronouncing his name correctly. Yeah. Yeah. Yeah. Yeah. That's very good.swyx [00:13:51]: It seems like the community is very, they move very quickly. Yeah. Like when there's a new model out, they just drop whatever the current one is. And they just all move wholesale over. Like they don't really stay to explore the full capabilities. Like if, if the stable cascade was that good, they would have AB tested a bit more. Instead they're like, okay, SD3 is out. Let's go. You know?Comfy [00:14:11]: Well, I find the opposite actually. The community doesn't like, they only jump on a new model when there's a significant improvement. Like if there's a, only like a incremental improvement, which is what, uh, most of these models are going to have, especially if you, cause, uh, stay the same parameter count. Yeah. Like you're not going to get a massive improvement, uh, into like, unless there's something big that, that changes. So, uh. Yeah.swyx [00:14:41]: And how are they evaluating these improvements? Like, um, because there's, it's a whole chain of, you know, comfy workflows. Yeah. How does, how does one part of the chain actually affect the whole process?Comfy [00:14:52]: Are you talking on the model side specific?swyx [00:14:54]: Model specific, right? But like once you have your whole workflow based on a model, it's very hard to move.Comfy [00:15:01]: Uh, not, well, not really. Well, it depends on your, uh, depends on their specific kind of the workflow. Yeah.swyx [00:15:09]: So I do a lot of like text and image. Yeah.Comfy [00:15:12]: When you do change, like most workflows are kind of going to be complete. Yeah. It's just like, you might have to completely change your prompt completely change. Okay.swyx [00:15:24]: Well, I mean, then maybe the question is really about evals. Like what does the comfy community do for evals? Just, you know,Comfy [00:15:31]: Well, that they don't really do that. It's more like, oh, I think this image is nice. So that's, uh,swyx [00:15:38]: They just subscribe to Fofr AI and just see like, you know, what Fofr is doing. Yeah.Comfy [00:15:43]: Well, they just, they just generate like it. Like, I don't see anyone really doing it. Like, uh, at least on the comfy side, comfy users, they, it's more like, oh, generate images and see, oh, this one's nice. It's like, yeah, it's not, uh, like the, the more, uh, like, uh, scientific, uh, like, uh, like checking that's more on specifically on like model side. If, uh, yeah, but there is a lot of, uh, vibes also, cause it is a like, uh, artistic, uh, you can create a very good model that doesn't generate nice images. Cause most images on the internet are ugly. So if you, if that's like, if you just, oh, I have the best model at 10th giant, it's super smart. I created on all the, like I've trained on just all the images on the internet. The images are not going to look good. So yeah.Alessio [00:16:42]: Yeah.Comfy [00:16:43]: They're going to be very consistent. But yeah. People like, it's not going to be like the, the look that people are going to be expecting from, uh, from a model. So. Yeah.swyx [00:16:54]: Can we talk about LoRa's? Cause we thought we talked about models then like the next step is probably LoRa's. Before, I actually, I'm kind of curious how LoRa's entered the tool set of the image community because the LoRa paper was 2021. And then like, there was like other methods like textual inversion that was popular at the early SD stage. Yeah.Comfy [00:17:13]: I can't even explain the difference between that. Yeah. Textual inversions. That's basically what you're doing is you're, you're training a, cause well, yeah. Stable diffusion. You have the diffusion model, you have text encoder. So basically what you're doing is training a vector that you're going to pass to the text encoder. It's basically you're training a new word. Yeah.swyx [00:17:37]: It's a little bit like representation engineering now. Yeah.Comfy [00:17:40]: Yeah. Basically. Yeah. You're just, so yeah, if you know how like the text encoder works, basically you have, you take your, your words of your product, you convert those into tokens with the tokenizer and those are converted into vectors. Basically. Yeah. Each token represents a different vector. So each word presents a vector. And those, depending on your words, that's the list of vectors that get passed to the text encoder, which is just. Yeah. Yeah. I'm just a stack of, of attention. Like basically it's a very close to LLM architecture. Yeah. Yeah. So basically what you're doing is just training a new vector. We're saying, well, I have all these images and I want to know which word does that represent? And it's going to get like, you train this vector and then, and then when you use this vector, it hopefully generates. Like something similar to your images. Yeah.swyx [00:18:43]: I would say it's like surprisingly sample efficient in picking up the concept that you're trying to train it on. Yeah.Comfy [00:18:48]: Well, people have kind of stopped doing that even though back as like when I was at Stability, we, we actually did train internally some like textual versions on like T5 XXL actually worked pretty well. But for some reason, yeah, people don't use them. And also they might also work like, like, yeah, this is something and probably have to test, but maybe if you train a textual version, like on T5 XXL, it might also work with all the other models that use T5 XXL because same thing with like, like the textual inversions that, that were trained for SD 1.5, they also kind of work on SDXL because SDXL has the, has two text encoders. And one of them is the same as the, as the SD 1.5 CLIP-L. So those, they actually would, they don't work as strongly because they're only applied to one of the text encoders. But, and the same thing for SD3. SD3 has three text encoders. So it works. It's still, you can still use your textual version SD 1.5 on SD3, but it's just a lot weaker because now there's three text encoders. So it gets even more diluted. Yeah.swyx [00:20:05]: Do people experiment a lot on, just on the CLIP side, there's like Siglip, there's Blip, like do people experiment a lot on those?Comfy [00:20:12]: You can't really replace. Yeah.swyx [00:20:14]: Because they're trained together, right? Yeah.Comfy [00:20:15]: They're trained together. So you can't like, well, what I've seen people experimenting with is a long CLIP. So basically someone fine tuned the CLIP model to accept longer prompts.swyx [00:20:27]: Oh, it's kind of like long context fine tuning. Yeah.Comfy [00:20:31]: So, so like it's, it's actually supported in Core Comfy.swyx [00:20:35]: How long is long?Comfy [00:20:36]: Regular CLIP is 77 tokens. Yeah. Long CLIP is 256. Okay. So, but the hack that like you've, if you use stable diffusion 1.5, you've probably noticed, oh, it still works if I, if I use long prompts, prompts longer than 77 words. Well, that's because the hack is to just, well, you split, you split it up in chugs of 77, your whole big prompt. Let's say you, you give it like the massive text, like the Bible or something, and it would split it up in chugs of 77 and then just pass each one through the CLIP and then just cut anything together at the end. It's not ideal, but it actually works.swyx [00:21:26]: Like the positioning of the words really, really matters then, right? Like this is why order matters in prompts. Yeah.Comfy [00:21:33]: Yeah. Like it, it works, but it's, it's not ideal, but it's what people expect. Like if, if someone gives a huge prompt, they expect at least some of the concepts at the end to be like present in the image. But usually when they give long prompts, they, they don't, they like, they don't expect like detail, I think. So that's why it works very well.swyx [00:21:58]: And while we're on this topic, prompts waiting, negative comments. Negative prompting all, all sort of similar part of this layer of the stack. Yeah.Comfy [00:22:05]: The, the hack for that, which works on CLIP, like it, basically it's just for SD 1.5, well, for SD 1.5, the prompt waiting works well because CLIP L is a, is not a very deep model. So you have a very high correlation between, you have the input token, the index of the input token vector. And the output token, they're very, the concepts are very close, closely linked. So that means if you interpolate the vector from what, well, the, the way Comfy UI does it is it has, okay, you have the vector, you have an empty prompt. So you have a, a chunk, like a CLIP output for the empty prompt, and then you have the one for your prompt. And then it interpolates from that, depending on your prompt. Yeah.Comfy [00:23:07]: So that's how it, how it does prompt waiting. But this stops working the deeper your text encoder is. So on T5X itself, it doesn't work at all. So. Wow.swyx [00:23:20]: Is that a problem for people? I mean, cause I'm used to just move, moving up numbers. Probably not. Yeah.Comfy [00:23:25]: Well.swyx [00:23:26]: So you just use words to describe, right? Cause it's a bigger language model. Yeah.Comfy [00:23:30]: Yeah. So. Yeah. So honestly it might be good, but I haven't seen many complaints on Flux that it's not working. So, cause I guess people can sort of get around it with, with language. So. Yeah.swyx [00:23:46]: Yeah. And then coming back to LoRa's, now the, the popular way to, to customize models is LoRa's. And I saw you also support Locon and LoHa, which I've never heard of before.Comfy [00:23:56]: There's a bunch of, cause what, what the LoRa is essentially is. Instead of like, okay, you have your, your model and then you want to fine tune it. So instead of like, what you could do is you could fine tune the entire thing, but that's a bit heavy. So to speed things up and make things less heavy, what you can do is just fine tune some smaller weights, like basically two, two matrices that when you multiply like two low rank matrices and when you multiply them together, gives a, represents a difference between trained weights and your base weights. So by training those two smaller matrices, that's a lot less heavy. Yeah.Alessio [00:24:45]: And they're portable. So you're going to share them. Yeah. It's like easier. And also smaller.Comfy [00:24:49]: Yeah. That's the, how LoRa's work. So basically, so when, when inferencing you, you get an inference with them pretty efficiently, like how ComputeWrite does it. It just, when you use a LoRa, it just applies it straight on the weights so that there's only a small delay at the base, like before the sampling to when it applies the weights and then it just same speed as, as before. So for, for inference, it's, it's not that bad, but, and then you have, so basically all the LoRa types like LoHa, LoCon, everything, that's just different ways of representing that like. Basically, you can call it kind of like compression, even though it's not really compression, it's just different ways of represented, like just, okay, I want to train a different on the difference on the weights. What's the best way to represent that difference? There's the basic LoRa, which is just, oh, let's multiply these two matrices together. And then there's all the other ones, which are all different algorithms. So. Yeah.Alessio [00:25:57]: So let's talk about LoRa. Let's talk about what comfy UI actually is. I think most people have heard of it. Some people might've seen screenshots. I think fewer people have built very complex workflows. So when you started, automatic was like the super simple way. What were some of the choices that you made? So the node workflow, is there anything else that stands out as like, this was like a unique take on how to do image generation workflows?Comfy [00:26:22]: Well, I feel like, yeah, back then everyone was trying to make like easy to use interface. Yeah. So I'm like, well, everyone's trying to make an easy to use interface.swyx [00:26:32]: Let's make a hard to use interface.Comfy [00:26:37]: Like, so like, I like, I don't need to do that, everyone else doing it. So let me try something like, let me try to make a powerful interface that's not easy to use. So.swyx [00:26:52]: So like, yeah, there's a sort of node execution engine. Yeah. Yeah. And it actually lists, it has this really good list of features of things you prioritize, right? Like let me see, like sort of re-executing from, from any parts of the workflow that was changed, asynchronous queue system, smart memory management, like all this seems like a lot of engineering that. Yeah.Comfy [00:27:12]: There's a lot of engineering in the back end to make things, cause I was always focused on making things work locally very well. Cause that's cause I was using it locally. So everything. So there's a lot of, a lot of thought and working by getting everything to run as well as possible. So yeah. ConfUI is actually more of a back end, at least, well, not all the front ends getting a lot more development, but, but before, before it was, I was pretty much only focused on the backend. Yeah.swyx [00:27:50]: So v0.1 was only August this year. Yeah.Comfy [00:27:54]: With the new front end. Before there was no versioning. So yeah. Yeah. Yeah.swyx [00:27:57]: And so what was the big rewrite for the 0.1 and then the 1.0?Comfy [00:28:02]: Well, that's more on the front end side. That's cause before that it was just like the UI, what, cause when I first wrote it, I just, I said, okay, how can I make, like, I can do web development, but I don't like doing it. Like what's the easiest way I can slap a node interface on this. And then I found this library. Yeah. Like JavaScript library.swyx [00:28:26]: Live graph?Comfy [00:28:27]: Live graph.swyx [00:28:28]: Usually people will go for like react flow for like a flow builder. Yeah.Comfy [00:28:31]: But that seems like too complicated. So I didn't really want to spend time like developing the front end. So I'm like, well, oh, light graph. This has the whole node interface. So, okay. Let me just plug that into, to my backend.swyx [00:28:49]: I feel like if Streamlit or Gradio offered something that you would have used Streamlit or Gradio cause it's Python. Yeah.Comfy [00:28:54]: Yeah. Yeah. Yeah.Comfy [00:29:00]: Yeah.Comfy [00:29:14]: Yeah. logic and your backend logic and just sticks them together.swyx [00:29:20]: It's supposed to be easy for you guys. If you're a Python main, you know, I'm a JS main, right? Okay. If you're a Python main, it's supposed to be easy.Comfy [00:29:26]: Yeah, it's easy, but it makes your whole software a huge mess.swyx [00:29:30]: I see, I see. So you're mixing concerns instead of separating concerns?Comfy [00:29:34]: Well, it's because... Like frontend and backend. Frontend and backend should be well separated with a defined API. Like that's how you're supposed to do it. Smart people disagree. It just sticks everything together. It makes it easy to like a huge mess. And also it's, there's a lot of issues with Gradio. Like it's very good if all you want to do is just get like slap a quick interface on your, like to show off your ML project. Like that's what it's made for. Yeah. Like there's no problem using it. Like, oh, I have my, I have my code. I just wanted a quick interface on it. That's perfect. Like use Gradio. But if you want to make something that's like a real, like real software that will last a long time and will be easy to maintain, then I would avoid it. Yeah.swyx [00:30:32]: So your criticism is Streamlit and Gradio are the same. I mean, those are the same criticisms.Comfy [00:30:37]: Yeah, Streamlit I haven't used as much. Yeah, I just looked a bit.swyx [00:30:43]: Similar philosophy.Comfy [00:30:44]: Yeah, it's similar. It's just, it just seems to me like, okay, for quick, like AI demos, it's perfect.swyx [00:30:51]: Yeah. Going back to like the core tech, like asynchronous queues, slow re-execution, smart memory management, you know, anything that you were very proud of or was very hard to figure out?Comfy [00:31:00]: Yeah. The thing that's the biggest pain in the ass is probably the memory management. Yeah.swyx [00:31:05]: Were you just paging models in and out or? Yeah.Comfy [00:31:08]: Before it was just, okay, load the model, completely unload it. Then, okay, that, that works well when you, your model are small, but if your models are big and it takes sort of like, let's say someone has a, like a, a 4090, and the model size is 10 gigabytes, that can take a few seconds to like load and load, load and load, so you want to try to keep things like in memory, in the GPU memory as much as possible. What Comfy UI does right now is it. It tries to like estimate, okay, like, okay, you're going to sample this model, it's going to take probably this amount of memory, let's remove the models, like this amount of memory that's been loaded on the GPU and then just execute it. But so there's a fine line between just because try to remove the least amount of models that are already loaded. Because as fans, like Windows drivers, and one other problem is the NVIDIA driver on Windows by default, because there's a way to, there's an option to disable that feature, but by default it, like, if you start loading, you can overflow your GPU memory and then it's, the driver's going to automatically start paging to RAM. But the problem with that is it's, it makes everything extremely slow. So when you see people complaining, oh, this model, it works, but oh, s**t, it starts slowing down a lot, that's probably what's happening. So it's basically you have to just try to get, use as much memory as possible, but not too much, or else things start slowing down, or people get out of memory, and then just find, try to find that line where, oh, like the driver on Windows starts paging and stuff. Yeah. And the problem with PyTorch is it's, it's high levels, don't have that much fine-grained control over, like, specific memory stuff, so kind of have to leave, like, the memory freeing to, to Python and PyTorch, which is, can be annoying sometimes.swyx [00:33:32]: So, you know, I think one thing is, as a maintainer of this project, like, you're designing for a very wide surface area of compute, like, you even support CPUs.Comfy [00:33:42]: Yeah, well, that's... That's just, for PyTorch, PyTorch supports CPUs, so, yeah, it's just, that's not, that's not hard to support.swyx [00:33:50]: First of all, is there a market share estimate, like, is it, like, 70% NVIDIA, like, 30% AMD, and then, like, miscellaneous on Apple, Silicon, or whatever?Comfy [00:33:59]: For Comfy? Yeah. Yeah, and, yeah, I don't know the market share.swyx [00:34:03]: Can you guess?Comfy [00:34:04]: I think it's mostly NVIDIA. Right. Because, because AMD, the problem, like, AMD works horribly on Windows. Like, on Linux, it works fine. It's, it's lower than the price equivalent NVIDIA GPU, but it works, like, you can use it, you generate images, everything works. On Linux, on Windows, you might have a hard time, so, that's the problem, and most people, I think most people who bought AMD probably use Windows. They probably aren't going to switch to Linux, so... Yeah. So, until AMD actually, like, ports their, like, raw cam to, to Windows properly, and then there's actually PyTorch, I think they're, they're doing that, they're in the process of doing that, but, until they get it, they get a good, like, PyTorch raw cam build that works on Windows, it's, like, they're going to have a hard time. Yeah.Alessio [00:35:06]: We got to get George on it. Yeah. Well, he's trying to get Lisa Su to do it, but... Let's talk a bit about, like, the node design. So, unlike all the other text-to-image, you have a very, like, deep, so you have, like, a separate node for, like, clip and code, you have a separate node for, like, the case sampler, you have, like, all these nodes. Going back to, like, the making it easy versus making it hard, but, like, how much do people actually play with all the settings, you know? Kind of, like, how do you guide people to, like, hey, this is actually going to be very impactful versus this is maybe, like, less impactful, but we still want to expose it to you?Comfy [00:35:40]: Well, I try to... I try to expose, like, I try to expose everything or, but, yeah, at least for the, but for things, like, for example, for the samplers, like, there's, like, yeah, four different sampler nodes, which go in easiest to most advanced. So, yeah, if you go, like, the easy node, the regular sampler node, that's, you have just the basic settings. But if you use, like, the sampler advanced... If you use, like, the custom advanced node, that, that one you can actually, you'll see you have, like, different nodes.Alessio [00:36:19]: I'm looking it up now. Yeah. What are, like, the most impactful parameters that you use? So, it's, like, you know, you can have more, but, like, which ones, like, really make a difference?Comfy [00:36:30]: Yeah, they all do. They all have their own, like, they all, like, for example, yeah, steps. Usually you want steps, you want them to be as low as possible. But you want, if you're optimizing your workflow, you want to, you lower the steps until, like, the images start deteriorating too much. Because that, yeah, that's the number of steps you're running the diffusion process. So, if you want things to be faster, lower is better. But, yeah, CFG, that's more, you can kind of see that as the contrast of the image. Like, if your image looks too bursty. Then you can lower the CFG. So, yeah, CFG, that's how, yeah, that's how strongly the, like, the negative versus positive prompt. Because when you sample a diffusion model, it's basically a negative prompt. It's just, yeah, positive prediction minus negative prediction.swyx [00:37:32]: Contrastive loss. Yeah.Comfy [00:37:34]: It's positive minus negative, and the CFG does the multiplier. Yeah. Yeah. Yeah, so.Alessio [00:37:41]: What are, like, good resources to understand what the parameters do? I think most people start with automatic, and then they move over, and it's, like, snap, CFG, sampler, name, scheduler, denoise. Read it.Comfy [00:37:53]: But, honestly, well, it's more, it's something you should, like, try out yourself. I don't know, you don't necessarily need to know how it works to, like, what it does. Because even if you know, like, CFGO, it's, like, positive minus negative prompt. Yeah. So the only thing you know at CFG is if it's 1.0, then that means the negative prompt isn't applied. It also means sampling is two times faster. But, yeah. But other than that, it's more, like, you should really just see what it does to the images yourself, and you'll probably get a more intuitive understanding of what these things do.Alessio [00:38:34]: Any other nodes or things you want to shout out? Like, I know the animate diff IP adapter. Those are, like, some of the most popular ones. Yeah. What else comes to mind?Comfy [00:38:44]: Not nodes, but there's, like, what I like is when some people, sometimes they make things that use ComfyUI as their backend. Like, there's a plugin for Krita that uses ComfyUI as its backend. So you can use, like, all the models that work in Comfy in Krita. And I think I've tried it once. But I know a lot of people use it, and it's probably really nice, so.Alessio [00:39:15]: What's the craziest node that people have built, like, the most complicated?Comfy [00:39:21]: Craziest node? Like, yeah. I know some people have made, like, video games in Comfy with, like, stuff like that. So, like, someone, like, I remember, like, yeah, last, I think it was last year, someone made, like, a, like, Wolfenstein 3D in Comfy. Of course. And then one of the inputs was, oh, you can generate a texture, and then it changes the texture in the game. So you can plug it to, like, the workflow. And there's a lot of, if you look there, there's a lot of crazy things people do, so. Yeah.Alessio [00:39:59]: And now there's, like, a node register that people can use to, like, download nodes. Yeah.Comfy [00:40:04]: Like, well, there's always been the, like, the ComfyUI manager. Yeah. But we're trying to make this more, like, I don't know, official, like, with, yeah, with the node registry. Because before the node registry, the, like, okay, how did your custom node get into ComfyUI manager? That's the guy running it who, like, every day he searched GitHub for new custom nodes and added dev annually to his custom node manager. So we're trying to make it less effortless. So we're trying to make it less effortless for him, basically. Yeah.Alessio [00:40:40]: Yeah. But I was looking, I mean, there's, like, a YouTube download node. There's, like, this is almost like, you know, a data pipeline more than, like, an image generation thing at this point. It's, like, you can get data in, you can, like, apply filters to it, you can generate data out.Comfy [00:40:54]: Yeah. You can do a lot of different things. Yeah. So I'm thinking, I think what I did is I made it easy to make custom nodes. So I think that helped a lot. I think that helped a lot for, like, the ecosystem because it is very easy to just make a node. So, yeah, a bit too easy sometimes. Then we have the issue where there's a lot of custom node packs which share similar nodes. But, well, that's, yeah, something we're trying to solve by maybe bringing some of the functionality into the core. Yeah. Yeah. Yeah.Alessio [00:41:36]: And then there's, like, video. People can do video generation. Yeah.Comfy [00:41:40]: Video, that's, well, the first video model was, like, stable video diffusion, which was last, yeah, exactly last year, I think. Like, one year ago. But that wasn't a true video model. So it was...swyx [00:41:55]: It was, like, moving images? Yeah.Comfy [00:41:57]: I generated video. What I mean by that is it's, like, it's still 2D Latents. It's basically what I'm trying to do. So what they did is they took SD2, and then they added some temporal attention to it, and then trained it on videos and all. So it's kind of, like, animated, like, same idea, basically. Why I say it's not a true video model is that you still have, like, the 2D Latents. Like, a true video model, like Mochi, for example, would have 3D Latents. Mm-hmm.Alessio [00:42:32]: Which means you can, like, move through the space, basically. It's the difference. You're not just kind of, like, reorienting. Yeah.Comfy [00:42:39]: And it's also, well, it's also because you have a temporal VAE. Mm-hmm. Also, like, Mochi has a temporal VAE that compresses on, like, the temporal direction, also. So that's something you don't have with, like, yeah, animated diff and stable video diffusion. They only, like, compress spatially, not temporally. Mm-hmm. Right. So, yeah. That's why I call that, like, true video models. There's, yeah, there's actually a few of them, but the one I've implemented in comfy is Mochi, because that seems to be the best one so far. Yeah.swyx [00:43:15]: We had AJ come and speak at the stable diffusion meetup. The other open one I think I've seen is COG video. Yeah.Comfy [00:43:21]: COG video. Yeah. That one's, yeah, it also seems decent, but, yeah. Chinese, so we don't use it. No, it's fine. It's just, yeah, I could. Yeah. It's just that there's a, it's not the only one. There's also a few others, which I.swyx [00:43:36]: The rest are, like, closed source, right? Like, Cling. Yeah.Comfy [00:43:39]: Closed source, there's a bunch of them. But I mean, open. I've seen a few of them. Like, I can't remember their names, but there's COG videos, the big, the big one. Then there's also a few of them that released at the same time. There's one that released at the same time as SSD 3.5, same day, which is why I don't remember the name.swyx [00:44:02]: We should have a release schedule so we don't conflict on each of these things. Yeah.Comfy [00:44:06]: I think SD 3.5 and Mochi released on the same day. So everything else was kind of drowned, completely drowned out. So for some reason, lots of people picked that day to release their stuff.Comfy [00:44:21]: Yeah. Which is, well, shame for those. And I think Omnijet also released the same day, which also seems interesting. Yeah. Yeah.Alessio [00:44:30]: What's Comfy? So you are Comfy. And then there's like, comfy.org. I know we do a lot of things for, like, news research and those guys also have kind of like a more open source thing going on. How do you work? Like you mentioned, you mostly work on like, the core piece of it. And then what...Comfy [00:44:47]: Maybe I should fade it in because I, yeah, I feel like maybe, yeah, I only explain part of the story. Right. Yeah. Maybe I should explain the rest. So yeah. So yeah. Basically, January, that's when the first January 2023, January 16, 2023, that's when Amphi was first released to the public. Then, yeah, did a Reddit post about the area composition thing somewhere in, I don't remember exactly, maybe end of January, beginning of February. And then someone, a YouTuber, made a video about it, like Olivio, he made a video about Amphi in March 2023. I think that's when it was a real burst of attention. And by that time, I was continuing to develop it and it was getting, people were starting to use it more, which unfortunately meant that I had first written it to do like experiments, but then my time to do experiments went down. It started going down, because people were actually starting to use it then. Like, I had to, and I said, well, yeah, time to add all these features and stuff. Yeah, and then I got hired by Stability June, 2023. Then I made, basically, yeah, they hired me because they wanted the SD-XL. So I got the SD-XL working very well withітhe UI, because they were experimenting withámphi.house.com. Actually, the SDX, how the SDXL released worked is they released, for some reason, like they released the code first, but they didn't release the model checkpoint. So they released the code. And then, well, since the research was related to code, I released the code in Compute 2. And then the checkpoints were basically early access. People had to sign up and they only allowed a lot of people from edu emails. Like if you had an edu email, like they gave you access basically to the SDXL 0.9. And, well, that leaked. Right. Of course, because of course it's going to leak if you do that. Well, the only way people could easily use it was with Comfy. So, yeah, people started using. And then I fixed a few of the issues people had. So then the big 1.0 release happened. And, well, Comfy UI was the only way a lot of people could actually run it on their computers. Because it just like automatic was so like inefficient and bad that most people couldn't actually, like it just wouldn't work. Like because he did a quick implementation. So people were forced. To use Comfy UI, and that's how it became popular because people had no choice.swyx [00:47:55]: The growth hack.Comfy [00:47:56]: Yeah.swyx [00:47:56]: Yeah.Comfy [00:47:57]: Like everywhere, like people who didn't have the 4090, they had like, who had just regular GPUs, they didn't have a choice.Alessio [00:48:05]: So yeah, I got a 4070. So think of me. And so today, what's, is there like a core Comfy team or?Comfy [00:48:13]: Uh, yeah, well, right now, um, yeah, we are hiring. Okay. Actually, so right now core, like, um, the core core itself, it's, it's me. Uh, but because, uh, the reason where folks like all the focus has been mostly on the front end right now, because that's the thing that's been neglected for a long time. So, uh, so most of the focus right now is, uh, all on the front end, but we are, uh, yeah, we will soon get, uh, more people to like help me with the actual backend stuff. Yeah. So, no, I'm not going to say a hundred percent because that's why once the, once we have our V one release, which is because it'd be the package, come fee-wise with the nice interface and easy to install on windows and hopefully Mac. Uh, yeah. Yeah. Once we have that, uh, we're going to have to, lots of stuff to do on the backend side and also the front end side, but, uh.Alessio [00:49:14]: What's the release that I'm on the wait list. What's the timing?Comfy [00:49:18]: Uh, soon. Uh, soon. Yeah, I don't want to promise a release date. We do have a release date we're targeting, but I'm not sure if it's public. Yeah, and we're still going to continue doing the open source, making MPUI the best way to run stable infusion models. At least the open source side, it's going to be the best way to run models locally. But we will have a few things to make money from it, like cloud inference or that type of thing. And maybe some things for some enterprises.swyx [00:50:08]: I mean, a few questions on that. How do you feel about the other comfy startups?Comfy [00:50:11]: I mean, I think it's great. They're using your name. Yeah, well, it's better they use comfy than they use something else. Yeah, that's true. It's fine. We're going to try not to... We don't want to... We want people to use comfy. Like I said, it's better that people use comfy than something else. So as long as they use comfy, I think it helps the ecosystem. Because more people, even if they don't contribute directly, the fact that they are using comfy means that people are more likely to join the ecosystem. So, yeah.swyx [00:50:57]: And then would you ever do text?Comfy [00:50:59]: Yeah, well, you can already do text with some custom nodes. So, yeah, it's something we like. Yeah, it's something I've wanted to eventually add to core, but it's more like not a very... It's a very high priority. But because a lot of people use text for prompt enhancement and other things like that. So, yeah, it's just that my focus has always been on diffusion models. Yeah, unless some text diffusion model comes out.swyx [00:51:30]: Yeah, David Holtz is investing a lot in text diffusion.Comfy [00:51:34]: Yeah, well, if a good one comes out, then we'll probably implement it since it fits with the whole...swyx [00:51:39]: Yeah, I mean, I imagine it's going to be a close source to Midjourney. Yeah.Comfy [00:51:43]: Well, if an open one comes out, then I'll probably implement it.Alessio [00:51:54]: Cool, comfy. Thanks so much for coming on. This was fun. Bye. Get full access to Latent Space at www.latent.space/subscribe

En este episodio de Charlas Técnicas de AWS, exploramos el mundo de Python con Denny Pérez, Directora de la Python Software Foundation. Desde su viaje de Dominicana a Canadá hasta su rol en la comunidad global de Python, Denny comparte su experiencia, sus contribuciones y las posibilidades de Python en distintas industrias.Este es el episodio 17 de la temporada 5ta.Tabla de Contenidos:01:10 Conociendo a nuestra invitada Denny Pérez, de Dominicana a Canadá03:20 ¿Por qué Python y no Java?04:52 De Python a Directora de la Python Software Foundation (PSF)06:06 ¿Qué funciones realiza la PSF?07:10 Funciones de los Directores de la PSF09:59 ¿Cómo contribuir en Python?12:00 Frameworks de aplicaciones en Python13:56 Python 3.13, novedades21:23 Usando Streamlit para construir aplicaciones23:20 Python en el mundo de la Inteligencia Artificial25:59 Python en el espacio32:02 Python en el mundo hispanohablante41:06 Historias de Terror y errores comunes con PythonEl infierno de las dependenciasExponiendo tus claves al mundoBucles sin finIntegración con terceros53:39 Cierre: El futuro de PythonRedes Sociales de nuestra invitada:LinkedIN: https://www.linkedin.com/in/dennyperez18/Twitter: https://x.com/dennyperez18Otras redes: https://linktr.ee/Dennyperez18Recursos sobre Python y la PSFPython: python.orgPython Software Foundation: https://www.python.org/psfPyladies: https://pyladies.com/Hablemos Python: https://hablemospython.dev/Frameworks de aplicacionesFastAPI: https://tert0.github.io/fastapi-framework/Django: https://docs.djangoproject.com/en/5.1/Django Software Foundation (DSF): https://www.djangoproject.com/foundation/Python en Inteligencia Artificial y Ciencia de Datosdeeplearning.ai - AI for Python Beginners: https://www.deeplearning.ai/short-courses/ai-python-for-beginners/Safety (Seguridad en paquetes Python): https://pypi.org/project/safety/Python en el Espacio y CienciaVisualizando el Espacio: App de Satélites con Streamlit: https://dev.to/aws-espanol/visualizando-el-espacio-como-construir-tu-propia-app-de-satelites-con-ia-y-streamlit-4pp4NASA Ingenuity Helicopter: https://github.com/readme/featured/nasa-ingenuity-helicopterFprime: https://github.com/nasa/fprimeAstropy: https://www.astropy.org/Proyecto SKA Astronomía: https://www.linkedin.com/posts/gruizesteban_cosmos-satellites-universe-activity-7248935127043493888-lnorLíneas Nazca: https://www.linkedin.com/posts/gruizesteban_ai-genai-drones-activity-7246470519204548608-5yWb✉️ Si quieren escribirnos pueden hacerlo a este correo: podcast-aws-espanol@amazon.comPodes encontrar el podcast en este link: https://aws-espanol.buzzsprout.com/O en tu plataforma de podcast favoritaMás información y tutoriales en el canal de youtube de Charlas Técnicas☆☆ NUESTRAS REDES SOCIALES ☆☆

Je vous présente Gourou Fabulus, version dématérialisée de moi-même, première étape importante dans mon processus de gourouification. Tester Gourou Fabulus : https://homofabulus.com/gourou Peut aussi servir de tutoriel pour coder un chatbot pouvant discuter avec vos documents (application RAG) avec les frameworks python Llamaindex et streamlit.

Хочете долучитися до команди MacPaw? Хутчіш сюди: https://macpaw.com/careers/?utm_source=youtube&utm_medium=free&utm_campaign=#application 

Ever wondered how to choose between fine-tuning an open-source LLM and using an existing model API? I had a blast chatting with Doris Lee and Yusuf Ozuysal from Snowflake on The Ravit Show. Here are the highlights of our discussion: - Introductions and factors influencing the choice between fine-tuning an open-source LLM and using an existing model API - Examples of how open source has impacted model bias awareness, code security, and talent attraction - Key considerations when assembling a scalable and reproducible open source gen AI stack - How accelerating pre-training contributes to AI application performance - Examples of using Modin and Streamlit in building open source AI applications Stay tuned for more insights from the Snowflake Summit! #data #ai #snowflakesummit #snowflakeflake2024 #theravitshow

Disclaimer: We recorded this episode ~1.5 months ago, timing for the FastHTML release. It then got bottlenecked by Llama3.1, Winds of AI Winter, and SAM2 episodes, so we're a little late. Since then FastHTML was released, swyx is building an app in it for AINews, and Anthropic has also released their prompt caching API. Remember when Dylan Patel of SemiAnalysis coined the GPU Rich vs GPU Poor war? (if not, see our pod with him). The idea was that if you're GPU poor you shouldn't waste your time trying to solve GPU rich problems (i.e. pre-training large models) and are better off working on fine-tuning, optimized inference, etc. Jeremy Howard (see our “End of Finetuning” episode to catchup on his background) and Eric Ries founded Answer.AI to do exactly that: “Practical AI R&D”, which is very in-line with the GPU poor needs. For example, one of their first releases was a system based on FSDP + QLoRA that let anyone train a 70B model on two NVIDIA 4090s. Since then, they have come out with a long list of super useful projects (in no particular order, and non-exhaustive):* FSDP QDoRA: this is just as memory efficient and scalable as FSDP/QLoRA, and critically is also as accurate for continued pre-training as full weight training.* Cold Compress: a KV cache compression toolkit that lets you scale sequence length without impacting speed.* colbert-small: state of the art retriever at only 33M params* JaColBERTv2.5: a new state-of-the-art retrievers on all Japanese benchmarks.* gpu.cpp: portable GPU compute for C++ with WebGPU.* Claudette: a better Anthropic API SDK. They also recently released FastHTML, a new way to create modern interactive web apps. Jeremy recently released a 1 hour “Getting started” tutorial on YouTube; while this isn't AI related per se, but it's close to home for any AI Engineer who are looking to iterate quickly on new products: In this episode we broke down 1) how they recruit 2) how they organize what to research 3) and how the community comes together. At the end, Jeremy gave us a sneak peek at something new that he's working on that he calls dialogue engineering: So I've created a new approach. It's not called prompt engineering. I'm creating a system for doing dialogue engineering. It's currently called AI magic. I'm doing most of my work in this system and it's making me much more productive than I was before I used it.He explains it a bit more ~44:53 in the pod, but we'll just have to wait for the public release to figure out exactly what he means.Timestamps* [00:00:00] Intro by Suno AI* [00:03:02] Continuous Pre-Training is Here* [00:06:07] Schedule-Free Optimizers and Learning Rate Schedules* [00:07:08] Governance and Structural Issues within OpenAI and Other AI Labs* [00:13:01] How Answer.ai works* [00:23:40] How to Recruit Productive Researchers* [00:27:45] Building a new BERT* [00:31:57] FSDP, QLoRA, and QDoRA: Innovations in Fine-Tuning Large Models* [00:36:36] Research and Development on Model Inference Optimization* [00:39:49] FastHTML for Web Application Development* [00:46:53] AI Magic & Dialogue Engineering* [00:52:19] AI wishlist & predictionsShow Notes* Jeremy Howard* Previously on Latent Space: The End of Finetuning, NeurIPS Startups* Answer.ai* Fast.ai* FastHTML* answerai-colbert-small-v1* gpu.cpp* Eric Ries* Aaron DeFazio* Yi Tai* Less Wright* Benjamin Warner* Benjamin Clavié* Jono Whitaker* Austin Huang* Eric Gilliam* Tim Dettmers* Colin Raffel* Sebastian Raschka* Carson Gross* Simon Willison* Sepp Hochreiter* Llama3.1 episode* Snowflake Arctic* Ranger Optimizer* Gemma.cpp* HTMX* UL2* BERT* DeBERTa* Efficient finetuning of Llama 3 with FSDP QDoRA* xLSTMTranscriptAlessio [00:00:00]: Hey everyone, welcome to the Latent Space podcast. This is Alessio, partner and CTO-in-Residence at Decibel Partners, and I'm joined by my co-host Swyx, founder of Smol AI.Swyx [00:00:14]: And today we're back with Jeremy Howard, I think your third appearance on Latent Space. Welcome.Jeremy [00:00:19]: Wait, third? Second?Swyx [00:00:21]: Well, I grabbed you at NeurIPS.Jeremy [00:00:23]: I see.Swyx [00:00:24]: Very fun, standing outside street episode.Jeremy [00:00:27]: I never heard that, by the way. You've got to send me a link. I've got to hear what it sounded like.Swyx [00:00:30]: Yeah. Yeah, it's a NeurIPS podcast.Alessio [00:00:32]: I think the two episodes are six hours, so there's plenty to listen, we'll make sure to send it over.Swyx [00:00:37]: Yeah, we're trying this thing where at the major ML conferences, we, you know, do a little audio tour of, give people a sense of what it's like. But the last time you were on, you declared the end of fine tuning. I hope that I sort of editorialized the title a little bit, and I know you were slightly uncomfortable with it, but you just own it anyway. I think you're very good at the hot takes. And we were just discussing in our pre-show that it's really happening, that the continued pre-training is really happening.Jeremy [00:01:02]: Yeah, absolutely. I think people are starting to understand that treating the three ULM FIT steps of like pre-training, you know, and then the kind of like what people now call instruction tuning, and then, I don't know if we've got a general term for this, DPO, RLHFE step, you know, or the task training, they're not actually as separate as we originally suggested they were in our paper, and when you treat it more as a continuum, and that you make sure that you have, you know, more of kind of the original data set incorporated into the later stages, and that, you know, we've also seen with LLAMA3, this idea that those later stages can be done for a lot longer. These are all of the things I was kind of trying to describe there. It wasn't the end of fine tuning, but more that we should treat it as a continuum, and we should have much higher expectations of how much you can do with an already trained model. You can really add a lot of behavior to it, you can change its behavior, you can do a lot. So a lot of our research has been around trying to figure out how to modify the model by a larger amount rather than starting from random weights, because I get very offended at the idea of starting from random weights.Swyx [00:02:14]: Yeah, I saw that in ICLR in Vienna, there was an outstanding paper about starting transformers from data-driven piers. I don't know if you saw that one, they called it sort of never trained from scratch, and I think it was kind of rebelling against like the sort of random initialization.Jeremy [00:02:28]: Yeah, I've, you know, that's been our kind of continuous message since we started Fast AI, is if you're training for random weights, you better have a really good reason, you know, because it seems so unlikely to me that nobody has ever trained on data that has any similarity whatsoever to the general class of data you're working with, and that's the only situation in which I think starting from random weights makes sense.Swyx [00:02:51]: The other trends since our last pod that I would point people to is I'm seeing a rise in multi-phase pre-training. So Snowflake released a large model called Snowflake Arctic, where they detailed three phases of training where they had like a different mixture of like, there was like 75% web in the first instance, and then they reduced the percentage of the web text by 10% each time and increased the amount of code in each phase. And I feel like multi-phase is being called out in papers more. I feel like it's always been a thing, like changing data mix is not something new, but calling it a distinct phase is new, and I wonder if there's something that you're seeingJeremy [00:03:32]: on your end. Well, so they're getting there, right? So the point at which they're doing proper continued pre-training is the point at which that becomes a continuum rather than a phase. So the only difference with what I was describing last time is to say like, oh, there's a function or whatever, which is happening every batch. It's not a huge difference. You know, I always used to get offended when people had learning rates that like jumped. And so one of the things I started doing early on in Fast.ai was to say to people like, no, you should actually have your learning rate schedule should be a function, not a list of numbers. So now I'm trying to give the same idea about training mix.Swyx [00:04:07]: There's been pretty public work from Meta on schedule-free optimizers. I don't know if you've been following Aaron DeFazio and what he's doing, just because you mentioned learning rate schedules, you know, what if you didn't have a schedule?Jeremy [00:04:18]: I don't care very much, honestly. I don't think that schedule-free optimizer is that exciting. It's fine. We've had non-scheduled optimizers for ages, like Less Wright, who's now at Meta, who was part of the Fast.ai community there, created something called the Ranger optimizer. I actually like having more hyperparameters. You know, as soon as you say schedule-free, then like, well, now I don't get to choose. And there isn't really a mathematically correct way of, like, I actually try to schedule more parameters rather than less. So like, I like scheduling my epsilon in my atom, for example. I schedule all the things. But then the other thing we always did with the Fast.ai library was make it so you don't have to set any schedules. So Fast.ai always supported, like, you didn't even have to pass a learning rate. Like, it would always just try to have good defaults and do the right thing. But to me, I like to have more parameters I can play with if I want to, but you don't have to.Alessio [00:05:08]: And then the more less technical side, I guess, of your issue, I guess, with the market was some of the large research labs taking all this innovation kind of behind closed doors and whether or not that's good, which it isn't. And now we could maybe make it more available to people. And then a month after we released the episode, there was the whole Sam Altman drama and like all the OpenAI governance issues. And maybe people started to think more, okay, what happens if some of these kind of labs, you know, start to break from within, so to speak? And the alignment of the humans is probably going to fall before the alignment of the models. So I'm curious, like, if you have any new thoughts and maybe we can also tie in some of the way that we've been building Answer as like a public benefit corp and some of those aspects.Jeremy [00:05:51]: Sure. So, yeah, I mean, it was kind of uncomfortable because two days before Altman got fired, I did a small public video interview in which I said, I'm quite sure that OpenAI's current governance structure can't continue and that it was definitely going to fall apart. And then it fell apart two days later and a bunch of people were like, what did you know, Jeremy?Alessio [00:06:13]: What did Jeremy see?Jeremy [00:06:15]: I didn't see anything. It's just obviously true. Yeah. So my friend Eric Ries and I spoke a lot before that about, you know, Eric's, I think probably most people would agree, the top expert in the world on startup and AI governance. And you know, we could both clearly see that this didn't make sense to have like a so-called non-profit where then there are people working at a company, a commercial company that's owned by or controlled nominally by the non-profit, where the people in the company are being given the equivalent of stock options, like everybody there was working there with expecting to make money largely from their equity. So the idea that then a board could exercise control by saying like, oh, we're worried about safety issues and so we're going to do something that decreases the profit of the company, when every stakeholder in the company, their remuneration pretty much is tied to their profit, it obviously couldn't work. So I mean, that was a huge oversight there by someone. I guess part of the problem is that the kind of people who work at non-profits and in this case the board, you know, who are kind of academics and, you know, people who are kind of true believers. I think it's hard for them to realize that 99.999% of the world is driven very heavily by money, especially huge amounts of money. So yeah, Eric and I had been talking for a long time before that about what could be done differently, because also companies are sociopathic by design and so the alignment problem as it relates to companies has not been solved. Like, companies become huge, they devour their founders, they devour their communities and they do things where even the CEOs, you know, often of big companies tell me like, I wish our company didn't do that thing. You know, I know that if I didn't do it, then I would just get fired and the board would put in somebody else and the board knows if they don't do it, then their shareholders can sue them because they're not maximizing profitability or whatever. So what Eric's spent a lot of time doing is trying to think about how do we make companies less sociopathic, you know, how to, or more, you know, maybe a better way to think of it is like, how do we make it so that the founders of companies can ensure that their companies continue to actually do the things they want them to do? You know, when we started a company, hey, we very explicitly decided we got to start a company, not a academic lab, not a nonprofit, you know, we created a Delaware Seacorp, you know, the most company kind of company. But when we did so, we told everybody, you know, including our first investors, which was you Alessio. They sound great. We are going to run this company on the basis of maximizing long-term value. And in fact, so when we did our second round, which was an angel round, we had everybody invest through a long-term SPV, which we set up where everybody had to agree to vote in line with long-term value principles. So like never enough just to say to people, okay, we're trying to create long-term value here for society as well as for ourselves and everybody's like, oh, yeah, yeah, I totally agree with that. But when it comes to like, okay, well, here's a specific decision we have to make, which will not maximize short-term value, people suddenly change their mind. So you know, it has to be written into the legal documents of everybody so that no question that that's the way the company has to be managed. So then you mentioned the PBC aspect, Public Benefit Corporation, which I never quite understood previously. And turns out it's incredibly simple, like it took, you know, like one paragraph added to our corporate documents to become a PBC. It was cheap, it was easy, but it's got this huge benefit, which is if you're not a public benefit corporation, then somebody can come along and offer to buy you with a stated description of like turning your company into the thing you most hate, right? And if they offer you more than the market value of your company and you don't accept it, then you are not necessarily meeting the kind of your fiduciary responsibilities. So the way like Eric always described it to me is like, if Philip Morris came along and said that you've got great technology for marketing cigarettes to children, so we're going to pivot your company to do that entirely, and we're going to pay you 50% more than the market value, you're going to have to say yes. If you have a PBC, then you are more than welcome to say no, if that offer is not in line with your stated public benefit. So our stated public benefit is to maximize the benefit to society through using AI. So given that more children smoking doesn't do that, then we can say like, no, we're not selling to you.Alessio [00:11:01]: I was looking back at some of our emails. You sent me an email on November 13th about talking and then on the 14th, I sent you an email working together to free AI was the subject line. And then that was kind of the start of the C round. And then two days later, someone got fired. So you know, you were having these thoughts even before we had like a public example of like why some of the current structures didn't work. So yeah, you were very ahead of the curve, so to speak. You know, people can read your awesome introduction blog and answer and the idea of having a R&D lab versus our lab and then a D lab somewhere else. I think to me, the most interesting thing has been hiring and some of the awesome people that you've been bringing on that maybe don't fit the central casting of Silicon Valley, so to speak. Like sometimes I got it like playing baseball cards, you know, people are like, oh, what teams was this person on, where did they work versus focusing on ability. So I would love for you to give a shout out to some of the awesome folks that you have on the team.Jeremy [00:11:58]: So, you know, there's like a graphic going around describing like the people at XAI, you know, Elon Musk thing. And like they are all connected to like multiple of Stanford, Meta, DeepMind, OpenAI, Berkeley, Oxford. Look, these are all great institutions and they have good people. And I'm definitely not at all against that, but damn, there's so many other people. And one of the things I found really interesting is almost any time I see something which I think like this is really high quality work and it's something I don't think would have been built if that person hadn't built the thing right now, I nearly always reach out to them and ask to chat. And I tend to dig in to find out like, okay, you know, why did you do that thing? Everybody else has done this other thing, your thing's much better, but it's not what other people are working on. And like 80% of the time, I find out the person has a really unusual background. So like often they'll have like, either they like came from poverty and didn't get an opportunity to go to a good school or had dyslexia and, you know, got kicked out of school in year 11, or they had a health issue that meant they couldn't go to university or something happened in their past and they ended up out of the mainstream. And then they kind of succeeded anyway. Those are the people that throughout my career, I've tended to kind of accidentally hire more of, but it's not exactly accidentally. It's like when I see somebody who's done, two people who have done extremely well, one of them did extremely well in exactly the normal way from the background entirely pointing in that direction and they achieved all the hurdles to get there. And like, okay, that's quite impressive, you know, but another person who did just as well, despite lots of constraints and doing things in really unusual ways and came up with different approaches. That's normally the person I'm likely to find useful to work with because they're often like risk-takers, they're often creative, they're often extremely tenacious, they're often very open-minded. So that's the kind of folks I tend to find myself hiring. So now at Answer.ai, it's a group of people that are strong enough that nearly every one of them has independently come to me in the past few weeks and told me that they have imposter syndrome and they're not convinced that they're good enough to be here. And I kind of heard it at the point where I was like, okay, I don't think it's possible that all of you are so far behind your peers that you shouldn't get to be here. But I think part of the problem is as an R&D lab, the great developers look at the great researchers and they're like, wow, these big-brained, crazy research people with all their math and s**t, they're too cool for me, oh my God. And then the researchers look at the developers and they're like, oh, they're killing it, making all this stuff with all these people using it and talking on Twitter about how great it is. I think they're both a bit intimidated by each other, you know. And so I have to kind of remind them like, okay, there are lots of things in this world where you suck compared to lots of other people in this company, but also vice versa, you know, for all things. And the reason you came here is because you wanted to learn about those other things from those other people and have an opportunity to like bring them all together into a single unit. You know, it's not reasonable to expect you're going to be better at everything than everybody else. I guess the other part of it is for nearly all of the people in the company, to be honest, they have nearly always been better than everybody else at nearly everything they're doing nearly everywhere they've been. So it's kind of weird to be in this situation now where it's like, gee, I can clearly see that I suck at this thing that I'm meant to be able to do compared to these other people where I'm like the worst in the company at this thing for some things. So I think that's a healthy place to be, you know, as long as you keep reminding each other about that's actually why we're here. And like, it's all a bit of an experiment, like we don't have any managers. We don't have any hierarchy from that point of view. So for example, I'm not a manager, which means I don't get to tell people what to do or how to do it or when to do it. Yeah, it's been a bit of an experiment to see how that would work out. And it's been great. So for instance, Ben Clavier, who you might have come across, he's the author of Ragatouille, he's the author of Rerankers, super strong information retrieval guy. And a few weeks ago, you know, this additional channel appeared on Discord, on our private Discord called Bert24. And these people started appearing, as in our collab sections, we have a collab section for like collaborating with outsiders. And these people started appearing, there are all these names that I recognize, like Bert24, and they're all talking about like the next generation of Bert. And I start following along, it's like, okay, Ben decided that I think, quite rightly, we need a new Bert. Because everybody, like so many people are still using Bert, and it's still the best at so many things, but it actually doesn't take advantage of lots of best practices. And so he just went out and found basically everybody who's created better Berts in the last four or five years, brought them all together, suddenly there's this huge collaboration going on. So yeah, I didn't tell him to do that. He didn't ask my permission to do that. And then, like, Benjamin Warner dived in, and he's like, oh, I created a whole transformers from scratch implementation designed to be maximally hackable. He originally did it largely as a teaching exercise to show other people, but he was like, I could, you know, use that to create a really hackable BERT implementation. In fact, he didn't say that. He said, I just did do that, you know, and I created a repo, and then everybody's like starts using it. They're like, oh my god, this is amazing. I can now implement all these other BERT things. And it's not just answer AI guys there, you know, there's lots of folks, you know, who have like contributed new data set mixes and blah, blah, blah. So, I mean, I can help in the same way that other people can help. So like, then Ben Clavier reached out to me at one point and said, can you help me, like, what have you learned over time about how to manage intimidatingly capable and large groups of people who you're nominally meant to be leading? And so, you know, I like to try to help, but I don't direct. Another great example was Kerem, who, after our FSTP QLORA work, decided quite correctly that it didn't really make sense to use LoRa in today's world. You want to use the normalized version, which is called Dora. Like two or three weeks after we did FSTP QLORA, he just popped up and said, okay, I've just converted the whole thing to Dora, and I've also created these VLLM extensions, and I've got all these benchmarks, and, you know, now I've got training of quantized models with adapters that are as fast as LoRa, and as actually better than, weirdly, fine tuning. Just like, okay, that's great, you know. And yeah, so the things we've done to try to help make these things happen as well is we don't have any required meetings, you know, but we do have a meeting for each pair of major time zones that everybody's invited to, and, you know, people see their colleagues doing stuff that looks really cool and say, like, oh, how can I help, you know, or how can I learn or whatever. So another example is Austin, who, you know, amazing background. He ran AI at Fidelity, he ran AI at Pfizer, he ran browsing and retrieval for Google's DeepMind stuff, created Jemma.cpp, and he's been working on a new system to make it easier to do web GPU programming, because, again, he quite correctly identified, yeah, so I said to him, like, okay, I want to learn about that. Not an area that I have much expertise in, so, you know, he's going to show me what he's working on and teach me a bit about it, and hopefully I can help contribute. I think one of the key things that's happened in all of these is everybody understands what Eric Gilliam, who wrote the second blog post in our series, the R&D historian, describes as a large yard with narrow fences. Everybody has total flexibility to do what they want. We all understand kind of roughly why we're here, you know, we agree with the premises around, like, everything's too expensive, everything's too complicated, people are building too many vanity foundation models rather than taking better advantage of fine-tuning, like, there's this kind of general, like, sense of we're all on the same wavelength about, you know, all the ways in which current research is fucked up, and, you know, all the ways in which we're worried about centralization. We all care a lot about not just research for the point of citations, but research that actually wouldn't have happened otherwise, and actually is going to lead to real-world outcomes. And so, yeah, with this kind of, like, shared vision, people understand, like, you know, so when I say, like, oh, well, you know, tell me, Ben, about BERT 24, what's that about? And he's like, you know, like, oh, well, you know, you can see from an accessibility point of view, or you can see from a kind of a actual practical impact point of view, there's far too much focus on decoder-only models, and, you know, like, BERT's used in all of these different places and industry, and so I can see, like, in terms of our basic principles, what we're trying to achieve, this seems like something important. And so I think that's, like, a really helpful that we have that kind of shared perspective, you know?Alessio [00:21:14]: Yeah. And before we maybe talk about some of the specific research, when you're, like, reaching out to people, interviewing them, what are some of the traits, like, how do these things come out, you know, usually? Is it working on side projects that you, you know, you're already familiar with? Is there anything, like, in the interview process that, like, helps you screen for people that are less pragmatic and more research-driven versus some of these folks that are just gonna do it, you know? They're not waiting for, like, the perfect process.Jeremy [00:21:40]: Everybody who comes through the recruiting is interviewed by everybody in the company. You know, our goal is 12 people, so it's not an unreasonable amount. So the other thing to say is everybody so far who's come into the recruiting pipeline, everybody bar one, has been hired. So which is to say our original curation has been good. And that's actually pretty easy, because nearly everybody who's come in through the recruiting pipeline are people I know pretty well. So Jono Whitaker and I, you know, he worked on the stable diffusion course we did. He's outrageously creative and talented, and he's super, like, enthusiastic tinkerer, just likes making things. Benjamin was one of the strongest parts of the fast.ai community, which is now the alumni. It's, like, hundreds of thousands of people. And you know, again, like, they're not people who a normal interview process would pick up, right? So Benjamin doesn't have any qualifications in math or computer science. Jono was living in Zimbabwe, you know, he was working on, like, helping some African startups, you know, but not FAANG kind of credentials. But yeah, I mean, when you actually see people doing real work and they stand out above, you know, we've got lots of Stanford graduates and open AI people and whatever in our alumni community as well. You know, when you stand out above all of those people anyway, obviously you've got something going for you. You know, Austin, him and I worked together on the masks study we did in the proceeding at the National Academy of Science. You know, we had worked together, and again, that was a group of, like, basically the 18 or 19 top experts in the world on public health and epidemiology and research design and so forth. And Austin, you know, one of the strongest people in that collaboration. So yeah, you know, like, I've been lucky enough to have had opportunities to work with some people who are great and, you know, I'm a very open-minded person, so I kind of am always happy to try working with pretty much anybody and some people stand out. You know, there have been some exceptions, people I haven't previously known, like Ben Clavier, actually, I didn't know before. But you know, with him, you just read his code, and I'm like, oh, that's really well-written code. And like, it's not written exactly the same way as everybody else's code, and it's not written to do exactly the same thing as everybody else's code. So yeah, and then when I chatted to him, it's just like, I don't know, I felt like we'd known each other for years, like we just were on the same wavelength, but I could pretty much tell that was going to happen just by reading his code. I think you express a lot in the code you choose to write and how you choose to write it, I guess. You know, or another example, a guy named Vic, who was previously the CEO of DataQuest, and like, in that case, you know, he's created a really successful startup. He won the first, basically, Kaggle NLP competition, which was automatic essay grading. He's got the current state-of-the-art OCR system, Surya. Again, he's just a guy who obviously just builds stuff, you know, he doesn't ask for permission, he doesn't need any, like, external resources. Actually, Karim's another great example of this, I mean, I already knew Karim very well because he was my best ever master's student, but it wasn't a surprise to me then when he then went off to create the world's state-of-the-art language model in Turkish on his own, in his spare time, with no budget, from scratch. This is not fine-tuning or whatever, he, like, went back to Common Crawl and did everything. Yeah, it's kind of, I don't know what I'd describe that process as, but it's not at all based on credentials.Swyx [00:25:17]: Assemble based on talent, yeah. We wanted to dive in a little bit more on, you know, turning from the people side of things into the technical bets that you're making. Just a little bit more on Bert. I was actually, we just did an interview with Yi Tay from Reka, I don't know if you're familiar with his work, but also another encoder-decoder bet, and one of his arguments was actually people kind of over-index on the decoder-only GPT-3 type paradigm. I wonder if you have thoughts there that is maybe non-consensus as well. Yeah, no, absolutely.Jeremy [00:25:45]: So I think it's a great example. So one of the people we're collaborating with a little bit with BERT24 is Colin Raffle, who is the guy behind, yeah, most of that stuff, you know, between that and UL2, there's a lot of really interesting work. And so one of the things I've been encouraging the BERT group to do, Colin has as well, is to consider using a T5 pre-trained encoder backbone as a thing you fine-tune, which I think would be really cool. You know, Colin was also saying actually just use encoder-decoder as your Bert, you know, why don't you like use that as a baseline, which I also think is a good idea. Yeah, look.Swyx [00:26:25]: What technical arguments are people under-weighting?Jeremy [00:26:27]: I mean, Colin would be able to describe this much better than I can, but I'll give my slightly non-expert attempt. Look, I mean, think about like diffusion models, right? Like in stable diffusion, like we use things like UNet. You have this kind of downward path and then in the upward path you have the cross connections, which it's not a tension, but it's like a similar idea, right? You're inputting the original encoding path into your decoding path. It's critical to make it work, right? Because otherwise in the decoding part, the model has to do so much kind of from scratch. So like if you're doing translation, like that's a classic kind of encoder-decoder example. If it's decoder only, you never get the opportunity to find the right, you know, feature engineering, the right feature encoding for the original sentence. And it kind of means then on every token that you generate, you have to recreate the whole thing, you know? So if you have an encoder, it's basically saying like, okay, this is your opportunity model to create a really useful feature representation for your input information. So I think there's really strong arguments for encoder-decoder models anywhere that there is this kind of like context or source thing. And then why encoder only? Well, because so much of the time what we actually care about is a classification, you know? It's like an output. It's like generating an arbitrary length sequence of tokens. So anytime you're not generating an arbitrary length sequence of tokens, decoder models don't seem to make much sense. Now the interesting thing is, you see on like Kaggle competitions, that decoder models still are at least competitive with things like Deberta v3. They have to be way bigger to be competitive with things like Deberta v3. And the only reason they are competitive is because people have put a lot more time and money and effort into training the decoder only ones, you know? There isn't a recent Deberta. There isn't a recent Bert. Yeah, it's a whole part of the world that people have slept on a little bit. And this is just what happens. This is how trends happen rather than like, to me, everybody should be like, oh, let's look at the thing that has shown signs of being useful in the past, but nobody really followed up with properly. That's the more interesting path, you know, where people tend to be like, oh, I need to get citations. So what's everybody else doing? Can I make it 0.1% better, you know, or 0.1% faster? That's what everybody tends to do. Yeah. So I think it's like, Itay's work commercially now is interesting because here's like a whole, here's a whole model that's been trained in a different way. So there's probably a whole lot of tasks it's probably better at than GPT and Gemini and Claude. So that should be a good commercial opportunity for them if they can figure out what those tasks are.Swyx [00:29:07]: Well, if rumors are to be believed, and he didn't comment on this, but, you know, Snowflake may figure out the commercialization for them. So we'll see.Jeremy [00:29:14]: Good.Alessio [00:29:16]: Let's talk about FSDP, Qlora, Qdora, and all of that awesome stuff. One of the things we talked about last time, some of these models are meant to run on systems that nobody can really own, no single person. And then you were like, well, what if you could fine tune a 70B model on like a 4090? And I was like, no, that sounds great, Jeremy, but like, can we actually do it? And then obviously you all figured it out. Can you maybe tell us some of the worst stories behind that, like the idea behind FSDP, which is kind of taking sharded data, parallel computation, and then Qlora, which is do not touch all the weights, just go quantize some of the model, and then within the quantized model only do certain layers instead of doing everything.Jeremy [00:29:57]: Well, do the adapters. Yeah.Alessio [00:29:59]: Yeah. Yeah. Do the adapters. Yeah. I will leave the floor to you. I think before you published it, nobody thought this was like a short term thing that we're just going to have. And now it's like, oh, obviously you can do it, but it's not that easy.Jeremy [00:30:12]: Yeah. I mean, to be honest, it was extremely unpleasant work to do. It's like not at all enjoyable. I kind of did version 0.1 of it myself before we had launched the company, or at least the kind of like the pieces. They're all pieces that are difficult to work with, right? So for the quantization, you know, I chatted to Tim Detmers quite a bit and, you know, he very much encouraged me by saying like, yeah, it's possible. He actually thought it'd be easy. It probably would be easy for him, but I'm not Tim Detmers. And, you know, so he wrote bits and bytes, which is his quantization library. You know, he wrote that for a paper. He didn't write that to be production like code. It's now like everybody's using it, at least the CUDA bits. So like, it's not particularly well structured. There's lots of code paths that never get used. There's multiple versions of the same thing. You have to try to figure it out. So trying to get my head around that was hard. And you know, because the interesting bits are all written in CUDA, it's hard to like to step through it and see what's happening. And then, you know, FSTP is this very complicated library and PyTorch, which not particularly well documented. So the only really, really way to understand it properly is again, just read the code and step through the code. And then like bits and bytes doesn't really work in practice unless it's used with PEF, the HuggingFace library and PEF doesn't really work in practice unless you use it with other things. And there's a lot of coupling in the HuggingFace ecosystem where like none of it works separately. You have to use it all together, which I don't love. So yeah, trying to just get a minimal example that I can play with was really hard. And so I ended up having to rewrite a lot of it myself to kind of create this like minimal script. One thing that helped a lot was Medec had this LlamaRecipes repo that came out just a little bit before I started working on that. And like they had a kind of role model example of like, here's how to train FSTP, LoRa, didn't work with QLoRa on Llama. A lot of the stuff I discovered, the interesting stuff would be put together by Les Wright, who's, he was actually the guy in the Fast.ai community I mentioned who created the Ranger Optimizer. So he's doing a lot of great stuff at Meta now. So yeah, I kind of, that helped get some minimum stuff going and then it was great once Benjamin and Jono joined full time. And so we basically hacked at that together and then Kerim joined like a month later or something. And it was like, gee, it was just a lot of like fiddly detailed engineering on like barely documented bits of obscure internals. So my focus was to see if it kind of could work and I kind of got a bit of a proof of concept working and then the rest of the guys actually did all the work to make it work properly. And, you know, every time we thought we had something, you know, we needed to have good benchmarks, right? So we'd like, it's very easy to convince yourself you've done the work when you haven't, you know, so then we'd actually try lots of things and be like, oh, and these like really important cases, the memory use is higher, you know, or it's actually slower. And we'd go in and we just find like all these things that were nothing to do with our library that just didn't work properly. And nobody had noticed they hadn't worked properly because nobody had really benchmarked it properly. So we ended up, you know, trying to fix a whole lot of different things. And even as we did so, new regressions were appearing in like transformers and stuff that Benjamin then had to go away and figure out like, oh, how come flash attention doesn't work in this version of transformers anymore with this set of models and like, oh, it turns out they accidentally changed this thing, so it doesn't work. You know, there's just, there's not a lot of really good performance type evals going on in the open source ecosystem. So there's an extraordinary amount of like things where people say like, oh, we built this thing and it has this result. And when you actually check it, so yeah, there's a shitload of war stories from getting that thing to work. And it did require a particularly like tenacious group of people and a group of people who don't mind doing a whole lot of kind of like really janitorial work, to be honest, to get the details right, to check them. Yeah.Alessio [00:34:09]: We had a trade out on the podcast and we talked about how a lot of it is like systems work to make some of these things work. It's not just like beautiful, pure math that you do on a blackboard. It's like, how do you get into the nitty gritty?Jeremy [00:34:22]: I mean, flash attention is a great example of that. Like it's, it basically is just like, oh, let's just take the attention and just do the tiled version of it, which sounds simple enough, you know, but then implementing that is challenging at lots of levels.Alessio [00:34:36]: Yeah. What about inference? You know, obviously you've done all this amazing work on fine tuning. Do you have any research you've been doing on the inference side, how to make local inference really fast on these models too?Jeremy [00:34:47]: We're doing quite a bit on that at the moment. We haven't released too much there yet. But one of the things I've been trying to do is also just to help other people. And one of the nice things that's happened is that a couple of folks at Meta, including Mark Seraphim, have done a nice job of creating this CUDA mode community of people working on like CUDA kernels or learning about that. And I tried to help get that going well as well and did some lessons to help people get into it. So there's a lot going on in both inference and fine tuning performance. And a lot of it's actually happening kind of related to that. So PyTorch team have created this Torch AO project on quantization. And so there's a big overlap now between kind of the FastAI and AnswerAI and CUDA mode communities of people working on stuff for both inference and fine tuning. But we're getting close now. You know, our goal is that nobody should be merging models, nobody should be downloading merged models, everybody should be using basically quantized plus adapters for almost everything and just downloading the adapters. And that should be much faster. So that's kind of the place we're trying to get to. It's difficult, you know, because like Karim's been doing a lot of work with VLM, for example. These inference engines are pretty complex bits of code. They have a whole lot of custom kernel stuff going on as well, as do the quantization libraries. So we've been working on, we're also quite a bit of collaborating with the folks who do HQQ, which is a really great quantization library and works super well. So yeah, there's a lot of other people outside AnswerAI that we're working with a lot who are really helping on all this performance optimization stuff, open source.Swyx [00:36:27]: Just to follow up on merging models, I picked up there that you said nobody should be merging models. That's interesting because obviously a lot of people are experimenting with this and finding interesting results. I would say in defense of merging models, you can do it without data. That's probably the only thing that's going for it.Jeremy [00:36:45]: To explain, it's not that you shouldn't merge models. You shouldn't be distributing a merged model. You should distribute a merged adapter 99% of the time. And actually often one of the best things happening in the model merging world is actually that often merging adapters works better anyway. The point is, Sean, that once you've got your new model, if you distribute it as an adapter that sits on top of a quantized model that somebody's already downloaded, then it's a much smaller download for them. And also the inference should be much faster because you're not having to transfer FB16 weights from HPM memory at all or ever load them off disk. You know, all the main weights are quantized and the only floating point weights are in the adapters. So that should make both inference and fine tuning faster. Okay, perfect.Swyx [00:37:33]: We're moving on a little bit to the rest of the fast universe. I would have thought that, you know, once you started Answer.ai, that the sort of fast universe would be kind of on hold. And then today you just dropped Fastlight and it looks like, you know, there's more activity going on in sort of Fastland.Jeremy [00:37:49]: Yeah. So Fastland and Answerland are not really distinct things. Answerland is kind of like the Fastland grown up and funded. They both have the same mission, which is to maximize the societal benefit of AI broadly. We want to create thousands of commercially successful products at Answer.ai. And we want to do that with like 12 people. So that means we need a pretty efficient stack, you know, like quite a few orders of magnitude more efficient, not just for creation, but for deployment and maintenance than anything that currently exists. People often forget about the D part of our R&D firm. So we've got to be extremely good at creating, deploying and maintaining applications, not just models. Much to my horror, the story around creating web applications is much worse now than it was 10 or 15 years ago in terms of, if I say to a data scientist, here's how to create and deploy a web application, you know, either you have to learn JavaScript or TypeScript and about all the complex libraries like React and stuff, and all the complex like details around security and web protocol stuff around how you then talk to a backend and then all the details about creating the backend. You know, if that's your job and, you know, you have specialists who work in just one of those areas, it is possible for that to all work. But compared to like, oh, write a PHP script and put it in the home directory that you get when you sign up to this shell provider, which is what it was like in the nineties, you know, here are those 25 lines of code and you're done and now you can pass that URL around to all your friends, or put this, you know, .pl file inside the CGI bin directory that you got when you signed up to this web host. So yeah, the thing I've been mainly working on the last few weeks is fixing all that. And I think I fixed it. I don't know if this is an announcement, but I tell you guys, so yeah, there's this thing called fastHTML, which basically lets you create a complete web application in a single Python file. Unlike excellent projects like Streamlit and Gradio, you're not working on top of a highly abstracted thing. That's got nothing to do with web foundations. You're working with web foundations directly, but you're able to do it by using pure Python. There's no template, there's no ginger, there's no separate like CSS and JavaScript files. It looks and behaves like a modern SPA web application. And you can create components for like daisy UI, or bootstrap, or shoelace, or whatever fancy JavaScript and or CSS tailwind etc library you like, but you can write it all in Python. You can pip install somebody else's set of components and use them entirely from Python. You can develop and prototype it all in a Jupyter notebook if you want to. It all displays correctly, so you can like interactively do that. And then you mentioned Fastlight, so specifically now if you're using SQLite in particular, it's like ridiculously easy to have that persistence, and all of your handlers will be passed database ready objects automatically, that you can just call dot delete dot update dot insert on. Yeah, you get session, you get security, you get all that. So again, like with most everything I do, it's very little code. It's mainly tying together really cool stuff that other people have written. You don't have to use it, but a lot of the best stuff comes from its incorporation of HTMX, which to me is basically the thing that changes your browser to make it work the way it always should have. So it just does four small things, but those four small things are the things that are basically unnecessary constraints that HTML should never have had, so it removes the constraints. It sits on top of Starlet, which is a very nice kind of lower level platform for building these kind of web applications. The actual interface matches as closely as possible to FastAPI, which is a really nice system for creating the kind of classic JavaScript type applications. And Sebastian, who wrote FastAPI, has been kind enough to help me think through some of these design decisions, and so forth. I mean, everybody involved has been super helpful. Actually, I chatted to Carson, who created HTMX, you know, so about it. Some of the folks involved in Django, like everybody in the community I've spoken to definitely realizes there's a big gap to be filled around, like, highly scalable, web foundation-based, pure Python framework with a minimum of fuss. So yeah, I'm getting a lot of support and trying to make sure that FastHTML works well for people.Swyx [00:42:38]: I would say, when I heard about this, I texted Alexio. I think this is going to be pretty huge. People consider Streamlit and Gradio to be the state of the art, but I think there's so much to improve, and having what you call web foundations and web fundamentals at the core of it, I think, would be really helpful.Jeremy [00:42:54]: I mean, it's based on 25 years of thinking and work for me. So like, FastML was built on a system much like this one, but that was of hell. And so I spent, you know, 10 years working on that. We had millions of people using that every day, really pushing it hard. And I really always enjoyed working in that. Yeah. So, you know, and obviously lots of other people have done like great stuff, and particularly HTMX. So I've been thinking about like, yeah, how do I pull together the best of the web framework I created for FastML with HTMX? There's also things like PicoCSS, which is the CSS system, which by default, FastHTML comes with. Although, as I say, you can pip install anything you want to, but it makes it like super easy to, you know, so we try to make it so that just out of the box, you don't have any choices to make. Yeah. You can make choices, but for most people, you just, you know, it's like the PHP in your home directory thing. You just start typing and just by default, you'll get something which looks and feels, you know, pretty okay. And if you want to then write a version of Gradio or Streamlit on top of that, you totally can. And then the nice thing is if you then write it in kind of the Gradio equivalent, which will be, you know, I imagine we'll create some kind of pip installable thing for that. Once you've outgrown, or if you outgrow that, it's not like, okay, throw that all away and start again. And this like whole separate language that it's like this kind of smooth, gentle path that you can take step-by-step because it's all just standard web foundations all the way, you know.Swyx [00:44:29]: Just to wrap up the sort of open source work that you're doing, you're aiming to create thousands of projects with a very, very small team. I haven't heard you mention once AI agents or AI developer tooling or AI code maintenance. I know you're very productive, but you know, what is the role of AI in your own work?Jeremy [00:44:47]: So I'm making something. I'm not sure how much I want to say just yet.Swyx [00:44:52]: Give us a nibble.Jeremy [00:44:53]: All right. I'll give you the key thing. So I've created a new approach. It's not called prompt engineering. It's called dialogue engineering. But I'm creating a system for doing dialogue engineering. It's currently called AI magic. I'm doing most of my work in this system and it's making me much more productive than I was before I used it. So I always just build stuff for myself and hope that it'll be useful for somebody else. Think about chat GPT with code interpreter, right? The basic UX is the same as a 1970s teletype, right? So if you wrote APL on a teletype in the 1970s, you typed onto a thing, your words appeared at the bottom of a sheet of paper and you'd like hit enter and it would scroll up. And then the answer from APL would be printed out, scroll up, and then you would type the next thing. And like, which is also the way, for example, a shell works like bash or ZSH or whatever. It's not terrible, you know, like we all get a lot done in these like very, very basic teletype style REPL environments, but I've never felt like it's optimal and everybody else has just copied chat GPT. So it's also the way BART and Gemini work. It's also the way the Claude web app works. And then you add code interpreter. And the most you can do is to like plead with chat GPT to write the kind of code I want. It's pretty good for very, very, very beginner users who like can't code at all, like by default now the code's even hidden away, so you never even have to see it ever happened. But for somebody who's like wanting to learn to code or who already knows a bit of code or whatever, it's, it seems really not ideal. So okay, that's one end of the spectrum. The other end of the spectrum, which is where Sean's work comes in, is, oh, you want to do more than chat GPT? No worries. Here is Visual Studio Code. I run it. There's an empty screen with a flashing cursor. Okay, start coding, you know, and it's like, okay, you can use systems like Sean's or like cursor or whatever to be like, okay, Apple K in cursors, like a creative form that blah, blah, blah. But in the end, it's like a convenience over the top of this incredibly complicated system that full-time sophisticated software engineers have designed over the past few decades in a totally different environment as a way to build software, you know. And so we're trying to like shoehorn in AI into that. And it's not easy to do. And I think there are like much better ways of thinking about the craft of software development in a language model world to be much more interactive, you know. So the thing that I'm building is neither of those things. It's something between the two. And it's built around this idea of crafting a dialogue, you know, where the outcome of the dialogue is the artifacts that you want, whether it be a piece of analysis or whether it be a Python library or whether it be a technical blog post or whatever. So as part of building that, I've created something called Claudette, which is a library for Claude. I've created something called Cosette, which is a library for OpenAI. They're libraries which are designed to make those APIs much more usable, much easier to use, much more concise. And then I've written AI magic on top of those. And that's been an interesting exercise because I did Claudette first, and I was looking at what Simon Willison did with his fantastic LLM library. And his library is designed around like, let's make something that supports all the LLM inference engines and commercial providers. I thought, okay, what if I did something different, which is like make something that's as Claude friendly as possible and forget everything else. So that's what Claudette was. So for example, one of the really nice things in Claude is prefill. So by telling the assistant that this is what your response started with, there's a lot of powerful things you can take advantage of. So yeah, I created Claudette to be as Claude friendly as possible. And then after I did that, and then particularly with GPT 4.0 coming out, I kind of thought, okay, now let's create something that's as OpenAI friendly as possible. And then I tried to look to see, well, where are the similarities and where are the differences? And now can I make them compatible in places where it makes sense for them to be compatible without losing out on the things that make each one special for what they are. So yeah, those are some of the things I've been working on in that space. And I'm thinking we might launch AI magic via a course called how to solve it with code. The name is based on the classic Polya book, if you know how to solve it, which is, you know, one of the classic math books of all time, where we're basically going to try to show people how to solve challenging problems that they didn't think they could solve without doing a full computer science course, by taking advantage of a bit of AI and a bit of like practical skills, as particularly for this like whole generation of people who are learning to code with and because of ChatGPT. Like I love it, I know a lot of people who didn't really know how to code, but they've created things because they use ChatGPT, but they don't really know how to maintain them or fix them or add things to them that ChatGPT can't do, because they don't really know how to code. And so this course will be designed to show you how you can like either become a developer who can like supercharge their capabilities by using language models, or become a language model first developer who can supercharge their capabilities by understanding a bit about process and fundamentals.Alessio [00:50:19]: Nice. That's a great spoiler. You know, I guess the fourth time you're going to be on learning space, we're going to talk about AI magic. Jeremy, before we wrap, this was just a great run through everything. What are the things that when you next come on the podcast in nine, 12 months, we're going to be like, man, Jeremy was like really ahead of it. Like, is there anything that you see in the space that maybe people are not talking enough? You know, what's the next company that's going to fall, like have drama internally, anything in your mind?Jeremy [00:50:47]: You know, hopefully we'll be talking a lot about fast HTML and hopefully the international community that at that point has come up around that. And also about AI magic and about dialogue engineering. Hopefully dialogue engineering catches on because I think it's the right way to think about a lot of this stuff. What else? Just trying to think about all on the research side. Yeah. I think, you know, I mean, we've talked about a lot of it. Like I think encoder decoder architectures, encoder only architectures, hopefully we'll be talking about like the whole re-interest in BERT that BERT 24 stimulated.Swyx [00:51:17]: There's a safe space model that came out today that might be interesting for this general discussion. One thing that stood out to me with Cartesia's blog posts was that they were talking about real time ingestion, billions and trillions of tokens, and keeping that context, obviously in the state space that they have.Jeremy [00:51:34]: Yeah.Swyx [00:51:35]: I'm wondering what your thoughts are because you've been entirely transformers the whole time.Jeremy [00:51:38]: Yeah. No. So obviously my background is RNNs and LSTMs. Of course. And I'm still a believer in the idea that state is something you can update, you know? So obviously Sepp Hochreiter came up, came out with xLSTM recently. Oh my God. Okay. Another whole thing we haven't talked about, just somewhat related. I've been going crazy for like a long time about like, why can I not pay anybody to save my KV cash? I just ingested the Great Gatsby or the documentation for Starlet or whatever, you know, I'm sending it as my prompt context. Why are you redoing it every time? So Gemini is about to finally come out with KV caching, and this is something that Austin actually in Gemma.cpp had had on his roadmap for years, well not years, months, long time. The idea that the KV cache is like a thing that, it's a third thing, right? So there's RAG, you know, there's in-context learning, you know, and prompt engineering, and there's KV cache creation. I think it creates like a whole new class almost of applications or as techniques where, you know, for me, for example, I very often work with really new libraries or I've created my own library that I'm now writing with rather than on. So I want all the docs in my new library to be there all the time. So I want to upload them once, and then we have a whole discussion about building this application using FastHTML. Well nobody's got FastHTML in their language model yet, I don't want to send all the FastHTML docs across every time. So one of the things I'm looking at doing in AI Magic actually is taking advantage of some of these ideas so that you can have the documentation of the libraries you're working on be kind of always available. Something over the next 12 months people will be spending time thinking about is how to like, where to use RAG, where to use fine-tuning, where to use KV cache storage, you know. And how to use state, because in state models and XLSTM, again, state is something you update. So how do we combine the best of all of these worlds?Alessio [00:53:46]: And Jeremy, I know before you talked about how some of the autoregressive models are not maybe a great fit for agents. Any other thoughts on like JEPA, diffusion for text, any interesting thing that you've seen pop up?Jeremy [00:53:58]: In the same way that we probably ought to have state that you can update, i.e. XLSTM and state models, in the same way that a lot of things probably should have an encoder, JEPA and diffusion both seem like the right conceptual mapping for a lot of things we probably want to do. So the idea of like, there should be a piece of the generative pipeline, which is like thinking about the answer and coming up with a sketch of what the answer looks like before you start outputting tokens. That's where it kind of feels like diffusion ought to fit, you know. And diffusion is, because it's not autoregressive, it's like, let's try to like gradually de-blur the picture of how to solve this. So this is also where dialogue engineering fits in, by the way. So with dialogue engineering, one of the reasons it's working so well for me is I use it to kind of like craft the thought process before I generate the code, you know. So yeah, there's a lot of different pieces here and I don't know how they'll all kind of exactly fit together. I don't know if JEPA is going to actually end up working in the text world. I don't know if diffusion will end up working in the text world, but they seem to be like trying to solve a class of problem which is currently unsolved.Alessio [00:55:13]: Awesome, Jeremy. This was great, as usual. Thanks again for coming back on the pod and thank you all for listening. Yeah, that was fantastic. Get full access to Latent Space at www.latent.space/subscribe

Join Brian and Marisa Smith as they dive into the world of developer advocacy, the challenges of agile methodologies in data engineering, and the vital role of open-source communities. Discover how to better support and communicate with your developers in this insightful episode! Overview In this episode, Brian Milner interviews developer relations expert Marisa Smith to explore the vital role of developer advocates in bridging the gap between companies and their users. Marisa shares her insights on the challenges of communicating with developers, emphasizing the need to create a welcoming environment for questions and feedback. She also discusses the unique difficulties developers face when implementing agile methodologies, particularly in the realm of data engineering. They highlight the significance of open-source communities in fostering innovation and collaboration and provide a preview of Marisa's upcoming talk at Agile 2024 on enhancing data pipelines with SQLMesh. Listen Now to Discover: [1:08] - Join Brian in an engaging conversation with Dr. Marisa Smith, PhD, Developer Relations Expert, Developer Advocate, and Speaker. [2:43] - Marisa Smith sheds light on the crucial role of a developer advocate, explaining how they bridge the gap between developers and the wider community. [3:49] - Brian digs into common mistakes in how we communicate with developers and poses the question: what are we getting wrong in our interactions? [5:57] - Marisa outlines the hurdles developers face in a Scrum team environment, shedding light on common obstacles. [12:00] - Marisa explores the hurdles in developer communication, offering insights into improving dialogue and understanding. [12:55] - Mountain Goat Software offers Working on a Scrum Team, a private class to help Scrum teams foster a team dynamic that supports the whole team, including bridging the gap in communicating with developer teams. [15:00] - Marisa discusses how SQLMesh has empowered data engineers to streamline their tasks, sparking a sense of 'Marie Kondoing' their work. [24:11] - Marisa emphasizes the vital importance of open-source developer communities for fostering innovation and teamwork. [26:51] - Brian shares a big thank you to Marisa for joining him on the show. [27:50] - We invite you to subscribe to the Agile Mentors Podcast. Do you have feedback or a great idea for an episode of the show? Great! Just send us an email. [27:54] - If you’d like to continue this discussion, join the Agile Mentors Community. You get a year of free membership into that site by taking any class with Mountain Goat Software, such as CSM or CSPO. We'd love to see you in one of Mountain Goat Software's classes, you can find the schedule here. References and resources mentioned in the show: Dr. Marisa Smith, PhD Join the SQLMesh Community Agile 2024 SQLMesh Working on a Scrum Team Subscribe to the Agile Mentors Podcast Mountain Goat Software’s Private Training Certified ScrumMaster® Training and Scrum Certification Certified Scrum Product Owner® Training Mountain Goat Software Certified Scrum and Agile Training Schedule Join the Agile Mentors Community Want to get involved? This show is designed for you, and we’d love your input. Enjoyed what you heard today? Please leave a rating and a review. It really helps, and we read every single one. Got an Agile subject you’d like us to discuss or a question that needs an answer? Share your thoughts with us at podcast@mountaingoatsoftware.com This episode’s presenters are: Brian Milner is SVP of coaching and training at Mountain Goat Software. He's passionate about making a difference in people's day-to-day work, influenced by his own experience of transitioning to Scrum and seeing improvements in work/life balance, honesty, respect, and the quality of work. Marisa Smith is a Developer Relations expert who bridges the gap between the community and development teams, addressing problems and promoting open-source software. With a Ph.D. in Computational & Theoretical Physical Chemistry, she has a background in simulating radiation effects in water. Auto-generated Transcript: Brian (00:00) Welcome in Agile Mentors. We're here for another episode of the Agile Mentors podcast. I'm with you as always, Brian Milner. And today I have the one, the only Marisa Smith with us. Welcome in Marisa. Marisa (00:13) Hi, thank you so much for having me. Brian (00:15) Very excited to have Marisa with us. If you're not familiar with Marisa, her title is Developer Relations Expert. So right there, that's an episode, right? We could talk just about that. And we'll get into that a little bit more, but there's a lot of really interesting stuff here about Marisa. She has her PhD in theoretical and computational physical chemistry. So... Marisa (00:41) Yeah. Brian (00:42) Again, wow, right? I mean, this is amazing stuff. She's worked at Streamlet. She was their very first developer advocate there. And she has since, Streamlet's been acquired by Snowflake. And you founded Tobacco Data, is that right? Marisa (01:07) Uh, no, I, um, I am their first developer advocate at Tupiqium data. Yeah. No words. Brian (01:11) OK, gotcha. Sorry about that. Messed that up. So very, very interesting background. And one of the things that caught our notice, Marisa spoke last year at Agile 2023 and is speaking again this year at Agile 2024. So again, if you're going to come out, I highly recommend you attend her talk. Her talk is called Marie Kondo. your data pipelines with SQLMesh, which I think is really, really interesting. But I'm talking too much, and I want to turn it over to Marisa here. Help us understand developer relations expert and developer advocate. What does that mean? Marisa (01:59) Yeah, so I am, what I always say is that I am the person that connects your company to the people who use your product. And it just so happens that the companies that I work for are companies that work in the tech industry. They're building some sort of piece of the tech stack. So the people that use it, their customers are other developer, developers essentially, or technical people. Brian (02:22) Yeah, so you're an expert in the... Marisa (02:27) in the art of, in the art of like, how do we communicate with other developers? How do we pass that information back and forth between the developers that are making a product and the developers that use a product. And how do we make sure that, you know, we're getting, we're, we're getting the best out of our, out of our users and that they're getting the best out of the technology that we're trying to build for them. Brian (02:49) That is so, so interesting. And so I'm sure product owners are listening going, yeah, help me. Help me. I want to understand. How do I talk to developers? So gosh, there's so many directions we can go with this. What do you think people misunderstand most when they try to communicate with developers? What do we get wrong? Marisa (02:55) hahahaha Oh, wow, that's a great question. Let me think about this for a second. I think, I think from, from, from my perspective, as somebody who spends a lot of time, like running different communities, especially open source communities, I think that people get the wrong idea in that. Yes, these developers are your customer, but a lot of the time they have very limited time. They come in, they look at, you know, maybe your open source product or, uh, you know, your free version or something that they're trying to see if they can integrate this in their own stack. And I think people can. or companies can come at them a little bit too quickly, a little bit too salesy, right? And then that ends up driving them away. They're like, no, no, no, I'm really not interested in any of this. I'm just trying to figure out if this is the right technology. A lot of developers like to iterate. They try things out, right? And so I think if you come at them too early with, oh, here's our sales process. Here's this, this is how much this costs. It's like, no, no, no, I'm like... way early, I'm MVP POC type of thing, trying to see if I can understand, or if this works with my team, my workflow, my current pipeline, the other technologies that we use, you know, that, that type of thing. I think that's one of the biggest mistakes that you can make, especially when you're talking about open source, which is kind of like my bread and butter. Brian (04:28) Right, right. Yeah. And you know, the communication, especially, I mean, because we talk about Scrum and Agile here on the podcast, you know, that relationship between the business side of the house and the developer side of the house, it's almost like, you know, Romeo and Juliet and the two houses, you know, they speak different languages. They want different things. They see the world in a very different way. Marisa (04:34) Mm -hmm. Mm -hmm. Brian (04:58) And yet somehow these two groups have to figure out how are we going to work together to really deliver something that is valuable, right? So you work with a lot of developers. You talk with a lot of developers. And I know there's lots of different kind of practices and things that are out there that developers are using these days. Marisa (05:09) Yes. Yeah. Brian (05:26) When you talk to them about something like Scrum, or when that kind of process comes up, what are some of the chief complaints that you hear from developers when they talk about working on a Scrum team? What are they not like about it? Marisa (05:44) Ooh, interesting. Yeah, I would say. I would say, I mean, in the area that I'm working in right now, I'm working pretty deep in the data pipelines. So Tobico data runs these two open source projects, SQLMesh and SQL glot. And it's essentially your T of your ETL pipeline, right? We're using SQL, we understand SQL and we're transforming your SQL queries into these tables and we're helping you manage these pipelines as they evolve over time. And so, um, and so I think, you know, in this space, what happens is when you're talking about, you know, working with scrum teams and stuff like that, one of the pieces of agile is trying out new things and iterating. And that can actually be super difficult for a lot of like these data engineers and developers to do, because you have accountability at the end of the day, right? If you're changing up your data, you're mutating some, some SQL query that changes some model that changes your data pipeline downstream. And then all of a sudden. you know, you've pushed it to production and, uh -oh, this data is not what you expected. It's not what you had like originally tested for. And that's because, you know, teams have to work across many different, um, they have many different like iterations that they're working on and many different teams are, you know, making changes potentially simultaneously. And so I think that for, for developers, when you talk about scrum and you talk about agile, particularly if you're talking about like, adjusting tooling or trying out something new again, coming back to this fact that like, you know, they're just there to test things out. They have very limited time and that's because they get stressed out. It's like, well, I don't want to break production. We have to protect production, your production environment as much as possible. And, you know, part of agile and part of doing all these things is trying, trying these new tools, trying these new companies, trying these new methods. And you can get a lot of resistance from that because. they know this method works, they know that they have accountability and they know that production will be fine if they use this methodology that they've set up. Sometimes it can be a little bit of a matchstick thing in the back end. Brian (07:47) Yeah. Well, and I mean, just hearing you talk about that and thinking about it, I know one of the big friction points between the business side and the developer side is take SQLMesh. If that's something that my team has never worked with and I have someone on the team who is interested in that and wants to, thinks it might give us some benefits, There's friction there with the product side to say, product has all these things they want, and they want to push these things forward. But I think this bit of adding this to our stack is going to improve things. But how do I communicate that to the business to give us the time to do this? Because it's not directly leading to a feature, but it will improve things moving on. How do you? How do you balance it? How do you have that conversation with a product person or the business side of the house to really say, Hey, this is worthwhile. Marisa (08:50) Yeah. Yeah. Honestly, it's super difficult. And I think it's really a balance of, you know, having to have these engineers dedicate some time to new tooling and testing things out. And then once you have done that in their schedule, you know, I don't know how frequently you may want to do it like once a month or something like that, where they, you know, take a day and just review what new is out there. What else should we be looking at? What other tools, what other... You know, with, especially with the emergence of AI and all of that recently, that changes a mile a minute, I think. So, you know, keeping on top of that is, is, is a huge burden on these engineering teams. And so time needs to be dedicated for actually getting that kind of stuff done and the freedom to actually try these things out and do like just a minimum viable product, right? Just a tiny POC with like a little Tinkit data set. And then on, I think there's some. Brian (09:23) Ha ha. Marisa (09:45) there's some weight of this on the company that's developing these open source products or new tooling, that they have to start communicating and figuring out their business value as well. Because those developers cannot, with this little example, show all of the benefits that you would get. What cost savings do you get? What efficiency savings? What increase in productivity? All of that has to be done by the company and you need to have that ready. so that you can back up your developers that are trying out your product and your open source projects so that they have something to go off of. They're like, hey, look, I made this. It took one week or something like that. And I got it to a place where it's really good and look at all these cool features. And this will make us so much faster in our pipeline. And then here is the company's documentation or case study or whatever it is that says this should increase our productivity approximately this much. And... that these other big companies that are similar to us have this sort of success with it. And, you know, it, I think those two combined can really help alleviate these pressure that the developers feel and give them the time to actually try out new tools, which in many cases they love to do. They love to try new things. They love open source, right. And they will be your best advocates. If you spend the time talking to them, communicating with them and giving them the things that they need to be successful. Brian (11:12) Yeah, I would think that's kind of a, there's a duality to, I would imagine your role in speaking with them about your product, because in one case you have to sell them on, hey, look how beneficial this is. This is, you should add this to your stack. But on the other hand, you've got to equip them with the, the, uh, the sales pitch, I guess, that they can then make to their business to say, hey, you should allow us to do this. You should give us the, whether it's finances to do this or the time or the resources to do this, that, you know, it does benefit you as well. So that's gotta be a really difficult part of that communication is kind of, you know, getting people who are not really salespeople, you know, having been a developer, I know I kind of get, you know, the personality type. Uh, and you know, we, we don't want to have to talk to people. We want to be able to put our headphones on and get our work done. Uh, but now, now I'm in the position where if I want to do this thing, I know it is the right thing to do. I've got to convince others and that's not really my strong point. So how do you, how do you help them with that? Marisa (12:24) Mm -hmm. Yeah. Yeah, it's definitely a long road. I would say, you know, it's not done in a split second, especially when you're talking about larger companies, like any sort of fan company. They will take a lot of time to make this decision. So you have to be really committed, I think, to each person that you're talking to and each person that you're trying to help get moving with your product to really make them successful. And... For us, what we've been doing recently is we go in and we help them with that communication point, right? Like our developers know our tool the best. And so if there needs be, like we'll stop and we will actually go in and do a presentation to the wider team and be like, okay, you guys have set up this POC. You've tried it out. You really like it. Here are the benefits. Here are, you know, here's how we describe it. Here's how, you know, We have seen other companies succeed with this and we have some decks that are basically ready to go. So we can go in and actually help them with that communication stage as well. Brian (13:34) Yeah, that's awesome. Well, then let's, because this is fascinating, and I really enjoy kind of the idea of trying to be an advocate for the developers. But I'm curious as well, with your upcoming talk at Agile 2024, by the way, just I don't think I've said the date, but July 22nd through 26th in Dallas, Texas, go to. AgileAlliance .org. You can find out more information about it. I think I've told everyone here on the podcast, I'm speaking there as well. So come and see both of us in my hometown in Dallas. So Marie Kondo, your data pipelines with SQLMesh. Tell us what was the idea behind this, where you got the idea for this talk, and what it is you're trying to get across in it. Marisa (14:18) Thank you. Yeah, absolutely. Well, here's the story. One of our users, I was doing case studies for our, for us, because we need to understand our business value. We need to show people that like they can get this, these, you know, these cost savings, these productivity savings and things like that. So I've been doing interviews with some of the companies that currently use SQLMesh. And one of the, one of the interviewees, we were just chatting and he was like, oh, You know, he brought Marie Kondo up and he was like, yeah, like SQLMesh just brings me joy. It brings joy back to my data engineering. And I'm like, well, wait a minute. What, wait a minute. What do you mean? mean? like, oh yeah. Well, I mean, you know, we spend all of this time. Fretting Fretting about these data pipelines, getting the correct data down the pipeline, getting the business needs on the timeline that they need them, you know, updating your production value or your production environments with, with anything new that's been requested. And. Brian (15:01) you Marisa (15:21) there isn't really a proper process for data deployments, right? Like for code, you know, we have GitHub, we have Git, we have all these things, right? You make some changes, you save it, you test it out, you make some more changes, you save that, you test that out, right? Like all of these iterations, they create all these versions and these checkpoints. But on the data side of that, there is nothing. It's just like match disks and toothpicks that hold up some of these. Brian (15:30) Yeah. Hahaha. Marisa (15:49) some of these pipelines, right? And that's become the standard. I'm not saying that this is particular companies that are doing this or something like that. It's pretty much everybody. And all of this falls on top of your DevOps engineers or your data engineers that are a very small percentage of your company. And they're treading through this escalating technical debt, right? Every time you add a new table, every time you add a new dashboard, that's a new backend that they are managing, right? It becomes very stressful for them. And... This individual was saying that he had lost a lot of the joy out of his work and you spend how many hours a day working, right? This is a huge chunk of your life that it's just like, oh my God, I don't want to do this anymore. I don't want to make that change because the pipeline currently works. It's not broken. It's not broken. Don't change that type of thing, right? But this isn't what business is. This isn't what agile is. You're supposed to iterate. You're supposed to make these changes. You're supposed to investigate. Brian (16:24) Yeah. Right. Marisa (16:42) find new things and find new insights. And you can't do that unless you start changing things. And so he had said that once they started implementing SQLMesh with the different features that it has with this virtual data environments and these data versionings, and we have these data contracts that essentially allow you to turn, have checkpoints for your data and have essentially unit tests for your data. He was like, oh my gosh, now I can not have to worry about it. I can just try something, see if it works. And then if it doesn't, it doesn't matter because I can roll it back super easily and things like that. So that's really where the inspiration came from. Brian (17:21) That's awesome. Yeah, that's such a huge hole and it's such a needed kind of component to the stack, as you said, because, you know, I mean, you're right in kind of a programmer world. And, you know, if you're outside of the database world, there's all these tools you can use and put in place and test and see how things are going. But that fear that you have when you work in the database world where if I make the wrong error here, It could mess up all our data. It's not just that it's going to present it in the wrong way, but it could actually damage, which is a valuable asset. It's a hugely valuable asset, the data that the company has, maybe one of their most valuable assets. So yeah, that's an amazing tool. And... Marisa (18:10) Yeah. Cool. And these days, yeah. Brian (18:16) So this is also an open source project, right? So tell me how that's been interacting with the community on this and working in a corporate environment, but also it's an open source environment on this product that you're a developer advocate for. Marisa (18:19) Mm -hmm. Yep. Yeah. Well, I love it. I love open source. As I mentioned before, open source is kind of my bread and butter because Streamlit, you know, was essentially the other end. I've gone from the dashboard side, which Streamlit is basically a free open source dashboarding tool that's built just in Python, to the side of the tables that make that, and the SQL that makes those dashboards possible in the backend. And so this, it's something that I really love about my job because... I've been lucky enough to work with a bunch of tools that are super useful and that people really love, uh, and that they have that they, you know, people who co -founded it, that there are co -founders all came from huge fang companies, right? SQLMesh was basically born out of these data problems specifically to solve them because they had been literally experiencing them at these companies themselves. And they, you know, went out and were like, okay, what's the solution out there? And, you know, uh, Brian (19:21) Yeah. Marisa (19:31) There's this, there's another company called dbt. They were pretty much the first one on the scene. They've been around for like, I want to say like 10 or 15 years and they changed the space. Like they really did make some huge advancements. But the thing is, is they came at it from a completely different angle than we're trying to come at it because they came at it from the almost like the data science and data analytics side. And they weren't necessarily thinking about future, how big these data sets were going to get with. Brian (19:52) Yeah. Marisa (19:58) these different like Netflix, Airbnb and stuff like that, right? Their data sets are huge. They're parsing huge amounts of data. And, you know, in the current tools, the systems that you have, you have to refresh your entire data warehouse every time you want to make a change to production. If you have a terabyte worth of data that you're trying to refresh every single time you make a change, that literally, you're just, you're twiddling your thumbs. Your analysts are just like sitting around waiting anxiously to find out. Brian (20:03) Yeah. Marisa (20:27) you know, if those changes they just made to the sequel is actually viable and good, right? So, sorry, I think I lost. I think they lost the train of the question I got all passionate about. Brian (20:32) Yeah. No, no, no, that's fine. That's fine. No, I mean, I was just, I was asking about, you know, kind of the interactivity with the community on that and working with, you know, these open source projects, you know, you have volunteers, you have people who are giving up their time for free, basically to improve your product. That's got to come with a whole just mess of other considerations and concerns and how has that been? Marisa (21:07) Yeah, yeah, yeah. So that's been good. And where I had been going with that point was that I've been lucky enough to work with companies and tools that are super useful, that were developed out of pain points that other people have experienced. So that side of things has been really great. When you're trying to develop a community, I just try to make the most welcoming space so that people feel comfortable in asking any sorts of questions. Right? Because that is the only way that you kind of surface some of these things that might've been otherwise hiding. Right? If people are nervous to ask questions, then you're not going to really have a proper conversation around, Oh, well, wait a minute. How did you get to that point? Like what led you to use it this way or to do it this way or something like that. Um, and so, yeah, there's lots of considerations, but we've been lucky enough that the, you know, a lot of our users are very happy with it, but also are. Brian (21:38) Yeah. Marisa (22:01) because of that, they're very vocal and they are very happy to, you know, take that five, 10 minutes, fill that survey in or meet with me and talk through, you know, a case study or like how they found SQLMesh and how it's going and stuff like that. So I think that the real balance comes in as to how much time do you want to dedicate my time? Do you want to dedicate to doing these interviews and getting this feedback versus Like, oh, we've already got like a pretty large signal that the next feature they need is this. Like, let's just run and do that feature and get that out for them, as opposed to continuing to bog up their time with interview questions or survey questions and bog up my time with it as well. So I think that's more the balancing point is when do you start acting on the signals that you're starting to see from your community versus like just constantly collecting data? Brian (22:56) Yeah. And I love your point there about kind of creating that welcoming environment. It's very similar to what we talk about with just teams of the whole psychological safety kind of aspect of it. And if I feel like I can't say something without being made fun of or feel like I'm made to look stupid for my question, then you're right. I don't surface the things that I should, even if it's, you know, just, hey, I'm not sure how to do this and getting help for that kind of thing. Marisa (23:22) Mm -hmm. Yeah. And I mean, I feel like with open source communities as well, they're all online. And this is a, I like that psychological safety and I like to try and promote this friendly environment because there's no guarantee that the person on the other end even speaks English. They might be running it through a translator, right? Like you have no idea. So they need to have like that safe place that they can just be like, oh, Hey, I tried this, but I couldn't get it to work. Right. And then that's when you can start to expand and be like, Oh, okay. Well, actually this person is from. Brian (23:39) Yep. Marisa (23:54) I don't know, like France, they speak French, right? So you're trying to translate kind of back and forth, either in your head or with a tool. And it's kind of like, okay, well, what else can I do to help them? Right? It's like, oh, well, what they need is a more in -depth, like getting started guide, right? We need to add these steps in or put this little note in there that's like, hey, if you get tripped up and you get this error, that's because of this and this is how you fix it type of thing, right? So that you just kind of like fill in those gaps of knowledge. Brian (24:23) Yeah. I mean, there's probably so much that we could extract just from the remoteness of the team that you work with, because open source is an area where there is no office. It's not like we all come into the same place. And yet, it works. So for people who think it can't work if you're remote, well, open source is proof that it can. Yeah. Marisa (24:37) Yeah. Yeah, absolutely. And not because of that, our team is a hundred percent remote as well. Right. Like our entire Tobacco team works async, right. And we're only able to do that because of different tools and processes that we have put in place and, and that we do have this community and our community is a, is a Slack community. We actually, we could put a link in or something so that people can check it out if they were interested. But, but yeah. And, and that community isn't just about. Brian (25:11) Sure, sure, sure. Marisa (25:16) that specific tool. And I think that's also another thing that you want to surface when you're talking to developers is that this isn't just a space to ask questions about SQLMesh or SQLGlot, it's a space to talk in general. What are some of the best practices around evolving your data pipelines? What, you know, do you have any questions about your DevOps? Like, you know, what, what are people using for their cloud service provider? Why? Like, you know, did you switch from this to this and, you know, What was your thoughts around that? And, you know, what do people do with a dataset that is, you know, 300 rows and, and like 700 columns, like, how do you deal with this size of data? How do you want to, like, how do others mutate it? Like, do you incrementally load it? So just like try and load in the newest data. Do you only re when do you refresh these, like all these questions and all of this conversation needs to happen because we need to start deciding on a proper process and. DBT had started doing that and we're trying to continue this conversation. Brian (26:16) Yeah, I would imagine that's working with the product as well, that that's very beneficial to even product development. Because you can hear from them, even if it's not part of your product, but if you hear those questions about, hey, well, how have you handled this, or what do you do in this kind of scenario, I can imagine that would lead you to maybe new product ideas based off of some of those conversations. Yeah. Marisa (26:43) Oh, absolutely. 100%. I don't have like a specific example at the top of my head, but I definitely... Brian (26:48) Right. No, no, no. Yeah. Yeah, it could come from this. So it's community, right? I mean, it's just the importance of having that community and not just being, no, we can only talk about this one product here, but no, we're a supportive community and we help each other here. Yeah. Awesome. Well, this is fascinating. And I could talk with you about this for a long, long time. I'm looking forward to your talk. Marisa (26:55) Mm -hmm. Mm -hmm. Yeah. Yeah. Brian (27:16) So I haven't checked the schedule yet, but I'm gonna, hopefully it's not on one of those times when like, normally, I don't know if you find this as well, but it's like the ones that I start and think, oh, that's the one I really wanna go to, turns out to be the one where you're speaking or somebody who's a lifelong friend is talking at that time. You're like, I can't miss that person's. So I'm hoping that's not the case, cause I really wanna come and hear this talk and hear more about it. Marisa (27:17) Great, thank you. Yeah. Brian (27:45) Thank you for giving us your time and helping us to understand this, Marisa. I really appreciate you coming on. Marisa (27:49) What? Yeah. Thank you so much for having me.

Join our Pybites Community for free hereWe coach people with their Python, developer and mindset skills, more info here.---This week we have an exciting interview with Pydantic's creator Samuel Colvin. ---NOTE that it's best to watch this episode on YouTube, because Samuel demos Pydantic's new Logfire product as well as a bit of FastUI.

Está no ar, o Data Hackers News !! Os assuntos mais quentes da semana, com as principais notícias da área de Dados, IA e Tecnologia, que você também encontra na nossa Newsletter semanal, agora no Podcast do Data Hackers !! Aperte o play e ouça agora, o Data Hackers News dessa semana ! Para saber tudo sobre o que está acontecendo na área de dados, se inscreva na Newsletter semanal: ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠https://www.datahackers.news/⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠ Baixe o relatório completo do State of Data Brazil e os highlights da pesquisa : ⁠⁠⁠⁠https://stateofdata.datahackers.com.br/⁠⁠⁠⁠ Conheça nossos comentaristas do Data Hackers News: ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠Monique Femme⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠ ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠Paulo Vasconcellos⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠ Demais canais do Data Hackers: ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠Site⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠ ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠Linkedin⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠ ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠Instagram⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠ ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠Tik Tok⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠ ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠You Tube⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠ Matérias/assuntos comentados: Participe do Supletivo Data Hackers; Microsoft demonstra nova tecnologia de AI que permite dar vida a fotos antigas e obras de arte de maneira muito realista; Meta anuncia nova versão do LLaMa com integração no Whatsapp e Instagram. Já aproveita, para nos seguir no Spotify, Apple Podcasts, ou no seu player de podcasts favoritos ! --- Send in a voice message: https://podcasters.spotify.com/pod/show/datahackers/message

The wonderful Liz Acosta joins us on this episode of the podcast for a slightly philosophical chat on community, the different iterations of DevRel, and why we as humans will always gravitate to the things we perceive as “real”. Liz is the new Developer Content Marketing Manager for Streamlit at Snowflake, and this provides us with a wonderful opportunity to learn from someone right in the process of determining their role in an organisation! Liz talks to us about her plans for Streamlit, all of which centre connection and community heavily. As someone who has worked heavily in DevRel, Liz has no intentions of abandoning the Developer Advocates at Streamlit to fend for themselves, and her passion for connecting with people is more than clear. Liz tells us all about the evolution of DevRel, and why she believes increased regulation and codifying of previously informal guidelines is a good thing. Not only that, but for those of you who enjoy a good philosophical discussion, she explains that while we all appreciate things that are “real”, the not-so-real can be just as valuable. Whether that be AI generated copy, or an artist who lip-syncs their way through a concert! Reach out to Liz here: https://www.linkedin.com/in/lizacostalinkedin/ Find out more and listen to previous podcasts here: https://www.voxgig.com/podcast Subscribe to our newsletter for weekly updates and information about upcoming meetups: https://voxgig.substack.com/ Join the Dublin DevRel Meetup group here: www.devrelmeetup.com

Welcome to The Nonlinear Library, where we use Text-to-Speech software to convert the best writing from the Rationalist and EA communities into audio. This is: Mech Interp Challenge: January - Deciphering the Caesar Cipher Model, published by CallumMcDougall on January 1, 2024 on The AI Alignment Forum. I'm writing this post to discuss solutions to the November challenge, and present the challenge for this January. If you've not read the first post in this sequence, I'd recommend starting there - it outlines the purpose behind these challenges, and recommended prerequisite material. January Problem The problem for this month is interpreting a model which has been trained to classify a sequence according to the Caeser cipher shift value which was used to encode it. The sequences have been generated by taking English sentences containing only lowercase letters & punctuation, and choosing a random value X between 0 and 25 to rotate the letters (e.g. if the value was 3, then a becomes d, b becomes e, and so on, finishing with z becoming c). The model was trained using cross entropy loss to predict the shift value X for the text it's been fed, at every sequence position (so for a single sequence, the correct value will be the same at every sequence position, but since the model has bidirectional attention, it will find it easier to predict the value of X at later sequence positions). There are 3 different modes to the problem, to give you some more options! Each mode corresponds to a different dataset, but the same task & same model architecture. Easy mode In easy mode, the data was generated by: Choosing the 100 most frequent 3-letter words in the English Language (as approximated from a text file containing the book "Hitchhiker's Guide To The Galaxy") Choosing words from this len-100 list, with probabilities proportional to their frequency in the book Separating these words with spaces The model uses single-character tokenization. The vocabulary size is 27: each lowercase letter, plus whitespace. Medium mode This is identical to easy, the only difference is that the words are drawn from this len-100 list uniformly, rather than according to their true frequencies. Hard mode In hard mode, the data was generated from random slices of OpenWebText (i.e. natural language text from the internet). It was processed by converting all uppercase characters to lowercase, then removing all characters except for the 26 lowercase letters plus the ten characters "n .,:;?!'" (i.e. newline, space, and 8 common punctuation characters). In all 3 modes, the model's architecture is the same, and it was trained the same way. The model is attention only. It has 2 attention layers, with 2 heads per layer. It was trained with weight decay, and an Adam optimizer with linearly decaying learning rate. I don't expect this problem to be as difficult as some of the others in this sequence, however the presence of MLPs does provide a different kind of challenge. You can find more details on the Streamlit page, or this Colab notebook. Feel free to reach out if you have any questions! November Problem - Solutions The single attention head implements uniform attention to all previous tokens in the sequence. The OV matrix is essentially one-dimensional: it projects each token with value s onto su, where u is some vector in the residual stream learned by the model. The component of the residual stream in this direction then represents the cumulative mean (note, the cumulative mean rather than the cumulative sum, because attention is finite - for example, we expect the component to be the same after the sequences (1, 1, 2) and (1, 1, 2, 1, 1, 2) because net attention to each different token value will be the same). The model's "positive cumsum prediction direction" aligns closely with u, and vice-versa for the "negative cumsum prediction direction" - this allows the model to already get >50% accuracy before the MLP even comes into play. But without the MLP, the mod...

Paul Millerd spent several years working in strategy consulting and on the “default path” before deciding to walk away to work on his own in 2017. His book, The Pathless Path, chronicles his own journey and deep dive into the history of work and has been read by more than 40,000 people. His podcast, The Pathless Path Podcast, highlights conversations with others following unconventional paths. He also runs the online training business StrategyU, helping people learn the skills of consulting without having to work in the industry. In our conversation, Paul shares:• An explanation of the “default path” and the “pathless path”• Signs you may be stuck on the default path• How to inch your way toward the pathless path• Why Paul suggests everyone should take a three-month sabbatical• Tips for embracing fear and betting on yourself• How to work through the fear of losing money and prestige—Brought to you by Sanity—The most customizable content layer to power your growth engine | Maui Nui Venison—The healthiest red meat on the planet delivered directly to your door | Wix Studio—The web creation platform built for agencies—Find the full transcript at: https://www.lennyspodcast.com/redefining-success-money-and-belonging-paul-millerd-the-pathless-path/—Where to find Paul Millerd:• X: https://twitter.com/p_millerd• LinkedIn: https://www.linkedin.com/in/paulmillerd/• Website: https://pathlesspath.com/• Podcast: https://think-boundless.com/podcast/• Email: paul@strategyu.com—Where to find Lenny:• Newsletter: https://www.lennysnewsletter.com• X: https://twitter.com/lennysan• LinkedIn: https://www.linkedin.com/in/lennyrachitsky/—In this episode, we cover:(00:00) Paul's background(04:33) An explanation of the “default path”(06:32) Questions to help clarify which path you are on(07:35) Paul's thoughts on “remixing your path”(09:57) An explanation of the “pathless path” (12:06) Examples of the pathless path(13:54) Why meaning is hard to find and sustain on a traditional career path (16:05) A case for the three-month sabbatical (18:16) A mindfulness and self-reflection exercise(20:18) Why Paul recommends three months(22:28) Advice to founders on offering sabbaticals(23:40) Other tactics for self-discovery(27:08) The variability of income in self-employed roles (29:12) Methods for staying afloat after leaving your job(30:42) Tips for reframing your thoughts around money(33:19) Why betting on yourself usually works out(34:46) The importance of setting aside time for creative pursuits(36:22) How to dip your toes in and find your path (37:53) Lenny's personal journey(39:27) Advice on dealing with the naysayers (40:22) How to acknowledge and tame your fears(44:52) The “ship, quit, and learn” framework(46:39) Why the pathless path is one of constant reinvention(51:27) Paul's response to criticism (58:02) First steps for getting started on your journey(55:42) Lightning round—Referenced:• The Pathless Path: Imagining a New Story for Work and Life: https://www.amazon.com/Pathless-Path-Imagining-Story-Work/dp/B09QF6Q421• David Autor on X: https://twitter.com/davidautor• Tim Ferriss's blog: https://tim.blog/• Why you should define your fears instead of your goals | Tim Ferriss: https://www.youtube.com/watch?v=5J6jAC6XxAI• How Lenny Rachitsky Got 531,000 Substack Subscribers: https://www.youtube.com/watch?v=DMZem1NYfpM• The Lindy effect: https://en.wikipedia.org/wiki/Lindy_effect• StrategyU: https://strategyu.co/• David Deming's website: https://www.daviddeming.com/nyt-columns• The Great Work of Your Life: A Guide for the Journey to Your True Calling: https://www.amazon.com/Great-Work-Your-Life-Journey/dp/0553386077• Wanting: The Power of Mimetic Desire in Everyday Life: https://www.amazon.com/Wanting-Power-Mimetic-Desire-Everyday/dp/1250262488• Roadrunner: A Film About Anthony Bourdain: https://www.hbo.com/movies/roadrunner-a-film-about-anthony-bourdain• Nuna travel stroller: https://nunababy.com/usa/trvl-easy-fold-compact-stroller• Build Your Own Chatbot with OpenAI GPT-3 and Streamlit: https://medium.com/@avra42/build-your-own-chatbot-with-openai-gpt-3-and-streamlit-6f1330876846—Production and marketing by https://penname.co/. For inquiries about sponsoring the podcast, email podcast@lennyrachitsky.com.—Lenny may be an investor in the companies discussed. Get full access to Lenny's Newsletter at www.lennysnewsletter.com/subscribe

Sezon na keynote w pełni. Za nami keynotes OpenAI i GitHub. Czy świat stanął na krawędzi? Czy wszystko się zmieni i Ziemią zawładnie AI? Gdzie dowieźli, co dowieźli i jakie są tego konsekwencje. W Brew #10 (oklaski za okrągły odcinek

⬇⬇⬇APRIMI⬇⬇ Python e Streamlit: creiamo un app. Auto Elettrica VS Auto a Combustione Link app: https://confronto-auto-combustione-contro-auto-elettrica.streamlit.app/ Link github: https://github.com/DrElegantia/BEV-VS-TRAD/blob/main/BEV-VS-TRAD.py Abbonati qui: https://www.youtube.com/economiaitalia/join https://www.patreon.com/join/EconomiaItalia? Questo codice esegue un'analisi comparativa dei costi tra un'auto elettrica e un'auto a combustione, tenendo conto di vari fattori come il costo iniziale, i costi di ricarica/combustibile, l'assicurazione, il bollo, la manutenzione, i tagliandi e altri costi. Ecco come funziona il codice: Vengono acquisite le informazioni relative ai costi e ai parametri delle auto, come il costo iniziale, i chilometri annuali, i costi di rifornimento/ricarica, l'assicurazione, il bollo, la manutenzione, i tagliandi e altri fattori. Vengono calcolati i costi annuali fissi e variabili per entrambe le auto elettrica e a combustione, tenendo conto dei chilometri annuali, i costi di rifornimento/ricarica, l'assicurazione, il bollo, la manutenzione, i tagliandi e altri fattori. Il calcolo dei costi totali include sia i costi fissi che quelli variabili. Viene calcolato il "Break Even Point" (BEP), che rappresenta il numero di anni necessari affinché i costi totali dell'auto elettrica diventino inferiori rispetto a quelli dell'auto a combustione. Viene determinato quale tipo di auto (elettrica o a combustione) è più conveniente in base al confronto tra i costi variabili annuali. Vengono creati due istogrammi in pila per rappresentare la suddivisione dei costi totali (fissi + variabili) delle auto elettrica e a combustione. Uno degli istogrammi mostra i costi fissi e variabili separatamente, mentre l'altro mostra i costi totali. Viene creato un grafico a linee che mostra l'evoluzione dei costi totali delle due auto nel corso degli anni. Tutti i risultati vengono stampati in output, incluso il confronto tra i costi variabili annuali, il BEP, i costi totali delle due auto e la conclusione su quale auto è più conveniente in base alle impostazioni. Alla fine, i grafici vengono mostrati utilizzando la libreria Matplotlib. In sintesi, il codice ti consente di valutare in modo dettagliato e visivo i costi totali delle auto elettrica e a combustione nel corso degli anni, aiutandoti a prendere decisioni informate sulla scelta del tipo di veicolo da acquistare. Qui per segnalare temi: https://tellonym.me/dr.elegantia Podcast (su tutte le piattaforme): https://www.spreaker.com/show/dr-elegantia-podcast COME SOSTENERCI: Il nostro nuovo libro sull'economia: Guida Terrestre per Autoeconomisti https://www.poliniani.com/product-page/guida-terrestre link acquisto Amazon: https://amzn.to/36XTXs8 Acquistando le nostre T-shirt dedicate ai dati stampate in Serigrafia Artigianale con passione e orgoglio dai detenuti del Carcere Lorusso e Cutugno di Torino https://bit.ly/3zNsdkd e HTTPS://urly.it/3nga1 Guida al VOTO 2022: https://amzn.to/3KflXHd DonazionI Paypal: https://paypal.me/appuntiUAB Vuoi sostenermi ma non sborsare nemmeno un euro? Usa questo link per per il tuo prossimo acquisto su Amazon: https://amzn.to/2JGRyGT Qui trovi i libri che consiglio per iniziare a capirne di più sull'economia: https://www.youtube.com/watch?v=uEaIk8wQ3z8 Dove ci trovi: https://www.umbertobertonelli.it/info/ https://linktr.ee/economiaitalia La mia postazione: Logitech streamcam https://amzn.to/3HR6xq0 Luci https://amzn.to/3n6qtgP Shure MV7https://amzn.to/3HRh7k1 Asta https://amzn.to/3HSRvzY #economiaitalia #drelegantia #economiaDiventa un supporter di questo podcast: https://www.spreaker.com/podcast/dr-elegantia-podcast--5692498/support.

⬇⬇⬇APRIMI⬇⬇ Python e Streamlit: creiamo un app. Auto Elettrica VS Auto a Combustione Link app: https://confronto-auto-combustione-contro-auto-elettrica.streamlit.app/ Link github: https://github.com/DrElegantia/BEV-VS-TRAD/blob/main/BEV-VS-TRAD.py Abbonati qui: https://www.youtube.com/economiaitalia/join https://www.patreon.com/join/EconomiaItalia? Questo codice esegue un'analisi comparativa dei costi tra un'auto elettrica e un'auto a combustione, tenendo conto di vari fattori come il costo iniziale, i costi di ricarica/combustibile, l'assicurazione, il bollo, la manutenzione, i tagliandi e altri costi. Ecco come funziona il codice: Vengono acquisite le informazioni relative ai costi e ai parametri delle auto, come il costo iniziale, i chilometri annuali, i costi di rifornimento/ricarica, l'assicurazione, il bollo, la manutenzione, i tagliandi e altri fattori. Vengono calcolati i costi annuali fissi e variabili per entrambe le auto elettrica e a combustione, tenendo conto dei chilometri annuali, i costi di rifornimento/ricarica, l'assicurazione, il bollo, la manutenzione, i tagliandi e altri fattori. Il calcolo dei costi totali include sia i costi fissi che quelli variabili. Viene calcolato il "Break Even Point" (BEP), che rappresenta il numero di anni necessari affinché i costi totali dell'auto elettrica diventino inferiori rispetto a quelli dell'auto a combustione. Viene determinato quale tipo di auto (elettrica o a combustione) è più conveniente in base al confronto tra i costi variabili annuali. Vengono creati due istogrammi in pila per rappresentare la suddivisione dei costi totali (fissi + variabili) delle auto elettrica e a combustione. Uno degli istogrammi mostra i costi fissi e variabili separatamente, mentre l'altro mostra i costi totali. Viene creato un grafico a linee che mostra l'evoluzione dei costi totali delle due auto nel corso degli anni. Tutti i risultati vengono stampati in output, incluso il confronto tra i costi variabili annuali, il BEP, i costi totali delle due auto e la conclusione su quale auto è più conveniente in base alle impostazioni. Alla fine, i grafici vengono mostrati utilizzando la libreria Matplotlib. In sintesi, il codice ti consente di valutare in modo dettagliato e visivo i costi totali delle auto elettrica e a combustione nel corso degli anni, aiutandoti a prendere decisioni informate sulla scelta del tipo di veicolo da acquistare. Qui per segnalare temi: https://tellonym.me/dr.elegantia Podcast (su tutte le piattaforme): https://www.spreaker.com/show/dr-elegantia-podcast COME SOSTENERCI: Il nostro nuovo libro sull'economia: Guida Terrestre per Autoeconomisti https://www.poliniani.com/product-page/guida-terrestre link acquisto Amazon: https://amzn.to/36XTXs8 Acquistando le nostre T-shirt dedicate ai dati stampate in Serigrafia Artigianale con passione e orgoglio dai detenuti del Carcere Lorusso e Cutugno di Torino https://bit.ly/3zNsdkd e HTTPS://urly.it/3nga1 Guida al VOTO 2022: https://amzn.to/3KflXHd DonazionI Paypal: https://paypal.me/appuntiUAB Vuoi sostenermi ma non sborsare nemmeno un euro? Usa questo link per per il tuo prossimo acquisto su Amazon: https://amzn.to/2JGRyGT Qui trovi i libri che consiglio per iniziare a capirne di più sull'economia: https://www.youtube.com/watch?v=uEaIk8wQ3z8 Dove ci trovi: https://www.umbertobertonelli.it/info/ https://linktr.ee/economiaitalia La mia postazione: Logitech streamcam https://amzn.to/3HR6xq0 Luci https://amzn.to/3n6qtgP Shure MV7https://amzn.to/3HRh7k1 Asta https://amzn.to/3HSRvzY #economiaitalia #drelegantia #economia

This week we are talking about turning data into web applications with Streamlit. We're joined by two folks at Moser, Sarah Graddy and Brian MacDonald, who have been working with Streamlit on projects.Sarah Graddy is a Data &  Analytics Intern at Moser Consulting. She works with Python and GitHub and graduated from Kennesaw State University in May 2023 with a computer science degree.Brian MacDonald is a technology leader and innovatorwith a diverse background in the technology space. He loves to build products/services that will stand the test of time and be scalable and dynamic in their use. He enjoys working with a variety of clients and organizations to drive innovative solutions that will get the most from their data. He is committed to giving organizations the best solution that fits their specific needs and that's of most value to them. Creating the tools so that organizations can make data-based decisions is what motivates Brian to be creative and be driven towards long-term solutions.  

The importance of data teams is undeniable. Most companies today use data to drive decision-making on anything from software feature development to product strategy, hiring and marketing. In some companies data is the product, which can make data teams even more vital. But there's a common problem – analyzing data is hard and time consuming. The post Streamlit with Amanda Kelly appeared first on Software Engineering Daily.

The importance of data teams is undeniable. Most companies today use data to drive decision-making on anything from software feature development to product strategy, hiring and marketing. In some companies data is the product, which can make data teams even more vital. But there’s a common problem – analyzing data is hard and time consuming. The post Streamlit with Amanda Kelly appeared first on Software Engineering Daily.

Today I had the pleasure of bringing Tyler Richards back on the show. Last time I talked with him, he was a data scientist at Meta who had just written a book on Streamlit. Now he actually works at Streamlit and is releasing a second edition of his book. He joined Streamlit right before the were purchased by Snowflake and in this episode we talk about his experience going through an acquisition, landing a job at streamlit, and why more data scientists don't pursue entrepreneurial projects. Podcast Sponsors, Affiliates, and Partners:- Pathrise - http://pathrise.com/KenJee | Career mentorship for job applicants (Free till you land a job)- Taro - http://jointaro.com/r/kenj308 (20% discount) | Career mentorship if you already have a job - 365 Data Science (57% discount) - https://365datascience.pxf.io/P0jbBY | Learn data science today- Interview Query (10% discount) - https://www.interviewquery.com/?ref=kenjee |  Interview prep questionsTyler's Links:Install Streamlit - https://docs.streamlit.io/library/get-started/installationTwitter - https://twitter.com/tylerjrichardsLinkedIn - https://www.linkedin.com/in/tylerjrichards/Website - https://www.tylerjrichards.com/Book - https://blog.streamlit.io/streamlit-for-data-science-book/

Welcome to The Nonlinear Library, where we use Text-to-Speech software to convert the best writing from the Rationalist and EA communities into audio. This is: Mech Interp Challenge: September - Deciphering the Addition Model, published by TheMcDouglas on September 13, 2023 on The AI Alignment Forum. I'm writing this post to discuss solutions to the August challenge, and present the challenge for this September. Apologies for this coming so late in the month (EAGxBerlin was taking up much of my focus in the first half of this month). If you've not read the first post in this sequence, I'd recommend starting there - it outlines the purpose behind these challenges, and recommended prerequisite material. September Problem The problem for this month (or at least as much of the month as remains!) is interpreting a model which has been trained to perform simple addition. The model was fed input in the form of a sequence of digits (plus special + and = characters with token ids 10 and 11), and was tasked with predicting the sum of digits one sequence position before they would appear. Cross entropy loss was only applied to these four token positions, so the model's output at other sequence positions is meaningless. The model is attention-only, with 2 layers, and 3 heads per layer. It was trained with layernorm, weight decay, and an Adam optimizer with linearly decaying learning rate. You can find more details on the Streamlit page. Feel free to reach out if you have any questions! August Problem - Solutions You can read full solutions on the Streamlit page, or on my personal website. Both of these sources host interactive charts (Plotly and Circuitsvis) so are more suitable than a LessWrong/AF post to discuss the solutions in depth. However, I've presented a much shorter version of the solution below. If you're interested in these problems, I'd recommend having a try before you read on! The key idea with this model is path decomposition (see the corresponding section of A Mathematical Framework for Transformer Circuits). There are several different important types of path in this model, with different interpretations & purposes. We might call these negative paths and positive paths. The negative paths are designed to suppress repeated tokens, and the positive paths are designed to boost tokens which are more likely to be the first unique token. Let's start with the negative paths. Some layer 0 heads are duplicate token heads; they're composing with layer 1 heads to cause those heads to attend to & suppress duplicated tokens. This is done both with K-composition (heads in layer 1 suppress duplicated tokens because they attend to them more), and V-composition (the actual outputs of the DTHs are used as value input to heads in layer 1 to suppress duplicated tokens). Below is an example, where the second and third instances of a attend back to the first instance of a in head 0.2, and this composes with head 1.0 which attends back to (and suppresses) all the duplicated a tokens. Now, let's move on to the positive paths. Heads in layer 0 will attend to early tokens which aren't the same as the current destination token, because both these bits of evidence correlate with this token being the first unique token at this position (this is most obvious with the second token, since the first token is the correct answer here if and only if it doesn't equal the second token). Additionally, the outputs of heads in layer 0 are used as value input to heads in layer 1 to boost these tokens, i.e. as a virtual OV circuit. These paths aren't as obviously visible in the attention probabilities, because they're distributed: many tokens will weakly attend to some early token in a layer-0 head, and then all of those tokens will be weakly attended to by some layer-1 head. But the paths can be seen when we plot all the OV circuits, coloring each value by how much the final logits for that token are affected at the destination position: Another interesting o...

Welcome to The Nonlinear Library, where we use Text-to-Speech software to convert the best writing from the Rationalist and EA communities into audio. This is: Mech Interp Challenge: August - Deciphering the First Unique Character Model, published by TheMcDouglas on August 9, 2023 on The AI Alignment Forum. I'm writing this post to advertise the second in the sequence of monthly mechanistic interpretability challenges (the first to be posted on LessWrong). They are designed in the spirit of Stephen Casper's challenges, but with the more specific aim of working well in the context of the rest of the ARENA material, and helping people put into practice all the things they've learned so far. In this post, I'll describe the algorithmic problem & model trained to solve it, as well as the logistics for this challenge and the motivation for creating it. However, the main place you should access this problem is on the Streamlit page, where you can find setup code & instructions (as well as a link to a Colab notebook with all setup code included, if you'd prefer to use this). Task The algorithmic task is as follows: the model is presented with a sequence of characters, and for each character it has to correctly identify the first character in the sequence (up to and including the current character) which is unique up to that point. Model Our model was trained by minimising cross-entropy loss between its predictions and the true labels, at every sequence position simultaneously (including the zeroth sequence position, which is trivial because the input and target are both always "?"). You can inspect the notebook training_model.ipynb in the GitHub repo to see how it was trained. I used the version of the model which achieved highest accuracy over 50 epochs (accuracy ~99%). The model is is a 2-layer transformer with 3 attention heads, and causal attention. It includes layernorm, but no MLP layers. Note - although this model was trained for long enough to get loss close to zero (you can test this for yourself), it's not perfect. There are some weaknesses that the model has which might make it vulnerable to adversarial examples, and I've decided to leave these in. The model is still very good at its intended task, and the main focus of this challenge is on figuring out how it solves the task, not dissecting the situations where it fails. However, you might find that the adversarial examples help you understand the model better. Recommended material Material equivalent to the following from the ARENA course is highly recommended: [1.1] Transformer from scratch (sections 1-3) [1.2] Intro to Mech Interp (sections 1-3) The following material isn't essential, but is also recommended: [1.2] Intro to Mech Interp (section 4) If you want some guidance on how to get started, I'd recommend reading the solutions for the July problem - I expect there to be a lot of overlap in the best way to tackle these two problems. You can also reuse some of that code! Motivation Neel Nanda's post 200 COP in MI: Interpreting Algorithmic Problems does a good job explaining the motivation behind solving algorithmic problems such as these. I'd strongly recommend reading the whole post, because it also gives some high-level advice for approaching such problems. The main purpose of these challenges isn't to break new ground in mech interp, rather they're designed to help you practice using & develop better understanding for standard MI tools (e.g. interpreting attention, direct logit attribution), and more generally working with libraries like TransformerLens. Also, they're hopefully pretty fun, because why shouldn't we have some fun while we're learning? Logistics The solution to this problem will be published on the Streamlit in the first few days of September, at the same time as the next problem in the sequence. There will also likely be another LessWrong post. If you try to attempt this challenge, you can send your attempt in any of the following formats: Colab n...

Frank Slootman, CEO of Snowflake Computing, joins Sarah Guo and Elad Gil this week on No Priors. Before scaling Snowflake to its blockbuster IPO and beyond, Frank was also the CEO from early to scale for landmark enterprise companies ServiceNow and Data Domain. Frank grew up in the Netherlands and is also the author of three books: Amp It Up, Rise of the Data Cloud, and Tape Sucks. In this episode, our hosts talk with Frank about the opportunity for generative AI in the enterprise, why Snowflake isn't really a data warehousing company, their acquisitions of Neeva and Streamlit, apps within Snowflake, and how AI relates to traditional analytics and BI. He also talks about his personal journey, why it's always a good time to do performance management, and why most leaders struggle to raise the bar for performance. ** No Priors is taking a summer break! The podcast will be back with new episodes in three weeks. Join us on July 20th for a conversation with Devi Parikh, Research Director in Generative AI at Meta. ** No Priors is now on YouTube! Subscribe to the channel on YouTube and like this episode. Show Links: Forbes: How CEO-For-Hire Frank Slootman Turned Snowflake Into Software's Biggest-Ever IPO Amp It Up: Leading for Hypergrowth by Raising Expectations, Increasing Urgency, and Elevating Intensity Rise of the Data Cloud (Audible Audio Edition): Frank Slootman, Steve Hamm, Zach Hoffman, Snowflake: Books TAPE SUCKS: Inside Data Domain, A Silicon Valley Growth Story eBook : Slootman, Frank: Kindle Store Frank Slootman's LinkedIn Sign up for new podcasts every week. Email feedback to show@no-priors.com Follow us on Twitter: @NoPriorsPod | @Saranormous | @EladGil | @SnowflakeDB Show Notes: [00:06] - Frank's Insights on Career Success as a three-time CEO [12:42] - The message of his book Amp It Up [25:01] - Future of Natural Language and Data [36:29] - Data Management and Industry Transformation Future [45:13] - Managing Resources in Changing Economic Environment [50:09] - Amping Up Energy and Intensity Amid Economic Headwinds

In the season four premiere, Amanda Kelly, Product Director at Snowflake and Co-Founder and COO of Streamlit, shares her expert advice on data applications, creating a data-driven company, pushing your data forward, and so much more.--------How you approach data will define what's possible for your organization. Data engineers, data scientists, application developers, and a host of other data professionals who depend on the Snowflake Data Cloud continue to thrive thanks to a decade of technology breakthroughs. But that journey is only the beginning.Attend Snowflake Summit 2023 in Las Vegas June 26-29 to learn how to access, build, and monetize data, tools, models, and applications in ways that were previously unimaginable. Enable seamless alignment and collaboration across these crucial functions in the Data Cloud to transform nearly every aspect of your organization.Learn more and register at www.snowflake.com/summit

In this episode Benji talks to Courtland Goldengate, the VP of Marketing at Streamlit and a Director of Marketing at Snowflake. PLG is everywhere in Saas these days, but what if it's not right for many Saas solutions? Discussed in this episode:How to decipher if PLG is right for youKeeping focus in a ever-evolving marketing landscapeWhat EVERY Saas solution should steal from PLG

Tech News & AMA #20 with our community members Mehdi, Youssouf, Abderrahim and Manal. During this episode, we discuss 2023 IT trends predictions, BlablaConf updates, Hacktoberfest, and much more. Guests Manal Benchrif Abderrahim soubai Mehdi Cheracher Notes 0:00:00 - Introduction and welcoming 0:03:30 - Guests learning during last months 0:06:00 - Hacktoberfest 0:13:45 - Blablaconf update and call for speakers 0:26:00 - Moroccan national programming contest 2022 0:58:00 - State of Java report 1:01:00 - React new async rendering 1:10:00 - react query, qraphql, fetch, axios 1:15:00 - QA 1:23:40 - 2023 IT predictions 1:48:50 - GeeksBlabla Picks 1:44:00 - warming up and goodbye Links cfp.blablaconf hacktoberfest-open-source-2021 hacktoberfest Build Your First Sentiment Analysis Web App with Streamlit with Manal Benchrif |BlaBlaConf 2021 MNPC 2022 Editorial 2022 State of the Java Ecosystem Report stackblitz Top 5 Biggest Technology Trends In 2023 (Aikyo) rytr stateofdev Prepared and Presented by Youssouf El Azizi

Nesse episódio, Luan Moreno e Mateus Oliveira trazem as novidades da conferência data Summit 2022, sobre o Snowflake a plataforma nativa da nuvem que elimina a necessidade de data warehouses, data lakes e data marts separados, permitindo o compartilhamento seguro de dados em toda a organização e as novidades são As melhorias no snowflake:Unistore;Snowflake e iceberg Tables;Replicação, Failover e disaster recover; e muito mais dentro do nosso Engenharia de Dados[Cast] fique agente ate o final!Anúncios e Novidades da Conferência do Snowflake Summit 2022, segue informações:https://www.snowflake.com/summit/ StreamLithttps://events.snowflake.com/summit/agenda/session/887881  Inovações na Plataformahttps://events.snowflake.com/summit/agenda/session/849842Inovação do Armazenamento de Dados com Unis torehttps://events.snowflake.com/summit/agenda/session/834016 Snowflake Governancehttps://events.snowflake.com/summit/agenda/session/834019 O Futuro da Colaboraçãohttps://events.snowflake.com/summit/agenda/session/834018 Replicação e Failoverhttps://events.snowflake.com/summit/agenda/session/834021 Expansão das Capacidades do Storage com Apache Iceberghttps://events.snowflake.com/summit/agenda/session/884559 No YouTube possuímos um canal de Engenharia de Dados com os tópicos mais importantes dessa área e com lives todas as quartas-feiras.https://www.youtube.com/channel/UCnErAicaumKqIo4sanLo7vQ Quer ficar por dentro dessa área com posts e updates semanais, então acesse o LinkedIN para não perder nenhuma notícia.https://www.linkedin.com/in/luanmoreno/ Disponível no Spotify e na Apple Podcasthttps://open.spotify.com/show/5n9mOmAcjra9KbhKYpOMqYhttps://podcasts.apple.com/br/podcast/engenharia-de-dados-cast/  Luan Moreno = https://www.linkedin.com/in/luanmoreno/

Chanin Nantasenamat is behind the very successful YouTube channel Data Professor. He is an actual professor with a background in working in academia in the field of bioinformatics. Since the beginning of the pandemic he has started his brand Data Professor and has been taking the data science scene by storm. We talked about his background, how he decided to start his YouTube channel, how he built his YouTube channel to the size it is now, what future plans he has, his role at Streamlit and advice he can give to aspiring data science practitioners. Data Professor YouTube channel: https://www.youtube.com/c/DataProfessor Data Professor on Twitter: https://twitter.com/thedataprof Chanin's Data Science Landscape: https://twitter.com/thedataprof/status/1557407932060663808

In this episode, Adrien Treuille, Co-founder of Streamlit (acquired by Snowflake), shares how business people can use data science and machine learning, how you should incorporate data scientists into your organizations, and so much more.--------The Data Cloud World Tour is making 21 stops around the globe, so you can learn about the latest innovations to Snowflake's Data Cloud at a venue near you. Join your fellow data leaders at one of our full-day events to network with Snowflake customers and technology partners, attend educational breakout sessions, and learn how to drive more value from your data. Find an event near you at: https://www.snowflake.com/data-cloud-world-tour/

Snowflake is een van de snelst groeiende IT-bedrijven van dit moment. Na een zeer succesvolle IPO in 2020, is de waarde van het bedrijf fors gestegen, net als het aantal klanten. Veel grote IT-bedrijven werken inmiddels met het cloud dataplatform van Snowflake. Er waren flink wat aankondigingen rondom dit platform, maar in een soort sidenote, kwam naar voren dat Snowflake ook de security industrie wil veranderen. In deze podcast bespreken we de updates en die security aankondiging. Medio juni vond de Snowflake Summit plaats in Las Vegas, Techzine was erbij en sprak met diverse topmensen bij het bedrijf om te horen welke kant Snowflake op beweegt en wat er nieuw is aan het cloud data platform. Onder de belangrijkste aankondigingen is de algemene beschikbaarheid van Python, data analisten kunnen nu ook met Python applicaties en functies ontwikkelen bovenop het cloud data platform van Snowflake. Python is op dit moment wereldwijd de populairste programmeertaal en daarmee een zeer welkome toevoeging. Snowflake heeft ook een complete sandbox gebouwd rondom de Python-ondersteuning, zodat men alle Python libraries kan gebruiken, maar bij het uitvoeren van de code gebeurt dit in een sandbox, zodat er geen verbinding naar buiten kan worden gelegd. Hiermee kan Snowflake garanderen dat het werken met Python veilig is en niet je data ineens op straat komt te liggen. Een andere belangrijkste aankondiging was de integratie van Streamlit binnen Snowflake. Streamlit werd eerder dit jaar door Snowflake overgenomen. Hiermee kan je simpele applicaties bouwen die interacteren met de data in Snowflake. Tot slot de security aankondiging, waarbij Snowflake het cloud data platform beschikbaar en gereedmaakt voor security data. Het idee is dat alle security-oplossingen binnen een bedrijf hun intelligence data en log data in Snowflake opslaan, zodat alle security vendoren die het bedrijf gebruikt, ook gebruik kunnen maken van al die data. Door de security-oplossingen niet allemaal met hun eigen data silo's te laten werken, maar gezamenlijk één datapool te creëren, hebben ze meer informatie tot hun beschikking en zouden ze betere beschermingen moeten kunnen bieden. 

Summary Building an ML model is getting easier than ever, but it is still a challenge to get that model in front of the people that you built it for. Baseten is a platform that helps you quickly generate a full stack application powered by your model. You can easily create a web interface and APIs powered by the model you created, or a pre-trained model from their library. In this episode Tuhin Srivastava, co-founder of Basten, explains how the platform empowers data scientists and ML engineers to get their work in production without having to negotiate for help from their application development colleagues. Announcements Hello and welcome to the Machine Learning Podcast, the podcast about machine learning and how to bring it from idea to delivery. Data powers machine learning, but poor data quality is the largest impediment to effective ML today. Galileo is a collaborative data bench for data scientists building Natural Language Processing (NLP) models to programmatically inspect, fix and track their data across the ML workflow (pre-training, post-training and post-production) – no more excel sheets or ad-hoc python scripts. Get meaningful gains in your model performance fast, dramatically reduce data labeling and procurement costs, while seeing 10x faster ML iterations. Galileo is offering listeners a free 30 day trial and a 30% discount on the product there after. This offer is available until Aug 31, so go to themachinelearningpodcast.com/galileo and request a demo today! Do you wish you could use artificial intelligence to drive your business the way Big Tech does, but don’t have a money printer? Graft is a cloud-native platform that aims to make the AI of the 1% accessible to the 99%. Wield the most advanced techniques for unlocking the value of data, including text, images, video, audio, and graphs. No machine learning skills required, no team to hire, and no infrastructure to build or maintain. For more information on Graft or to schedule a demo, visit themachinelearningpodcast.com/graft today and tell them Tobias sent you. Predibase is a low-code ML platform without low-code limits. Built on top of our open source foundations of Ludwig and Horovod, our platform allows you to train state-of-the-art ML and deep learning models on your datasets at scale. Our platform works on text, images, tabular, audio and multi-modal data using our novel compositional model architecture. We allow users to operationalize models on top of the modern data stack, through REST and PQL – an extension of SQL that puts predictive power in the hands of data practitioners. Go to themachinelearningpodcast.com/predibase today to learn more and try it out! Your host is Tobias Macey and today I’m interviewing Tuhin Srivastava about Baseten, an ML Application Builder for data science and machine learning teams Interview Introduction How did you get involved in machine learning? Can you describe what Baseten is and the story behind it? Who are the target users for Baseten and what problems are you solving for them? What are some of the typical technical requirements for an application that is powered by a machine learning model? In the absence of Baseten, what are some of the common utilities/patterns that teams might rely on? What kinds of challenges do teams run into when serving a model in the context of an application? There are a number of projects that aim to reduce the overhead of turning a model into a usable product (e.g. Streamlit, Hex, etc.). What is your assessment of the current ecosystem for lowering the barrier to product development for ML and data science teams? Can you describe how the Baseten platform is designed? How have the design and goals of the project changed or evolved since you started working on it? How do you handle sandboxing of arbitrary user-managed code to ensure security and stability of the platform? How did you approach the system design to allow for mapping application development paradigms into a structure that was accessible to ML professionals? Can you describe the workflow for building an ML powered application? What types of models do you support? (e.g. NLP, computer vision, timeseries, deep neural nets vs. linear regression, etc.) How do the monitoring requirements shift for these different model types? What other challenges are presented by these different model types? What are the limitations in size/complexity/operational requirements that you have to impose to ensure a stable platform? What is the process for deploying model updates? For organizations that are relying on Baseten as a prototyping platform, what are the options for taking a successful application and handing it off to a product team for further customization? What are the most interesting, innovative, or unexpected ways that you have seen Baseten used? What are the most interesting, unexpected, or challenging lessons that you have learned while working on Baseten? When is Baseten the wrong choice? What do you have planned for the future of Baseten? Contact Info @tuhinone on Twitter LinkedIn Parting Question From your perspective, what is the biggest barrier to adoption of machine learning today? Closing Announcements Thank you for listening! Don’t forget to check out our other shows. The Data Engineering Podcast covers the latest on modern data management. Podcast.__init__ covers the Python language, its community, and the innovative ways it is being used. Visit the site to subscribe to the show, sign up for the mailing list, and read the show notes. If you’ve learned something or tried out a project from the show then tell us about it! Email hosts@themachinelearningpodcast.com) with your story. To help other people find the show please leave a review on iTunes and tell your friends and co-workers Links Baseten Gumroad scikit-learn Tensorflow Keras Streamlit Podcast.__init__ Episode Retool Hex Podcast.__init__ Episode Kubernetes React Monaco Huggingface Airtable Dall-E 2 GPT-3 Weights and Biases The intro and outro music is from Hitman’s Lovesong feat. Paola Graziano by The Freak Fandango Orchestra/CC BY-SA 3.0

Would you like a fast way to share your data science project results as an interactive dashboard instead of a Jupyter notebook? Streamlit is a library for creating simple web apps and dashboards using just Python. This week on the show, Christopher Trudeau is here, bringing another batch of PyCoder's Weekly articles and projects.

Are you working on or considering a machine learning project? On this episode, we'll meet three people from the MLOps community: Demetrios Brinkmann, Kate Kuznecova, and Vishnu Rachakonda. They are here to tell us about the lifecycle of a machine learning project. We'll talk about getting started with prototypes and choosing frameworks, the development process, and finally moving into deployment and production. Links from the show Demetrios Brinkmann: @DPBrinkm Kate Kuznecova: linkedin.com Vishnu Rachakonda: linkedin.com MLOps Community: mlops.community Feature stores: mlops.community Great Expectations: github.com source control: DVC: dvc.org StreamLit: streamlit.io MLOps Jobs: mlops.pallet.com Made With ML Apps: madewithml.com Banana.dev: banana.dev FastAPI: fastapi.tiangolo.com MLOps without too much Ops: towardsdatascience.com NBDev: nbdev.fast.ai The "Works on My Machine" Certification Program: codinghorror.com Watch this episode on YouTube: youtube.com Episode transcripts: talkpython.fm --- Stay in touch with us --- Subscribe on YouTube: youtube.com Follow Talk Python on Twitter: @talkpython Follow Michael on Twitter: @mkennedy Sponsors Sentry Error Monitoring, Code TALKPYTHON Stack Overflow Talk Python Training

Adrien Treuille is Co-founder & CEO of Streamlit, the open-source app framework for Machine Learning and Data Science teams to build data apps. The company's underlying open-source project, also named Streamlit, has over 18K stars and a community of 1.4K+ on Discord. Streamlit raised $60M+ from Sequoia, GGV, and Gradient Ventures and since the recording of this podcast was acquired by Snowflake for a reported $800M.

Most Searched Tickers Over The Last 24 Hours on Benzinga Pro CISOStraight from Benzinga newsdesk, hosts Brent Slava and Steve Krause bring you the market news and stocks to watch.Steve and Brent focus on the following today:SNOW - analysts are maintaining their buy ratings. a Morgan Stanley analyst said "By improving price performance for the end customer, SNOW takes a hit on near-term monetization, but should yield a higher volume of workloads moving to the platform and more durable growth longer-term. We would take advantage of the dislocation to build positions in this core software franchise"NRDY - a warning on the 10b5-1 trading plan when viewing form 4 filingsBenzinga Pro's Top 5 Stocks To Watch For Thursday, Mar. 3, 2022: SNOW, ZUO, RKDA, CISO, NRDYToday's 5 Stock Ideas:Snowflake (SNOW) - One of Thursday morning's biggest decliners. The stock was down 18% despite better-than-expected quarterly results. Snowflake also announced the purchase of a software framework company named Streamlit.Zuora (ZUO) - The company announced a $400 million investment from private-equity firm Silver Lake. Zuora also announced strong Q4 earnings and sales but gave guidance that was light.Arcadia Biosciences (RKDA) - A play on increased investor focus on crop and wheat prices. On Wednesday, traders circulated word Russia produces 66% of fertilizer used to grow corn and wheat. Arcadia describes itself as a producer of high-value crop improvements primarily in wheat, soy, and hemp.Cerberus Cyber Sentinel (CISO) -One of the most-searched tickers on Benzinga Pro Thursday morning. The company is a play on cybersecurity.Nerdy (NRDY) - A play on recent insider buying by the company's CEO, Charles Cohn. The exec bought about $1 million shares in company stock on Wednesday of this week.Hosts:Steve KrauseSr. Reporter Benzinga NewsdeskBrent SlavaSr. Reporter Head of Benzinga Newsdeskpro.benzinga.comFree 2-week trial, no credit card requiredUse coupon code YOUTUBE20 to get 20% offDisclaimer: All of the information, material, and/or content contained in this program is for informational purposes only. Investing in stocks, options, and futures is risky and not suitable for all investors. Please consult your own independent financial adviser before making any investment decisions.Advertising Inquiries: https://redcircle.com/brandsPrivacy & Opt-Out: https://redcircle.com/privacy

New entrepreneurs face many challenges at the start of their founding careers. Some of these challenges range from getting adequate resources, budgeting, market demand, and a fool-proof work process, to mention a few. But one of the first major challenges faced by startups is employee hiring and management. A company and its values are made up of its workers; “Your vibe attracts your tribe.” As such, it is imperative to have one's recruitment process figured out as it can make or mar the organization, its culture, and work output. The burning question now is, “How?”

Data science and machine learning developments can't have an impact if they don't get into everyone's hands. In this episode, Amanda Kelly of Streamlit joins Chris Grundemann and Stephen Foskett to talk about the challenges and opportunities in bringing data science to everyone's hands. How can we enable marketing, sales, marketing, and other elements of the business to access data and make informed decisions themselves? Data science teams have to meet business people where they are to better answer their questions rather than trying to create a perfect model in a vacuum. Streamlit helps to productize python scripts with a complete and flexible front-end and easy deployment, making it easy to share and iterate. These micro apps foster collaboration and interaction between data science and the business. Three Questions Stephen: Is AI just a new aspect of data science or is it truly a unique field? Chris: Can you think of an application for ML that has not yet been rolled out but will make a major impact in the future? Leon Adato: What responsibility do you think IT folks have to insure the things that we build are ethical? Gests and Hosts Amanda Kelly, Co-Founder of Streamlit. Follow her thoughts on the Streamlit Blog. Chris Grundemann, Gigaom Analyst and Managing Director at Grundemann Technology Solutions. Connect with Chris on ChrisGrundemann.com on Twitter at @ChrisGrundemann. Stephen Foskett, Publisher of Gestalt IT and Organizer of Tech Field Day. Find Stephen's writing at GestaltIT.com and on Twitter at @SFoskett. Date: 11/16/2021 Tags: @streamlit, @SFoskett, @ChrisGrundemann

We are actually giving away 4 copies of Tyler's book! Comment on the YouTube video with why you want to learn streamlit for a chance to win one of the copies! You can also win by commenting on the twitter post, my instagram post, or my linkedin post related to this podcast! Tyler is a data scientist at Facebook who recently published a book on the Python library Streamlit called 'Getting Started with Streamlit for Data Science'. He graduated from the University of Florida in 2018, and worked on election integrity problems for nonprofits and research labs while there.

Omar Sanseviero - Machine Learning Engineer en HuggingFace. https://huggingface.co/ https://twitter.com/osanseviero Temas clave: Cómo HuggingFace y Spacy colaboraron para crecer el ecosistema open source NLP. No existe una ruta de oro para entrar a aprender IA. Busca resolver un problema y a partir de ahí encuentra las herramientas para resolverlo. Tip: PyTorch para comenzar. Haz proyectos interesantes para ti. El material está gratis en internet. Conocimientos necesarios para hacer Machine Learning de frontera. Sin tantas matemáticas harás maravillas. Crea tu portafolio y gana exposición en el HUB de HuggingFace. Sube modelos y demos al HUB con solo Python. Streamlit como frontend/backend para tus aplicaciones. Luego súbelo a HuggingFace. Todo gratis. Tendencia: Machine learning aplicado a nuevas áreas. Aplicar las nuevas técnicas a aplicaciones como 3D, audio, química, etc. Transformers: en el centro de la innovación. Están mostrando alta aplicabilidad a otras áreas como imágenes y audio. Cómo un latinoamericano puede llegar a una empresa que esté avanzando la frontera del machine learning. Aplica sin miedo a trabajos, aprende inglés, comparte con la comunidad (Twitter, Discord), crea contenido, obtén visibilidad. Actitud de crecimiento. ¿Es necesaria la licenciatura/maestría en computación para hacer machine learning? El síndrome del impostor. Todos lo hemos tenido. Producido por ELIA - Escuela Latinoamericana de Inteligencia Artificial: ELIA (@elia_latam) Omar Espejel (@espejelomar) Estefanía Arias (@yharyarias5) Yeder Laura (@yederlvicente)

Alex Reece is a friend and manager at Streamlit (https://streamlit.io/). He joins us to talk about work politics and how much someone in the IC career path has to care about either of them. We also got into the path people take into management vs. staying in engineering forever, and Critter takes every opportunity to whine about imposter syndrome along the way. Leave us a review on Apple Podcasts! https://podcasts.apple.com/us/podcast/we-should-be-working/id1545522072 Homepage of the poddy: https://anchor.fm/jace-and-critter/ Critter's email: mikecrittenden@gmail.com Critter's Twitter: https://twitter.com/mcrittenden Critter's blog: https://critter.blog Jace just wants you to leave him alone.

Streamlit turns data scripts into shareable web apps in minutes. All in Python. All for free. No front‑end experience required. I speak with Adrien Treuille and Amanda Kelly to find out more about the company and how it fits into data pipelines. --- Send in a voice message: https://anchor.fm/theweeklysqueak/message

Today I had the pleasure of speaking with Adrien Treuille. He is co-founder and CEO of Streamlit which is pioneering next-generation tools for machine learning engineers. Adrien has been VP of Simulation Zoox, lead a Google X project, and was a Professor of Computer Science and Robotics at Carnegie Mellon. He gives talks around the world, including to the President's Council of Advisors on Science and Technology, and has won numerous scientific awards, including the MIT Top innovators under 35. Adrien and his work have been featured in the documentaries "What Will the Future Be Like" by PBS/NOVA, and "Lo and Behold" by Werner Herzog. In todays episode we talk about Adrien's experience moving from academia to industry, how Streamlit was inspired by hardware used to create electronic music, and the philosophy that guided most of his career decisions.

Big oof. Tulsa once again got throttled on the road at Houston, 86–59. With the postponement of the Tulane game we focus almost entirely on the offensive struggles that Tulsa has shown as of late (hint - it involves the number three, and r*bounding). Starting this week the schedule gets easier as we preview the road matchups against Temple (1/26) and ECU (1/30) CFB Recruiting Map: app · Streamlit (cfb-recruiting.herokuapp.com) --- Support this podcast: https://anchor.fm/thegoldenhurricast/support

今回から「最近の気になるトピック」コーナーを始めました！ ▶︎ 最近の気になるトピック @takapy Orchest というWebでpythonスクリプトのパイプラインが組めるライブラリ。前処理、モデリング、推論、などのスクリプト（notebookも可能っぽい）を用意しておいて、それらをブラウザ上でつなぎ合わせてパイプラインを作れる。 @yaginuuun BigQuery のアップデート 情報。 BigQuery UIのマルチタブ編集、クエリ補完など良さそうな機能がアナウンスされた。 ▶︎ 今回のテーマ：「Streamlit」 Streamlit というPythonでフロントエンドを構築できるフレームワークを触ってみたので、それについてお話しました。See Privacy Policy at https://art19.com/privacy and California Privacy Notice at https://art19.com/privacy#do-not-sell-my-info.

Streamlit is an open source framework that wants to revolutionize building machine learning and data science applications, and just secured a $21 million Series A funding to try and do this. Streamlit wants to be for data science what business intelligence tools have been for databases: a quick way to get results, without bothering much with the details. In today's episode, we welcome Streamlit CEO and Founder, Adrien Treuille. We discuss what makes Streamlit special, how it works, and where data-driven applications at large are going next. Article published on ZDNet in June 2020.