Podcasts about Equivalence

From Palantir and Two Sigma to building Goodfire into the poster-child for actionable mechanistic interpretability, Mark Bissell (Member of Technical Staff) and Myra Deng (Head of Product) are trying to turn “peeking inside the model” into a repeatable production workflow by shipping APIs, landing real enterprise deployments, and now scaling the bet with a recent $150M Series B funding round at a $1.25B valuation.In this episode, we go far beyond the usual “SAEs are cool” take. We talk about Goodfire's core bet: that the AI lifecycle is still fundamentally broken because the only reliable control we have is data and we post-train, RLHF, and fine-tune by “slurping supervision through a straw,” hoping the model picks up the right behaviors while quietly absorbing the wrong ones. Goodfire's answer is to build a bi-directional interface between humans and models: read what's happening inside, edit it surgically, and eventually use interpretability during training so customization isn't just brute-force guesswork.Mark and Myra walk through what that looks like when you stop treating interpretability like a lab demo and start treating it like infrastructure: lightweight probes that add near-zero latency, token-level safety filters that can run at inference time, and interpretability workflows that survive messy constraints (multilingual inputs, synthetic→real transfer, regulated domains, no access to sensitive data). We also get a live window into what “frontier-scale interp” means operationally (i.e. steering a trillion-parameter model in real time by targeting internal features) plus why the same tooling generalizes cleanly from language models to genomics, medical imaging, and “pixel-space” world models.We discuss:* Myra + Mark's path: Palantir (health systems, forward-deployed engineering) → Goodfire early team; Two Sigma → Head of Product, translating frontier interpretability research into a platform and real-world deployments* What “interpretability” actually means in practice: not just post-hoc poking, but a broader “science of deep learning” approach across the full AI lifecycle (data curation → post-training → internal representations → model design)* Why post-training is the first big wedge: “surgical edits” for unintended behaviors likereward hacking, sycophancy, noise learned during customization plus the dream of targeted unlearning and bias removal without wrecking capabilities* SAEs vs probes in the real world: why SAE feature spaces sometimes underperform classifiers trained on raw activations for downstream detection tasks (hallucination, harmful intent, PII), and what that implies about “clean concept spaces”* Rakuten in production: deploying interpretability-based token-level PII detection at inference time to prevent routing private data to downstream providers plus the gnarly constraints: no training on real customer PII, synthetic→real transfer, English + Japanese, and tokenization quirks* Why interp can be operationally cheaper than LLM-judge guardrails: probes are lightweight, low-latency, and don't require hosting a second large model in the loop* Real-time steering at frontier scale: a demo of steering Kimi K2 (~1T params) live and finding features via SAE pipelines, auto-labeling via LLMs, and toggling a “Gen-Z slang” feature across multiple layers without breaking tool use* Hallucinations as an internal signal: the case that models have latent uncertainty / “user-pleasing” circuitry you can detect and potentially mitigate more directly than black-box methods* Steering vs prompting: the emerging view that activation steering and in-context learning are more closely connected than people think, including work mapping between the two (even for jailbreak-style behaviors)* Interpretability for science: using the same tooling across domains (genomics, medical imaging, materials) to debug spurious correlations and extract new knowledge up to and including early biomarker discovery work with major partners* World models + “pixel-space” interpretability: why vision/video models make concepts easier to see, how that accelerates the feedback loop, and why robotics/world-model partners are especially interesting design partners* The north star: moving from “data in, weights out” to intentional model design where experts can impart goals and constraints directly, not just via reward signals and brute-force post-training—Goodfire AI* Website: https://goodfire.ai* LinkedIn: https://www.linkedin.com/company/goodfire-ai/* X: https://x.com/GoodfireAIMyra Deng* Website: https://myradeng.com/* LinkedIn: https://www.linkedin.com/in/myra-deng/* X: https://x.com/myra_dengMark Bissell* LinkedIn: https://www.linkedin.com/in/mark-bissell/* X: https://x.com/MarkMBissellFull Video EpisodeTimestamps00:00:00 Introduction00:00:05 Introduction to the Latent Space Podcast and Guests from Goodfire00:00:29 What is Goodfire? Mission and Focus on Interpretability00:01:01 Goodfire's Practical Approach to Interpretability00:01:37 Goodfire's Series B Fundraise Announcement00:02:04 Backgrounds of Mark and Myra from Goodfire00:02:51 Team Structure and Roles at Goodfire00:05:13 What is Interpretability? Definitions and Techniques00:05:30 Understanding Errors00:07:29 Post-training vs. Pre-training Interpretability Applications00:08:51 Using Interpretability to Remove Unwanted Behaviors00:10:09 Grokking, Double Descent, and Generalization in Models00:10:15 404 Not Found Explained00:12:06 Subliminal Learning and Hidden Biases in Models00:14:07 How Goodfire Chooses Research Directions and Projects00:15:00 Troubleshooting Errors00:16:04 Limitations of SAEs and Probes in Interpretability00:18:14 Rakuten Case Study: Production Deployment of Interpretability00:20:45 Conclusion00:21:12 Efficiency Benefits of Interpretability Techniques00:21:26 Live Demo: Real-Time Steering in a Trillion Parameter Model00:25:15 How Steering Features are Identified and Labeled00:26:51 Detecting and Mitigating Hallucinations Using Interpretability00:31:20 Equivalence of Activation Steering and Prompting00:34:06 Comparing Steering with Fine-Tuning and LoRA Techniques00:36:04 Model Design and the Future of Intentional AI Development00:38:09 Getting Started in Mechinterp: Resources, Programs, and Open Problems00:40:51 Industry Applications and the Rise of Mechinterp in Practice00:41:39 Interpretability for Code Models and Real-World Usage00:43:07 Making Steering Useful for More Than Stylistic Edits00:46:17 Applying Interpretability to Healthcare and Scientific Discovery00:49:15 Why Interpretability is Crucial in High-Stakes Domains like Healthcare00:52:03 Call for Design Partners Across Domains00:54:18 Interest in World Models and Visual Interpretability00:57:22 Sci-Fi Inspiration: Ted Chiang and Interpretability01:00:14 Interpretability, Safety, and Alignment Perspectives01:04:27 Weak-to-Strong Generalization and Future Alignment Challenges01:05:38 Final Thoughts and Hiring/Collaboration Opportunities at GoodfireTranscriptShawn Wang [00:00:05]: So welcome to the Latent Space pod. We're back in the studio with our special MechInterp co-host, Vibhu. Welcome. Mochi, Mochi's special co-host. And Mochi, the mechanistic interpretability doggo. We have with us Mark and Myra from Goodfire. Welcome. Thanks for having us on. Maybe we can sort of introduce Goodfire and then introduce you guys. How do you introduce Goodfire today?Myra Deng [00:00:29]: Yeah, it's a great question. So Goodfire, we like to say, is an AI research lab that focuses on using interpretability to understand, learn from, and design AI models. And we really believe that interpretability will unlock the new generation, next frontier of safe and powerful AI models. That's our description right now, and I'm excited to dive more into the work we're doing to make that happen.Shawn Wang [00:00:55]: Yeah. And there's always like the official description. Is there an understatement? Is there an unofficial one that sort of resonates more with a different audience?Mark Bissell [00:01:01]: Well, being an AI research lab that's focused on interpretability, there's obviously a lot of people have a lot that they think about when they think of interpretability. And I think we have a pretty broad definition of what that means and the types of places that can be applied. And in particular, applying it in production scenarios, in high stakes industries, and really taking it sort of from the research world into the real world. Which, you know. It's a new field, so that hasn't been done all that much. And we're excited about actually seeing that sort of put into practice.Shawn Wang [00:01:37]: Yeah, I would say it wasn't too long ago that Anthopic was like still putting out like toy models or superposition and that kind of stuff. And I wouldn't have pegged it to be this far along. When you and I talked at NeurIPS, you were talking a little bit about your production use cases and your customers. And then not to bury the lead, today we're also announcing the fundraise, your Series B. $150 million. $150 million at a 1.25B valuation. Congrats, Unicorn.Mark Bissell [00:02:02]: Thank you. Yeah, no, things move fast.Shawn Wang [00:02:04]: We were talking to you in December and already some big updates since then. Let's dive, I guess, into a bit of your backgrounds as well. Mark, you were at Palantir working on health stuff, which is really interesting because the Goodfire has some interesting like health use cases. I don't know how related they are in practice.Mark Bissell [00:02:22]: Yeah, not super related, but I don't know. It was helpful context to know what it's like. Just to work. Just to work with health systems and generally in that domain. Yeah.Shawn Wang [00:02:32]: And Mara, you were at Two Sigma, which actually I was also at Two Sigma back in the day. Wow, nice.Myra Deng [00:02:37]: Did we overlap at all?Shawn Wang [00:02:38]: No, this is when I was briefly a software engineer before I became a sort of developer relations person. And now you're head of product. What are your sort of respective roles, just to introduce people to like what all gets done in Goodfire?Mark Bissell [00:02:51]: Yeah, prior to Goodfire, I was at Palantir for about three years as a forward deployed engineer, now a hot term. Wasn't always that way. And as a technical lead on the health care team and at Goodfire, I'm a member of the technical staff. And honestly, that I think is about as specific as like as as I could describe myself because I've worked on a range of things. And, you know, it's it's a fun time to be at a team that's still reasonably small. I think when I joined one of the first like ten employees, now we're above 40, but still, it looks like there's always a mix of research and engineering and product and all of the above. That needs to get done. And I think everyone across the team is, you know, pretty, pretty switch hitter in the roles they do. So I think you've seen some of the stuff that I worked on related to image models, which was sort of like a research demo. More recently, I've been working on our scientific discovery team with some of our life sciences partners, but then also building out our core platform for more of like flexing some of the kind of MLE and developer skills as well.Shawn Wang [00:03:53]: Very generalist. And you also had like a very like a founding engineer type role.Myra Deng [00:03:58]: Yeah, yeah.Shawn Wang [00:03:59]: So I also started as I still am a member of technical staff, did a wide range of things from the very beginning, including like finding our office space and all of this, which is we both we both visited when you had that open house thing. It was really nice.Myra Deng [00:04:13]: Thank you. Thank you. Yeah. Plug to come visit our office.Shawn Wang [00:04:15]: It looked like it was like 200 people. It has room for 200 people. But you guys are like 10.Myra Deng [00:04:22]: For a while, it was very empty. But yeah, like like Mark, I spend. A lot of my time as as head of product, I think product is a bit of a weird role these days, but a lot of it is thinking about how do we take our frontier research and really apply it to the most important real world problems and how does that then translate into a platform that's repeatable or a product and working across, you know, the engineering and research teams to make that happen and also communicating to the world? Like, what is interpretability? What is it used for? What is it good for? Why is it so important? All of these things are part of my day-to-day as well.Shawn Wang [00:05:01]: I love like what is things because that's a very crisp like starting point for people like coming to a field. They all do a fun thing. Vibhu, why don't you want to try tackling what is interpretability and then they can correct us.Vibhu Sapra [00:05:13]: Okay, great. So I think like one, just to kick off, it's a very interesting role to be head of product, right? Because you guys, at least as a lab, you're more of an applied interp lab, right? Which is pretty different than just normal interp, like a lot of background research. But yeah. You guys actually ship an API to try these things. You have Ember, you have products around it, which not many do. Okay. What is interp? So basically you're trying to have an understanding of what's going on in model, like in the model, in the internal. So different approaches to do that. You can do probing, SAEs, transcoders, all this stuff. But basically you have an, you have a hypothesis. You have something that you want to learn about what's happening in a model internals. And then you're trying to solve that from there. You can do stuff like you can, you know, you can do activation mapping. You can try to do steering. There's a lot of stuff that you can do, but the key question is, you know, from input to output, we want to have a better understanding of what's happening and, you know, how can we, how can we adjust what's happening on the model internals? How'd I do?Mark Bissell [00:06:12]: That was really good. I think that was great. I think it's also a, it's kind of a minefield of a, if you ask 50 people who quote unquote work in interp, like what is interpretability, you'll probably get 50 different answers. And. Yeah. To some extent also like where, where good fire sits in the space. I think that we're an AI research company above all else. And interpretability is a, is a set of methods that we think are really useful and worth kind of specializing in, in order to accomplish the goals we want to accomplish. But I think we also sort of see some of the goals as even more broader as, as almost like the science of deep learning and just taking a not black box approach to kind of any part of the like AI development life cycle, whether that. That means using interp for like data curation while you're training your model or for understanding what happened during post-training or for the, you know, understanding activations and sort of internal representations, what is in there semantically. And then a lot of sort of exciting updates that were, you know, are sort of also part of the, the fundraise around bringing interpretability to training, which I don't think has been done all that much before. A lot of this stuff is sort of post-talk poking at models as opposed to. To actually using this to intentionally design them.Shawn Wang [00:07:29]: Is this post-training or pre-training or is that not a useful.Myra Deng [00:07:33]: Currently focused on post-training, but there's no reason the techniques wouldn't also work in pre-training.Shawn Wang [00:07:38]: Yeah. It seems like it would be more active, applicable post-training because basically I'm thinking like rollouts or like, you know, having different variations of a model that you can tweak with the, with your steering. Yeah.Myra Deng [00:07:50]: And I think in a lot of the news that you've seen in, in, on like Twitter or whatever, you've seen a lot of unintended. Side effects come out of post-training processes, you know, overly sycophantic models or models that exhibit strange reward hacking behavior. I think these are like extreme examples. There's also, you know, very, uh, mundane, more mundane, like enterprise use cases where, you know, they try to customize or post-train a model to do something and it learns some noise or it doesn't appropriately learn the target task. And a big question that we've always had is like, how do you use your understanding of what the model knows and what it's doing to actually guide the learning process?Shawn Wang [00:08:26]: Yeah, I mean, uh, you know, just to anchor this for people, uh, one of the biggest controversies of last year was 4.0 GlazeGate. I've never heard of GlazeGate. I didn't know that was what it was called. The other one, they called it that on the blog post and I was like, well, how did OpenAI call it? Like officially use that term. And I'm like, that's funny, but like, yeah, I guess it's the pitch that if they had worked a good fire, they wouldn't have avoided it. Like, you know what I'm saying?Myra Deng [00:08:51]: I think so. Yeah. Yeah.Mark Bissell [00:08:53]: I think that's certainly one of the use cases. I think. Yeah. Yeah. I think the reason why post-training is a place where this makes a lot of sense is a lot of what we're talking about is surgical edits. You know, you want to be able to have expert feedback, very surgically change how your model is doing, whether that is, you know, removing a certain behavior that it has. So, you know, one of the things that we've been looking at or is, is another like common area where you would want to make a somewhat surgical edit is some of the models that have say political bias. Like you look at Quen or, um, R1 and they have sort of like this CCP bias.Shawn Wang [00:09:27]: Is there a CCP vector?Mark Bissell [00:09:29]: Well, there's, there are certainly internal, yeah. Parts of the representation space where you can sort of see where that lives. Yeah. Um, and you want to kind of, you know, extract that piece out.Shawn Wang [00:09:40]: Well, I always say, you know, whenever you find a vector, a fun exercise is just like, make it very negative to see what the opposite of CCP is.Mark Bissell [00:09:47]: The super America, bald eagles flying everywhere. But yeah. So in general, like lots of post-training tasks where you'd want to be able to, to do that. Whether it's unlearning a certain behavior or, you know, some of the other kind of cases where this comes up is, are you familiar with like the, the grokking behavior? I mean, I know the machine learning term of grokking.Shawn Wang [00:10:09]: Yeah.Mark Bissell [00:10:09]: Sort of this like double descent idea of, of having a model that is able to learn a generalizing, a generalizing solution, as opposed to even if memorization of some task would suffice, you want it to learn the more general way of doing a thing. And so, you know, another. A way that you can think about having surgical access to a model's internals would be learn from this data, but learn in the right way. If there are many possible, you know, ways to, to do that. Can make interp solve the double descent problem?Shawn Wang [00:10:41]: Depends, I guess, on how you. Okay. So I, I, I viewed that double descent as a problem because then you're like, well, if the loss curves level out, then you're done, but maybe you're not done. Right. Right. But like, if you actually can interpret what is a generalizing or what you're doing. What is, what is still changing, even though the loss is not changing, then maybe you, you can actually not view it as a double descent problem. And actually you're just sort of translating the space in which you view loss and like, and then you have a smooth curve. Yeah.Mark Bissell [00:11:11]: I think that's certainly like the domain of, of problems that we're, that we're looking to get.Shawn Wang [00:11:15]: Yeah. To me, like double descent is like the biggest thing to like ML research where like, if you believe in scaling, then you don't need, you need to know where to scale. And. But if you believe in double descent, then you don't, you don't believe in anything where like anything levels off, like.Vibhu Sapra [00:11:30]: I mean, also tendentially there's like, okay, when you talk about the China vector, right. There's the subliminal learning work. It was from the anthropic fellows program where basically you can have hidden biases in a model. And as you distill down or, you know, as you train on distilled data, those biases always show up, even if like you explicitly try to not train on them. So, you know, it's just like another use case of. Okay. If we can interpret what's happening in post-training, you know, can we clear some of this? Can we even determine what's there? Because yeah, it's just like some worrying research that's out there that shows, you know, we really don't know what's going on.Mark Bissell [00:12:06]: That is. Yeah. I think that's the biggest sentiment that we're sort of hoping to tackle. Nobody knows what's going on. Right. Like subliminal learning is just an insane concept when you think about it. Right. Train a model on not even the logits, literally the output text of a bunch of random numbers. And now your model loves owls. And you see behaviors like that, that are just, they defy, they defy intuition. And, and there are mathematical explanations that you can get into, but. I mean.Shawn Wang [00:12:34]: It feels so early days. Objectively, there are a sequence of numbers that are more owl-like than others. There, there should be.Mark Bissell [00:12:40]: According to, according to certain models. Right. It's interesting. I think it only applies to models that were initialized from the same starting Z. Usually, yes.Shawn Wang [00:12:49]: But I mean, I think that's a, that's a cheat code because there's not enough compute. But like if you believe in like platonic representation, like probably it will transfer across different models as well. Oh, you think so?Mark Bissell [00:13:00]: I think of it more as a statistical artifact of models initialized from the same seed sort of. There's something that is like path dependent from that seed that might cause certain overlaps in the latent space and then sort of doing this distillation. Yeah. Like it pushes it towards having certain other tendencies.Vibhu Sapra [00:13:24]: Got it. I think there's like a bunch of these open-ended questions, right? Like you can't train in new stuff during the RL phase, right? RL only reorganizes weights and you can only do stuff that's somewhat there in your base model. You're not learning new stuff. You're just reordering chains and stuff. But okay. My broader question is when you guys work at an interp lab, how do you decide what to work on and what's kind of the thought process? Right. Because we can ramble for hours. Okay. I want to know this. I want to know that. But like, how do you concretely like, you know, what's the workflow? Okay. There's like approaches towards solving a problem, right? I can try prompting. I can look at chain of thought. I can train probes, SAEs. But how do you determine, you know, like, okay, is this going anywhere? Like, do we have set stuff? Just, you know, if you can help me with all that. Yeah.Myra Deng [00:14:07]: It's a really good question. I feel like we've always at the very beginning of the company thought about like, let's go and try to learn what isn't working in machine learning today. Whether that's talking to customers or talking to researchers at other labs, trying to understand both where the frontier is going and where things are really not falling apart today. And then developing a perspective on how we can push the frontier using interpretability methods. And so, you know, even our chief scientist, Tom, spends a lot of time talking to customers and trying to understand what real world problems are and then taking that back and trying to apply the current state of the art to those problems and then seeing where they fall down basically. And then using those failures or those shortcomings to understand what hills to climb when it comes to interpretability research. So like on the fundamental side, for instance, when we have done some work applying SAEs and probes, we've encountered, you know, some shortcomings in SAEs that we found a little bit surprising. And so have gone back to the drawing board and done work on that. And then, you know, we've done some work on better foundational interpreter models. And a lot of our team's research is focused on what is the next evolution beyond SAEs, for instance. And then when it comes to like control and design of models, you know, we tried steering with our first API and realized that it still fell short of black box techniques like prompting or fine tuning. And so went back to the drawing board and we're like, how do we make that not the case and how do we improve it beyond that? And one of our researchers, Ekdeep, who just joined is actually Ekdeep and Atticus are like steering experts and have spent a lot of time trying to figure out like, what is the research that enables us to actually do this in a much more powerful, robust way? So yeah, the answer is like, look at real world problems, try to translate that into a research agenda and then like hill climb on both of those at the same time.Shawn Wang [00:16:04]: Yeah. Mark has the steering CLI demo queued up, which we're going to go into in a sec. But I always want to double click on when you drop hints, like we found some problems with SAEs. Okay. What are they? You know, and then we can go into the demo. Yeah.Myra Deng [00:16:19]: I mean, I'm curious if you have more thoughts here as well, because you've done it in the healthcare domain. But I think like, for instance, when we do things like trying to detect behaviors within models that are harmful or like behaviors that a user might not want to have in their model. So hallucinations, for instance, harmful intent, PII, all of these things. We first tried using SAE probes for a lot of these tasks. So taking the feature activation space from SAEs and then training classifiers on top of that, and then seeing how well we can detect the properties that we might want to detect in model behavior. And we've seen in many cases that probes just trained on raw activations seem to perform better than SAE probes, which is a bit surprising if you think that SAEs are actually also capturing the concepts that you would want to capture cleanly and more surgically. And so that is an interesting observation. I don't think that is like, I'm not down on SAEs at all. I think there are many, many things they're useful for, but we have definitely run into cases where I think the concept space described by SAEs is not as clean and accurate as we would expect it to be for actual like real world downstream performance metrics.Mark Bissell [00:17:34]: Fair enough. Yeah. It's the blessing and the curse of unsupervised methods where you get to peek into the AI's mind. But sometimes you wish that you saw other things when you walked inside there. Although in the PII instance, I think weren't an SAE based approach actually did prove to be the most generalizable?Myra Deng [00:17:53]: It did work well in the case that we published with Rakuten. And I think a lot of the reasons it worked well was because we had a noisier data set. And so actually the blessing of unsupervised learning is that we actually got to get more meaningful, generalizable signal from SAEs when the data was noisy. But in other cases where we've had like good data sets, it hasn't been the case.Shawn Wang [00:18:14]: And just because you named Rakuten and I don't know if we'll get it another chance, like what is the overall, like what is Rakuten's usage or production usage? Yeah.Myra Deng [00:18:25]: So they are using us to essentially guardrail and inference time monitor their language model usage and their agent usage to detect things like PII so that they don't route private user information.Myra Deng [00:18:41]: And so that's, you know, going through all of their user queries every day. And that's something that we deployed with them a few months ago. And now we are actually exploring very early partnerships, not just with Rakuten, but with other people around how we can help with potentially training and customization use cases as well. Yeah.Shawn Wang [00:19:03]: And for those who don't know, like it's Rakuten is like, I think number one or number two e-commerce store in Japan. Yes. Yeah.Mark Bissell [00:19:10]: And I think that use case actually highlights a lot of like what it looks like to deploy things in practice that you don't always think about when you're doing sort of research tasks. So when you think about some of the stuff that came up there that's more complex than your idealized version of a problem, they were encountering things like synthetic to real transfer of methods. So they couldn't train probes, classifiers, things like that on actual customer data of PII. So what they had to do is use synthetic data sets. And then hope that that transfer is out of domain to real data sets. And so we can evaluate performance on the real data sets, but not train on customer PII. So that right off the bat is like a big challenge. You have multilingual requirements. So this needed to work for both English and Japanese text. Japanese text has all sorts of quirks, including tokenization behaviors that caused lots of bugs that caused us to be pulling our hair out. And then also a lot of tasks you'll see. You might make simplifying assumptions if you're sort of treating it as like the easiest version of the problem to just sort of get like general results where maybe you say you're classifying a sentence to say, does this contain PII? But the need that Rakuten had was token level classification so that you could precisely scrub out the PII. So as we learned more about the problem, you're sort of speaking about what that looks like in practice. Yeah. A lot of assumptions end up breaking. And that was just one instance where you. A problem that seems simple right off the bat ends up being more complex as you keep diving into it.Vibhu Sapra [00:20:41]: Excellent. One of the things that's also interesting with Interp is a lot of these methods are very efficient, right? So where you're just looking at a model's internals itself compared to a separate like guardrail, LLM as a judge, a separate model. One, you have to host it. Two, there's like a whole latency. So if you use like a big model, you have a second call. Some of the work around like self detection of hallucination, it's also deployed for efficiency, right? So if you have someone like Rakuten doing it in production live, you know, that's just another thing people should consider.Mark Bissell [00:21:12]: Yeah. And something like a probe is super lightweight. Yeah. It's no extra latency really. Excellent.Shawn Wang [00:21:17]: You have the steering demos lined up. So we were just kind of see what you got. I don't, I don't actually know if this is like the latest, latest or like alpha thing.Mark Bissell [00:21:26]: No, this is a pretty hacky demo from from a presentation that someone else on the team recently gave. So this will give a sense for, for technology. So you can see the steering and action. Honestly, I think the biggest thing that this highlights is that as we've been growing as a company and taking on kind of more and more ambitious versions of interpretability related problems, a lot of that comes to scaling up in various different forms. And so here you're going to see steering on a 1 trillion parameter model. This is Kimi K2. And so it's sort of fun that in addition to the research challenges, there are engineering challenges that we're now tackling. Cause for any of this to be sort of useful in production, you need to be thinking about what it looks like when you're using these methods on frontier models as opposed to sort of like toy kind of model organisms. So yeah, this was thrown together hastily, pretty fragile behind the scenes, but I think it's quite a fun demo. So screen sharing is on. So I've got two terminal sessions pulled up here. On the left is a forked version that we have of the Kimi CLI that we've got running to point at our custom hosted Kimi model. And then on the right is a set up that will allow us to steer on certain concepts. So I should be able to chat with Kimi over here. Tell it hello. This is running locally. So the CLI is running locally, but the Kimi server is running back to the office. Well, hopefully should be, um, that's too much to run on that Mac. Yeah. I think it's, uh, it takes a full, like each 100 node. I think it's like, you can. You can run it on eight GPUs, eight 100. So, so yeah, Kimi's running. We can ask it a prompt. It's got a forked version of our, uh, of the SG line code base that we've been working on. So I'm going to tell it, Hey, this SG line code base is slow. I think there's a bug. Can you try to figure it out? There's a big code base, so it'll, it'll spend some time doing this. And then on the right here, I'm going to initialize in real time. Some steering. Let's see here.Mark Bissell [00:23:33]: searching for any. Bugs. Feature ID 43205.Shawn Wang [00:23:38]: Yeah.Mark Bissell [00:23:38]: 20, 30, 40. So let me, uh, this is basically a feature that we found that inside Kimi seems to cause it to speak in Gen Z slang. And so on the left, it's still sort of thinking normally it might take, I don't know, 15 seconds for this to kick in, but then we're going to start hopefully seeing him do this code base is massive for real. So we're going to start. We're going to start seeing Kimi transition as the steering kicks in from normal Kimi to Gen Z Kimi and both in its chain of thought and its actual outputs.Mark Bissell [00:24:19]: And interestingly, you can see, you know, it's still able to call tools, uh, and stuff. It's um, it's purely sort of it's it's demeanor. And there are other features that we found for interesting things like concision. So that's more of a practical one. You can make it more concise. Um, the types of programs, uh, programming languages that uses, but yeah, as we're seeing it come in. Pretty good. Outputs.Shawn Wang [00:24:43]: Scheduler code is actually wild.Vibhu Sapra [00:24:46]: Yo, this code is actually insane, bro.Vibhu Sapra [00:24:53]: What's the process of training in SAE on this, or, you know, how do you label features? I know you guys put out a pretty cool blog post about, um, finding this like autonomous interp. Um, something. Something about how agents for interp is different than like coding agents. I don't know while this is spewing up, but how, how do we find feature 43, two Oh five. Yeah.Mark Bissell [00:25:15]: So in this case, um, we, our platform that we've been building out for a long time now supports all the sort of classic out of the box interp techniques that you might want to have like SAE training, probing things of that kind, I'd say the techniques for like vanilla SAEs are pretty well established now where. You take your model that you're interpreting, run a whole bunch of data through it, gather activations, and then yeah, pretty straightforward pipeline to train an SAE. There are a lot of different varieties. There's top KSAEs, batch top KSAEs, um, normal ReLU SAEs. And then once you have your sparse features to your point, assigning labels to them to actually understand that this is a gen Z feature, that's actually where a lot of the kind of magic happens. Yeah. And the most basic standard technique is look at all of your d input data set examples that cause this feature to fire most highly. And then you can usually pick out a pattern. So for this feature, If I've run a diverse enough data set through my model feature 43, two Oh five. Probably tends to fire on all the tokens that sounds like gen Z slang. You know, that's the, that's the time of year to be like, Oh, I'm in this, I'm in this Um, and, um, so, you know, you could have a human go through all 43,000 concepts andVibhu Sapra [00:26:34]: And I've got to ask the basic question, you know, can we get examples where it hallucinates, pass it through, see what feature activates for hallucinations? Can I just, you know, turn hallucination down?Myra Deng [00:26:51]: Oh, wow. You really predicted a project we're already working on right now, which is detecting hallucinations using interpretability techniques. And this is interesting because hallucinations is something that's very hard to detect. And it's like a kind of a hairy problem and something that black box methods really struggle with. Whereas like Gen Z, you could always train a simple classifier to detect that hallucinations is harder. But we've seen that models internally have some... Awareness of like uncertainty or some sort of like user pleasing behavior that leads to hallucinatory behavior. And so, yeah, we have a project that's trying to detect that accurately. And then also working on mitigating the hallucinatory behavior in the model itself as well.Shawn Wang [00:27:39]: Yeah, I would say most people are still at the level of like, oh, I would just turn temperature to zero and that turns off hallucination. And I'm like, well, that's a fundamental misunderstanding of how this works. Yeah.Mark Bissell [00:27:51]: Although, so part of what I like about that question is you, there are SAE based approaches that might like help you get at that. But oftentimes the beauty of SAEs and like we said, the curse is that they're unsupervised. So when you have a behavior that you deliberately would like to remove, and that's more of like a supervised task, often it is better to use something like probes and specifically target the thing that you're interested in reducing as opposed to sort of like hoping that when you fragment the latent space, one of the vectors that pops out.Vibhu Sapra [00:28:20]: And as much as we're training an autoencoder to be sparse, we're not like for sure certain that, you know, we will get something that just correlates to hallucination. You'll probably split that up into 20 other things and who knows what they'll be.Mark Bissell [00:28:36]: Of course. Right. Yeah. So there's no sort of problems with like feature splitting and feature absorption. And then there's the off target effects, right? Ideally, you would want to be very precise where if you reduce the hallucination feature, suddenly maybe your model can't write. Creatively anymore. And maybe you don't like that, but you want to still stop it from hallucinating facts and figures.Shawn Wang [00:28:55]: Good. So Vibhu has a paper to recommend there that we'll put in the show notes. But yeah, I mean, I guess just because your demo is done, any any other things that you want to highlight or any other interesting features you want to show?Mark Bissell [00:29:07]: I don't think so. Yeah. Like I said, this is a pretty small snippet. I think the main sort of point here that I think is exciting is that there's not a whole lot of inter being applied to models quite at this scale. You know, Anthropic certainly has some some. Research and yeah, other other teams as well. But it's it's nice to see these techniques, you know, being put into practice. I think not that long ago, the idea of real time steering of a trillion parameter model would have sounded.Shawn Wang [00:29:33]: Yeah. The fact that it's real time, like you started the thing and then you edited the steering vector.Vibhu Sapra [00:29:38]: I think it's it's an interesting one TBD of what the actual like production use case would be on that, like the real time editing. It's like that's the fun part of the demo, right? You can kind of see how this could be served behind an API, right? Like, yes, you're you only have so many knobs and you can just tweak it a bit more. And I don't know how it plays in. Like people haven't done that much with like, how does this work with or without prompting? Right. How does this work with fine tuning? Like, there's a whole hype of continual learning, right? So there's just so much to see. Like, is this another parameter? Like, is it like parameter? We just kind of leave it as a default. We don't use it. So I don't know. Maybe someone here wants to put out a guide on like how to use this with prompting when to do what?Mark Bissell [00:30:18]: Oh, well, I have a paper recommendation. I think you would love from Act Deep on our team, who is an amazing researcher, just can't say enough amazing things about Act Deep. But he actually has a paper that as well as some others from the team and elsewhere that go into the essentially equivalence of activation steering and in context learning and how those are from a he thinks of everything in a cognitive neuroscience Bayesian framework, but basically how you can precisely show how. Prompting in context, learning and steering exhibit similar behaviors and even like get quantitative about the like magnitude of steering you would need to do to induce a certain amount of behavior similar to certain prompting, even for things like jailbreaks and stuff. It's a really cool paper. Are you saying steering is less powerful than prompting? More like you can almost write a formula that tells you how to convert between the two of them.Myra Deng [00:31:20]: And so like formally equivalent actually in the in the limit. Right.Mark Bissell [00:31:24]: So like one case study of this is for jailbreaks there. I don't know. Have you seen the stuff where you can do like many shot jailbreaking? You like flood the context with examples of the behavior. And the topic put out that paper.Shawn Wang [00:31:38]: A lot of people were like, yeah, we've been doing this, guys.Mark Bissell [00:31:40]: Like, yeah, what's in this in context learning and activation steering equivalence paper is you can like predict the number. Number of examples that you will need to put in there in order to jailbreak the model. That's cool. By doing steering experiments and using this sort of like equivalence mapping. That's cool. That's really cool. It's very neat. Yeah.Shawn Wang [00:32:02]: I was going to say, like, you know, I can like back rationalize that this makes sense because, you know, what context is, is basically just, you know, it updates the KV cache kind of and like and then every next token inference is still like, you know, the sheer sum of everything all the way. It's plus all the context. It's up to date. And you could, I guess, theoretically steer that with you probably replace that with your steering. The only problem is steering typically is on one layer, maybe three layers like like you did. So it's like not exactly equivalent.Mark Bissell [00:32:33]: Right, right. There's sort of you need to get precise about, yeah, like how you sort of define steering and like what how you're modeling the setup. But yeah, I've got the paper pulled up here. Belief dynamics reveal the dual nature. Yeah. The title is Belief Dynamics Reveal the Dual Nature of Incompetence. And it's an exhibition of the practical context learning and activation steering. So Eric Bigelow, Dan Urgraft on the who are doing fellowships at Goodfire, Ekt Deep's the final author there.Myra Deng [00:32:59]: I think actually to your question of like, what is the production use case of steering? I think maybe if you just think like one level beyond steering as it is today. Like imagine if you could adapt your model to be, you know, an expert legal reasoner. Like in almost real time, like very quickly. efficiently using human feedback or using like your semantic understanding of what the model knows and where it knows that behavior. I think that while it's not clear what the product is at the end of the day, it's clearly very valuable. Thinking about like what's the next interface for model customization and adaptation is a really interesting problem for us. Like we have heard a lot of people actually interested in fine-tuning an RL for open weight models in production. And so people are using things like Tinker or kind of like open source libraries to do that, but it's still very difficult to get models fine-tuned and RL'd for exactly what you want them to do unless you're an expert at model training. And so that's like something we'reShawn Wang [00:34:06]: looking into. Yeah. I never thought so. Tinker from Thinking Machines famously uses rank one LoRa. Is that basically the same as steering? Like, you know, what's the comparison there?Mark Bissell [00:34:19]: Well, so in that case, you are still applying updates to the parameters, right?Shawn Wang [00:34:25]: Yeah. You're not touching a base model. You're touching an adapter. It's kind of, yeah.Mark Bissell [00:34:30]: Right. But I guess it still is like more in parameter space then. I guess it's maybe like, are you modifying the pipes or are you modifying the water flowing through the pipes to get what you're after? Yeah. Just maybe one way.Mark Bissell [00:34:44]: I like that analogy. That's my mental map of it at least, but it gets at this idea of model design and intentional design, which is something that we're, that we're very focused on. And just the fact that like, I hope that we look back at how we're currently training models and post-training models and just think what a primitive way of doing that right now. Like there's no intentionalityShawn Wang [00:35:06]: really in... It's just data, right? The only thing in control is what data we feed in.Mark Bissell [00:35:11]: So, so Dan from Goodfire likes to use this analogy of, you know, he has a couple of young kids and he talks about like, what if I could only teach my kids how to be good people by giving them cookies or like, you know, giving them a slap on the wrist if they do something wrong, like not telling them why it was wrong or like what they should have done differently or something like that. Just figure it out. Right. Exactly. So that's RL. Yeah. Right. And, and, you know, it's sample inefficient. There's, you know, what do they say? It's like slurping feedback. It's like, slurping supervision. Right. And so you'd like to get to the point where you can have experts giving feedback to their models that are, uh, internalized and, and, you know, steering is an inference time way of sort of getting that idea. But ideally you're moving to a world whereVibhu Sapra [00:36:04]: it is much more intentional design in perpetuity for these models. Okay. This is one of the questions we asked Emmanuel from Anthropic on the podcast a few months ago. Basically the question, was you're at a research lab that does model training, foundation models, and you're on an interp team. How does it tie back? Right? Like, does this, do ideas come from the pre-training team? Do they go back? Um, you know, so for those interested, you can, you can watch that. There wasn't too much of a connect there, but it's still something, you know, it's something they want toMark Bissell [00:36:33]: push for down the line. It can be useful for all of the above. Like there are certainly post-hocVibhu Sapra [00:36:39]: use cases where it doesn't need to touch that. I think the other thing a lot of people forget is this stuff isn't too computationally expensive, right? Like I would say, if you're interested in getting into research, MechInterp is one of the most approachable fields, right? A lot of this train an essay, train a probe, this stuff, like the budget for this one, there's already a lot done. There's a lot of open source work. You guys have done some too. Um, you know,Shawn Wang [00:37:04]: There's like notebooks from the Gemini team for Neil Nanda or like, this is how you do it. Just step through the notebook.Vibhu Sapra [00:37:09]: Even if you're like, not even technical with any of this, you can still make like progress. There, you can look at different activations, but, uh, if you do want to get into training, you know, training this stuff, correct me if I'm wrong is like in the thousands of dollars, not even like, it's not that high scale. And then same with like, you know, applying it, doing it for post-training or all this stuff is fairly cheap in scale of, okay. I want to get into like model training. I don't have compute for like, you know, pre-training stuff. So it's, it's a very nice field to get into. And also there's a lot of like open questions, right? Um, some of them have to go with, okay, I want a product. I want to solve this. Like there's also just a lot of open-ended stuff that people could work on. That's interesting. Right. I don't know if you guys have any calls for like, what's open questions, what's open work that you either open collaboration with, or like, you'd just like to see solved or just, you know, for people listening that want to get into McInturk because people always talk about it. What are, what are the things they should check out? Start, of course, you know, join you guys as well. I'm sure you're hiring.Myra Deng [00:38:09]: There's a paper, I think from, was it Lee, uh, Sharky? It's open problems and, uh, it's, it's a bit of interpretability, which I recommend everyone who's interested in the field. Read. I'm just like a really comprehensive overview of what are the things that experts in the field think are the most important problems to be solved. I also think to your point, it's been really, really inspiring to see, I think a lot of young people getting interested in interpretability, actually not just young people also like scientists to have been, you know, experts in physics for many years and in biology or things like this, um, transitioning into interp, because the barrier of, of what's now interp. So it's really cool to see a number to entry is, you know, in some ways low and there's a lot of information out there and ways to get started. There's this anecdote of like professors at universities saying that all of a sudden every incoming PhD student wants to study interpretability, which was not the case a few years ago. So it just goes to show how, I guess, like exciting the field is, how fast it's moving, how quick it is to get started and things like that.Mark Bissell [00:39:10]: And also just a very welcoming community. You know, there's an open source McInturk Slack channel. There are people are always posting questions and just folks in the space are always responsive if you ask things on various forums and stuff. But yeah, the open paper, open problems paper is a really good one.Myra Deng [00:39:28]: For other people who want to get started, I think, you know, MATS is a great program. What's the acronym for? Machine Learning and Alignment Theory Scholars? It's like the...Vibhu Sapra [00:39:40]: Normally summer internship style.Myra Deng [00:39:42]: Yeah, but they've been doing it year round now. And actually a lot of our full-time staff have come through that program or gone through that program. And it's great for anyone who is transitioning into interpretability. There's a couple other fellows programs. We do one as well as Anthropic. And so those are great places to get started if anyone is interested.Mark Bissell [00:40:03]: Also, I think been seen as a research field for a very long time. But I think engineering... I think engineers are sorely wanted for interpretability as well, especially at Goodfire, but elsewhere, as it does scale up.Shawn Wang [00:40:18]: I should mention that Lee actually works with you guys, right? And in the London office and I'm adding our first ever McInturk track at AI Europe because I see this industry applications now emerging. And I'm pretty excited to, you know, help push that along. Yeah, I was looking forward to that. It'll effectively be the first industry McInturk conference. Yeah. I'm so glad you added that. You know, it's still a little bit of a bet. It's not that widespread, but I can definitely see this is the time to really get into it. We want to be early on things.Mark Bissell [00:40:51]: For sure. And I think the field understands this, right? So at ICML, I think the title of the McInturk workshop this year was actionable interpretability. And there was a lot of discussion around bringing it to various domains. Everyone's adding pragmatic, actionable, whatever.Shawn Wang [00:41:10]: It's like, okay, well, we weren't actionable before, I guess. I don't know.Vibhu Sapra [00:41:13]: And I mean, like, just, you know, being in Europe, you see the Interp room. One, like old school conferences, like, I think they had a very tiny room till they got lucky and they got it doubled. But there's definitely a lot of interest, a lot of niche research. So you see a lot of research coming out of universities, students. We covered the paper last week. It's like two unknown authors, not many citations. But, you know, you can make a lot of meaningful work there. Yeah. Yeah. Yeah.Shawn Wang [00:41:39]: Yeah. I think people haven't really mentioned this yet. It's just Interp for code. I think it's like an abnormally important field. We haven't mentioned this yet. The conspiracy theory last two years ago was when the first SAE work came out of Anthropic was they would do like, oh, we just used SAEs to turn the bad code vector down and then turn up the good code. And I think like, isn't that the dream? Like, you know, like, but basically, I guess maybe, why is it funny? Like, it's... If it was realistic, it would not be funny. It would be like, no, actually, we should do this. But it's funny because we know there's like, we feel there's some limitations to what steering can do. And I think a lot of the public image of steering is like the Gen Z stuff. Like, oh, you can make it really love the Golden Gate Bridge, or you can make it speak like Gen Z. To like be a legal reasoner seems like a huge stretch. Yeah. And I don't know if that will get there this way. Yeah.Myra Deng [00:42:36]: I think, um, I will say we are announcing. Something very soon that I will not speak too much about. Um, but I think, yeah, this is like what we've run into again and again is like, we, we don't want to be in the world where steering is only useful for like stylistic things. That's definitely not, not what we're aiming for. But I think the types of interventions that you need to do to get to things like legal reasoning, um, are much more sophisticated and require breakthroughs in, in learning algorithms. And that's, um...Shawn Wang [00:43:07]: And is this an emergent property of scale as well?Myra Deng [00:43:10]: I think so. Yeah. I mean, I think scale definitely helps. I think scale allows you to learn a lot of information and, and reduce noise across, you know, large amounts of data. But I also think we think that there's ways to do things much more effectively, um, even, even at scale. So like actually learning exactly what you want from the data and not learning things that you do that you don't want exhibited in the data. So we're not like anti-scale, but we are also realizing that scale is not going to get us anywhere. It's not going to get us to the type of AI development that we want to be at in, in the future as these models get more powerful and get deployed in all these sorts of like mission critical contexts. Current life cycle of training and deploying and evaluations is, is to us like deeply broken and has opportunities to, to improve. So, um, more to come on that very, very soon.Mark Bissell [00:44:02]: And I think that that's a use basically, or maybe just like a proof point that these concepts do exist. Like if you can manipulate them in the precise best way, you can get the ideal combination of them that you desire. And steering is maybe the most coarse grained sort of peek at what that looks like. But I think it's evocative of what you could do if you had total surgical control over every concept, every parameter. Yeah, exactly.Myra Deng [00:44:30]: There were like bad code features. I've got it pulled up.Vibhu Sapra [00:44:33]: Yeah. Just coincidentally, as you guys are talking.Shawn Wang [00:44:35]: This is like, this is exactly.Vibhu Sapra [00:44:38]: There's like specifically a code error feature that activates and they show, you know, it's not, it's not typo detection. It's like, it's, it's typos in code. It's not typical typos. And, you know, you can, you can see it clearly activates where there's something wrong in code. And they have like malicious code, code error. They have a whole bunch of sub, you know, sub broken down little grain features. Yeah.Shawn Wang [00:45:02]: Yeah. So, so the, the rough intuition for me, the, why I talked about post-training was that, well, you just, you know, have a few different rollouts with all these things turned off and on and whatever. And then, you know, you can, that's, that's synthetic data you can kind of post-train on. Yeah.Vibhu Sapra [00:45:13]: And I think we make it sound easier than it is just saying, you know, they do the real hard work.Myra Deng [00:45:19]: I mean, you guys, you guys have the right idea. Exactly. Yeah. We replicated a lot of these features in, in our Lama models as well. I remember there was like.Vibhu Sapra [00:45:26]: And I think a lot of this stuff is open, right? Like, yeah, you guys opened yours. DeepMind has opened a lot of essays on Gemma. Even Anthropic has opened a lot of this. There's, there's a lot of resources that, you know, we can probably share of people that want to get involved.Shawn Wang [00:45:41]: Yeah. And special shout out to like Neuronpedia as well. Yes. Like, yeah, amazing piece of work to visualize those things.Myra Deng [00:45:49]: Yeah, exactly.Shawn Wang [00:45:50]: I guess I wanted to pivot a little bit on, onto the healthcare side, because I think that's a big use case for you guys. We haven't really talked about it yet. This is a bit of a crossover for me because we are, we are, we do have a separate science pod that we're starting up for AI, for AI for science, just because like, it's such a huge investment category and also I'm like less qualified to do it, but we actually have bio PhDs to cover that, which is great, but I need to just kind of recover, recap your work, maybe on the evil two stuff, but then, and then building forward.Mark Bissell [00:46:17]: Yeah, for sure. And maybe to frame up the conversation, I think another kind of interesting just lens on interpretability in general is a lot of the techniques that were described. are ways to solve the AI human interface problem. And it's sort of like bidirectional communication is the goal there. So what we've been talking about with intentional design of models and, you know, steering, but also more advanced techniques is having humans impart our desires and control into models and over models. And the reverse is also very interesting, especially as you get to superhuman models, whether that's narrow superintelligence, like these scientific models that work on genomics, data, medical imaging, things like that. But down the line, you know, superintelligence of other forms as well. What knowledge can the AIs teach us as sort of that, that the other direction in that? And so some of our life science work to date has been getting at exactly that question, which is, well, some of it does look like debugging these various life sciences models, understanding if they're actually performing well, on tasks, or if they're picking up on spurious correlations, for instance, genomics models, you would like to know whether they are sort of focusing on the biologically relevant things that you care about, or if it's using some simpler correlate, like the ancestry of the person that it's looking at. But then also in the instances where they are superhuman, and maybe they are understanding elements of the human genome that we don't have names for or specific, you know, yeah, discoveries that they've made that that we don't know about, that's, that's a big goal. And so we're already seeing that, right, we are partnered with organizations like Mayo Clinic, leading research health system in the United States, our Institute, as well as a startup called Prima Menta, which focuses on neurodegenerative disease. And in our partnership with them, we've used foundation models, they've been training and applied our interpretability techniques to find novel biomarkers for Alzheimer's disease. So I think this is just the tip of the iceberg. But it's, that's like a flavor of some of the things that we're working on.Shawn Wang [00:48:36]: Yeah, I think that's really fantastic. Obviously, we did the Chad Zuckerberg pod last year as well. And like, there's a plethora of these models coming out, because there's so much potential and research. And it's like, very interesting how it's basically the same as language models, but just with a different underlying data set. But it's like, it's the same exact techniques. Like, there's no change, basically.Mark Bissell [00:48:59]: Yeah. Well, and even in like other domains, right? Like, you know, robotics, I know, like a lot of the companies just use Gemma as like the like backbone, and then they like make it into a VLA that like takes these actions. It's, it's, it's transformers all the way down. So yeah.Vibhu Sapra [00:49:15]: Like we have Med Gemma now, right? Like this week, even there was Med Gemma 1.5. And they're training it on this stuff, like 3d scans, medical domain knowledge, and all that stuff, too. So there's a push from both sides. But I think the thing that, you know, one of the things about McInturpp is like, you're a little bit more cautious in some domains, right? So healthcare, mainly being one, like guardrails, understanding, you know, we're more risk adverse to something going wrong there. So even just from a basic understanding, like, if we're trusting these systems to make claims, we want to know why and what's going on.Myra Deng [00:49:51]: Yeah, I think there's totally a kind of like deployment bottleneck to actually using. foundation models for real patient usage or things like that. Like, say you're using a model for rare disease prediction, you probably want some explanation as to why your model predicted a certain outcome, and an interpretable explanation at that. So that's definitely a use case. But I also think like, being able to extract scientific information that no human knows to accelerate drug discovery and disease treatment and things like that actually is a really, really big unlock for science, like scientific discovery. And you've seen a lot of startups, like say that they're going to accelerate scientific discovery. And I feel like we actually are doing that through our interp techniques. And kind of like, almost by accident, like, I think we got reached out to very, very early on from these healthcare institutions. And none of us had healthcare.Shawn Wang [00:50:49]: How did they even hear of you? A podcast.Myra Deng [00:50:51]: Oh, okay. Yeah, podcast.Vibhu Sapra [00:50:53]: Okay, well, now's that time, you know.Myra Deng [00:50:55]: Everyone can call us.Shawn Wang [00:50:56]: Podcasts are the most important thing. Everyone should listen to podcasts.Myra Deng [00:50:59]: Yeah, they reached out. They were like, you know, we have these really smart models that we've trained, and we want to know what they're doing. And we were like, really early that time, like three months old, and it was a few of us. And we were like, oh, my God, we've never used these models. Let's figure it out. But it's also like, great proof that interp techniques scale pretty well across domains. We didn't really have to learn too much about.Shawn Wang [00:51:21]: Interp is a machine learning technique, machine learning skills everywhere, right? Yeah. And it's obviously, it's just like a general insight. Yeah. Probably to finance too, I think, which would be fun for our history. I don't know if you have anything to say there.Mark Bissell [00:51:34]: Yeah, well, just across the science. Like, we've also done work on material science. Yeah, it really runs the gamut.Vibhu Sapra [00:51:40]: Yeah. Awesome. And, you know, for those that should reach out, like, you're obviously experts in this, but like, is there a call out for people that you're looking to partner with, design partners, people to use your stuff outside of just, you know, the general developer that wants to. Plug and play steering stuff, like on the research side more so, like, are there ideal design partners, customers, stuff like that?Myra Deng [00:52:03]: Yeah, I can talk about maybe non-life sciences, and then I'm curious to hear from you on the life sciences side. But we're looking for design partners across many domains, language, anyone who's customizing language models or trying to push the frontier of code or reasoning models is really interesting to us. And then also interested in the frontier of modeling. There's a lot of models that work in, like, pixel space, as we call it. So if you're doing world models, video models, even robotics, where there's not a very clean natural language interface to interact with, I think we think that Interp can really help and are looking for a few partners in that space.Shawn Wang [00:52:43]: Just because you mentioned the keyword

In this episode, Justin and Rob discuss captive equivalence as a welfare standard w/ Billy SveenWho will win? You decide. Reptile Fight Club!Billy Sveen-https://www.instagram.com/creepersherpetoculture/Billy's paperhttps://static1.squarespace.com/static/5f99c047fa9ceb071696ee9a/t/67ec2f45295d144305dc58f1/1743531845626/V7.1_001_Opinion_Sveen.pdf?utm_source=ig&utm_medium=social&utm_content=link_in_bioFollow Justin Julander @Australian Addiction Reptiles-http://www.australianaddiction.comIG https://www.instagram.com/jgjulander/Follow Rob @ https://www.instagram.com/highplainsherp/Follow MPR Network @FB: https://www.facebook.com/MoreliaPythonRadioIG: https://www.instagram.com/mpr_network/YouTube: https://www.youtube.com/channel/UCtrEaKcyN8KvC3pqaiYc0RQSwag store: https://teespring.com/stores/mprnetworkPatreon: https://www.patreon.com/moreliapythonradio

I am Rolf Claessen and together with my co-host Ken Suzan I am welcoming you to episode 169 of our podcast IP Fridays! Today's interview guest is Prof. Aloys Hüttermann, co-founder of my patent law firm Michalski Hüttermann & Partner and a true expert on the Unified Patent Court. He has written several books about the new system and we talk about all the things that plaintiffs and defendants can learn from the first decisions of the court and what they mean for strategic decisions of the parties involved. But before we jump into this very interesting interview, I have news for you! The US Patent and Trademark Office (USPTO) is planning rule changes that would make it virtually impossible for third parties to challenge invalid patents before the patent office. Criticism has come from the EFF and other inventor rights advocates: the new rules would play into the hands of so-called non-practicing entities (NPEs), as those attacked would have few cost-effective ways to have questionable patents deleted. The World Intellectual Property Organization (WIPO) reports a new record in international patent applications: in 2024, around 3.7 million patent applications were filed worldwide – an increase of 4.9% over the previous year. The main drivers were Asian countries (China alone accounted for 1.8 million), while demand for trademark protection has stabilized after the pandemic decline. US rapper Eminem is taking legal action in Australia against a company that sells swimwear under the name “Swim Shady.” He believes this infringes on his famous “Slim Shady” brand. The case illustrates that even humorous allusions to well-known brand names can lead to legal conflicts. A new ruling by the Unified Patent Court (UPC) demonstrates its cross-border impact. In “Fujifilm v. Kodak,” the local chamber in Mannheim issued an injunction that extends to the UK despite Brexit. The UPC confirmed its jurisdiction over the UK parts of a European patent, as the defendant Kodak is based in a UPC member state. A dispute over standard patents is looming at the EU level: the Legal Affairs Committee (JURI) of the European Parliament voted to take the European Commission to the European Court of Justice. The reason for this is the Commission’s controversial withdrawal of a draft regulation on the licensing of standard-essential patents (SEPs). Parliament President Roberta Metsola is to decide by mid-November whether to file the lawsuit. In trademark law, USPTO Director Squires reported on October 31, 2025, that a new unit (“Trademark Registration Protection Office”) had removed approximately 61,000 invalid trademark applications from the registries. This cleanup of the backlog relieved the examining authority and accelerated the processing of legitimate applications. Now let's jump into the interview with Aloys Hüttermann: The Unified Patent Court Comes of Age – Insights from Prof. Aloys Hüttermann The Unified Patent Court (UPC) has moved from a long-discussed project to a living, breathing court system that already shapes patent enforcement in Europe. In a recent IP Fridays interview, Prof. Aloys Hüttermann – founder and equity partner at Michalski · Hüttermann & Partner and one of the earliest commentators on the UPC – shared his experiences from the first years of practice, as well as his view on how the UPC fits into the global patent litigation landscape. This article summarises the key points of that conversation and is meant as an accessible overview for in-house counsel, patent attorneys and business leaders who want to understand what the UPC means for their strategy. How Prof. Hüttermann Became “Mr. UPC” Prof. Hüttermann has been closely involved with the UPC for more than a decade. When it became clear, around 13 years ago, that the European project of a unified patent court and a unitary patent was finally going to happen, he recognised that this would fundamentally change patent enforcement in Europe. He started to follow the legislative and political developments in detail and went beyond mere observation. As author and editor of several books and a major commentary on the UPC, he helped shape the discussion around the new system. His first book on the UPC appeared in 2016 – years before the court finally opened its doors in 2023. What fascinated him from the beginning was the unique opportunity to witness the creation of an entirely new court system, to analyse how it would be built and, where possible, to contribute to its understanding and development. It was clear to him that this system would be a “game changer” for European patent enforcement. UPC in the Global Triangle: Europe, the US and China In practice, most international patent disputes revolve around three major regions: the UPC territory in Europe, the United States and China. Each of these regions has its own procedural culture, cost structure and strategic impact. From a territorial perspective, the UPC is particularly attractive because it can, under the right conditions, grant pan-European injunctions that cover a broad range of EU Member States with a single decision. This consolidation of enforcement is something national courts in Europe simply cannot offer. From a cost perspective, the UPC is significantly cheaper than US litigation, especially if one compares the cost of one UPC action with a bundle of separate national cases in large European markets. When viewed against the territorial reach and procedural speed, the “bang for the buck” is very compelling. China is again a different story. The sheer volume of cases there is enormous, with tens of thousands of patent infringement cases per year. Chinese courts are known for their speed; first-instance decisions within about a year are common. In this respect they resemble the UPC more than the US does. The UPC also aims at a roughly 12 to 15 month time frame for first-instance cases where validity is at issue. The US, by contrast, features extensive discovery, occasionally jury trials and often longer timelines. The procedural culture is very different. The UPC, like Chinese courts, operates without discovery in the US sense, which makes proceedings more focused on the written record and expert evidence that the parties present, and less on pre-trial disclosure battles. Whether a company chooses to litigate in the US, the UPC, China, or some combination of these forums will depend on where the key markets and assets are. However, in Prof. Hüttermann's view, once Europe is an important market, it is hard to justify ignoring the UPC. He expects the court's caseload and influence to grow strongly over the coming years. A Landmark UPC Case: Syngenta v. Sumitomo A particularly important case in which Prof. Hüttermann was involved is the Syngenta v. Sumitomo matter, concerning a composition patent. This case has become a landmark in UPC practice for several reasons. First, the Court of Appeal clarified a central point about the reach of UPC injunctions. It made clear that once infringement is established in one Member State, this will usually be sufficient to justify a pan-European injunction covering all UPC countries designated by the patent. That confirmation gave patent owners confidence that the UPC can in fact deliver broad, cross-border relief in one go. Second, the facts of the case raised novel issues about evidence and territorial reach. The allegedly infringing product had been analysed based on a sample from the Czech Republic, which is not part of the UPC system. Later, the same product with the same name was marketed in Bulgaria, which is within UPC territory. The Court of Appeal held that the earlier analysis of the Czech sample could be relied on for enforcement in Bulgaria. This showed that evidence from outside the UPC territory can be sufficient, as long as it is properly linked to the products marketed within the UPC. Third, the Court of Appeal took the opportunity to state its view on inventive step. It confirmed that combining prior-art documents requires a “pointer”, in line with the EPO's problem-solution approach. The mere theoretical possibility of extracting a certain piece of information from a document does not suffice to justify an inventive-step attack. This is one of several decisions where the UPC has shown a strong alignment with EPO case law on substantive patentability. For Prof. Hüttermann personally, the case was also a lesson in oral advocacy before the UPC. During the two appeal hearings, the presiding judge asked unexpected questions that required quick and creative responses while the hearing continued. His practical takeaway is that parties should appear with a small, well-coordinated team: large enough to allow someone to work on a tricky question in the background, but small enough to remain agile. Two or three lawyers seem ideal; beyond that, coordination becomes difficult and “too many cooks spoil the broth”. A Game-Changing CJEU Decision: Bosch Siemens Hausgeräte v. Electrolux Surprisingly, one of the most important developments for European patent litigation in the past year did not come from the UPC at all, but from the Court of Justice of the European Union. In Bosch Siemens Hausgeräte v. Electrolux, the CJEU revisited the rules on cross-border jurisdiction under the Brussels I Recast Regulation (Brussels Ia). Previously, under what practitioners often referred to as the GAT/LuK regime, a court in one EU country was largely prevented from granting relief for alleged infringement in another country if the validity of the foreign patent was contested there. This significantly limited the possibilities for cross-border injunctions. In Bosch, the CJEU changed course. Without going into all procedural details, the essence is that courts in the EU now have broader powers to grant cross-border relief when certain conditions are met, particularly when at least one defendant is domiciled in the forum state. The concept of an “anchor defendant” plays a central role: if you sue one group company in its home forum, other group companies in other countries, including outside the EU, can be drawn into the case. This has already had practical consequences. German courts, for example, have issued pan-European injunctions covering around twenty countries in pharmaceutical cases. There are even attempts to sue European companies for infringement of US patents based on acts in the US, using the logic of Bosch as a starting point. How far courts will ultimately go remains to be seen, but the potential is enormous. For the UPC, this development is highly relevant. The UPC operates in the same jurisdictional environment as national courts, and many defendants in UPC cases will be domiciled in UPC countries. This increases the likelihood that the UPC, too, can leverage the broadened possibilities for cross-border relief. In addition, we have already seen UPC decisions that include non-EU countries such as the UK within the scope of injunctions, in certain constellations. The interaction between UPC practice and the Bosch jurisprudence of the CJEU is only beginning to unfold. Does the UPC Follow EPO Case Law? A key concern for many patent owners and practitioners is whether the UPC will follow the EPO's Boards of Appeal or develop its own, possibly divergent, case law on validity. On procedural matters, the UPC is naturally different from the EPO. It has its own rules of procedure, its own timelines and its own tools, such as “front-loaded” pleadings and tight limits on late-filed material. On substantive law, however, Prof. Hüttermann's conclusion is clear: there is “nothing new under the sun”. The UPC's approach to novelty, inventive step and added matter is very close to that of the EPO. The famous “gold standard” for added matter appears frequently in UPC decisions. Intermediate generalisations are treated with the same suspicion as at the EPO. In at least one case, the UPC revoked a patent for added matter even though the EPO had granted it in exactly that form. The alignment is not accidental. The UPC only deals with European patents granted by the EPO; it does not hear cases on purely national patents. If the UPC were more generous than the EPO, many patents would never reach it. If it were systematically stricter, patentees would be more tempted to opt out of the system. In practice, the UPC tends to apply the EPO's standards and, where anything differs, it is usually a matter of factual appreciation rather than a different legal test. For practitioners, this has a very practical implication: if you want to predict how the UPC will decide on validity, the best starting point is to ask how the EPO would analyse the case. The UPC may not always reach the same result in parallel EPO opposition proceedings, but the conceptual framework is largely the same. Trends in UPC Practice: PIs, Equivalents and Division-Specific Styles Even in its early years, certain trends and differences between UPC divisions can be observed. On preliminary injunctions, the local division in Düsseldorf has taken a particularly proactive role. It has been responsible for most of the ex parte PIs granted so far and applies a rather strict notion of urgency, often considering one month after knowledge of the infringement as still acceptable, but treating longer delays with scepticism. Other divisions tend to see two months as still compatible with urgency, and they are much more cautious with ex parte measures. Munich, by contrast, has indicated a strong preference for inter partes PI proceedings and appears reluctant to grant ex parte relief at all. A judge from Munich has even described the main action as the “fast” procedure and the inter partes PI as the “very fast” one, leaving little room for an even faster ex parte track. There are also differences in how divisions handle amendments and auxiliary requests in PI proceedings. Munich has suggested that if a patentee needs to rely on claim amendments or auxiliary requests in a PI, the request is unlikely to succeed. Other divisions have been more open to considering auxiliary requests. The doctrine of equivalents is another area where practice is not yet harmonised. The Hague division has explicitly applied a test taken from Dutch law in at least one case and found infringement by equivalence. However, the Court of Appeal has not yet endorsed a specific test, and in another recent Hague case the same division did not apply that Dutch-law test again. The Mannheim division has openly called for the development of an autonomous, pan-European equivalence test, but has not yet fixed such a test in a concrete decision. This is clearly an area to watch. Interim conferences are commonly used in most divisions to clarify issues early on, but Düsseldorf often dispenses with them to save time. In practice, interim conferences can be very helpful for narrowing down the issues, though parties should not expect to be able to predict the final decision from what is discussed there. Sometimes topics that dominate the interim conference play little or no role in the main oral hearing. A Front-Loaded System and Typical Strategic Mistakes UPC proceedings are highly front-loaded and very fast. A defendant usually has three months from service of the statement of claim to file a full statement of defence and any counterclaim for revocation. This is manageable, but only if the time is used wisely. One common strategic problem is that parties lose time at the beginning and only develop a clear strategy late in the three-month period. According to Prof. Hüttermann, it is crucial to have a firm strategy within the first two or three weeks and then execute it consistently. Constantly changing direction is a recipe for failure in such a compressed system. Another characteristic is the strict attitude towards late-filed material. It is difficult to introduce new documents or new inventive-step attacks later in the procedure. In some cases even alternative combinations of already-filed prior-art documents have been viewed as “new” attacks and rejected as late. At the appeal stage, the Court of Appeal has even considered new arguments based on different parts of a book already in the file as potentially late-filed. This does not mean that parties should flood the court with dozens of alternative attacks in the initial brief. In one revocation action, a plaintiff filed about fifty different inventive-step attacks, only to be told by the court that this was not acceptable and that the attacks had to be reduced and structured. The UPC is not a body conducting ex officio examination. It is entitled to manage the case actively and to ask parties to focus on the most relevant issues. Evidence Gathering, Protective Letters and the Defendant's Perspective The UPC provides powerful tools for both sides. Evidence inspection is becoming more common, not only at trade fairs but also at company premises. This can be a valuable tool for patentees, but it also poses a serious risk for defendants who may suddenly face court-ordered inspections. From the perspective of potential defendants, protective letters are an important instrument, especially in divisions like Düsseldorf where ex parte PIs are possible. A well-written protective letter, filed in advance, can significantly reduce the risk of a surprise injunction. The court fees are moderate, but the content of the protective letter must be carefully prepared; a poor submission can cause more harm than good. Despite the strong tools available to patentees, Prof. Hüttermann does not view the UPC as unfair to defendants. If a defendant files a solid revocation counterclaim, the pressure shifts to the patentee, who then has only two months to reply, prepare all auxiliary requests and adapt the enforcement strategy. This is even more demanding than at the EPO, because the patentee must not only respond to validity attacks but also ensure that any amended claims still capture the allegedly infringing product. It is entirely possible to secure the survival of a patent with an auxiliary request that no longer covers the defendant's product. In that scenario, the patentee has “won” on validity but lost the infringement case. Managing this tension under tight time limits is a key challenge of UPC practice. The Future Role of the UPC and How to Prepare Today the UPC hears a few hundred cases per year, compared with several thousand patent cases in the US and tens of thousands in China. Nevertheless, both the court itself and experienced practitioners see significant growth potential. Prof. Hüttermann expects case numbers to multiply in the medium term. Whether the UPC will become the first choice forum in global disputes or remain one pillar in parallel proceedings alongside the US and China will depend on the strategies of large patentees and the evolution of case law. However, the court is well equipped: it covers a large, economically important territory, is comparatively cost-effective and offers fast procedures with robust remedies. For companies that may end up before the UPC, preparation is essential. On the offensive side, that means building strong evidence and legal arguments before filing, being ready to proceed quickly and structured, and understanding the specific styles of the relevant divisions. On the defensive side, it may mean filing protective letters in risk-exposed markets, preparing internal processes for rapid reaction if a statement of claim arrives, and taking inspection requests seriously. Conclusion The Unified Patent Court has quickly moved from theory to practice. It offers pan-European relief, fast and front-loaded procedures, and a substantive approach that closely mirrors the EPO's case law. At the same time, national and EU-level developments like the Bosch Siemens Hausgeräte v. Electrolux decision are reshaping the jurisdictional framework in which the UPC operates, opening the door for far-reaching cross-border injunctions. For patent owners and potential defendants alike, the message is clear: the UPC is here to stay and will become more important year by year. Those who invest the time to understand its dynamics now – including its alignment with the EPO, the differences between divisions, and the strategic implications of its procedures – will be in a much better position when the first UPC dispute lands on their desk. Here is the full transcript of the interview: Rolf Claessen:Today's interview guest is Prof. Aloys Hüttermann. He is founder and equity partner of my firm, Michalski · Hüttermann & Partner. More importantly for today's interview, he has written several books about the Unified Patent Court. The first one already came out in 2016. He is co-editor and author of one of the leading commentaries on the UPC and has gained substantial experience in UPC cases so far – one of them even together with me. Thank you very much for being on IP Fridays again, Aloys. Aloys Hüttermann:Thank you for inviting me, it's an honour. How did you get so deeply involved in the UPC? Rolf Claessen:Before we dive into the details, how did you end up so deeply involved in the Unified Patent Court? And what personally fascinates you about this court? Aloys Hüttermann:This goes back quite a while – roughly 13 years. At that time it became clear that, after several failed attempts, Europe would really get a pan-European court and a pan-European patent, and that this time it was serious. I thought: this is going to be the future. That interested me a lot, both intellectually and practically. A completely new system was being built. You could watch how it evolved – and, if possible, even help shape it a bit. It was also obvious to me that this would be a complete game changer. Nobody expected that it would take until 2023 before the system actually started operating, but now it is here. I became heavily interested early on. As you mentioned, my first book on the UPC was published in 2016, in the expectation that the system would start soon. It took a bit longer, but now we finally have it. UPC vs. US and China – speed, cost and impact Rolf Claessen:Before we go deeper into the UPC, let's zoom out. If you compare litigation before the UPC with patent litigation in the US and in China – in terms of speed, cost and the impact of decisions – what are the key differences that a business leader should understand? Aloys Hüttermann:If you look at the three big regions – the UPC territory in Europe, the US and China – these are the major economic areas for many technology companies. One important point is territorial reach. In the UPC, if the conditions are met, you can get pan-European injunctions that cover many EU Member States in one go. We will talk about this later in more detail. On costs there is a huge difference between the US and the UPC. The UPC is much cheaper than US litigation, especially once you look at the number of countries you can cover with one case if the patent has been validated widely. China is different again. The number of patent infringement cases there is enormous. I have seen statistics of around 40,000 infringement cases per year in China. That is huge – compared with roughly 164 UPC infringement cases in the first year and maybe around 200 in the current year. On speed, Chinese courts are known to be very fast. You often get a first-instance decision in about a year. The UPC is comparable: if there is a counterclaim for revocation, you are looking at something like 12 to 15 months for a first-instance decision. The US can be slower, and the procedure is very different. You have full discovery, you may have juries. None of that exists at the UPC. From that perspective, Chinese and UPC proceedings are more similar to each other than either is to the US. The UPC is still a young court. We have to see how influential its case law will be worldwide in the long run. What we already see, at least in Germany, is a clear trend away from purely national patent litigation and towards the UPC. That is inside Europe. The global impact will develop over time. When is the UPC the most powerful tool? Rolf Claessen:Let's take the perspective of a global company. It has significant sales in Europe and in the US and production or key suppliers in China. In which situations would you say the UPC is your most powerful tool? And when might the US or China be the more strategic battleground? Aloys Hüttermann:To be honest, I would almost always consider bringing a case before the UPC. The “bang for the buck” is very good. The UPC is rather fast. That alone already gives you leverage in negotiations. The threat of a quick, wide-reaching injunction is a strong negotiation tool. Whether you litigate in the US instead of the UPC, or in addition, or whether you also go to China – that depends heavily on the individual case: where the products are sold, where the key markets are, where the defendant has assets, and so on. But in my view, once you have substantial sales in Europe, you should seriously consider the UPC. And for that reason alone I expect case numbers at the UPC to increase significantly in the coming years. A landmark UPC case: Syngenta vs. Sumitomo (composition patent) Rolf Claessen:You have already been involved in several UPC cases – and one of them together with me, which was great fun. Looking at the last 12 to 18 months, is there a case, decision or development that you find particularly noteworthy – something that really changed how you think about UPC litigation or how companies should prepare? Aloys Hüttermann:The most important UPC case I have been involved in so far is the Syngenta v. Sumitomo case on a composition patent. It has become a real landmark and was even mentioned in the UPC's annual report. It is important for several reasons. First, it was one of the first cases in which the Court of Appeal said very clearly: if you have established infringement in one Member State, that will usually be enough for a pan-European injunction covering all UPC countries designated by the patent. That is a powerful statement about the reach of UPC relief. Second, the facts were interesting. The patent concerned a composition. We had analysed a sample that had been obtained in the Czech Republic, which is not a UPC country. Later, the same product was marketed under the same name in Bulgaria, which is in the UPC. The question was whether the analysis of the Czech sample could be used as a basis for enforcement in Bulgaria. The Court of Appeal said yes, that was sufficient. Third, the Court of Appeal took the opportunity to say something about inventive step. It more or less confirmed that the UPC's approach is very close to the EPO's problem-solution approach. It emphasised that, if you want to combine prior-art documents, you need a “pointer” to do so. The mere theoretical possibility that a skilled person could dig a particular piece of information out of a document is not enough. For me personally, the most memorable aspect of this case was not the outcome – that was largely in line with what we had expected – but the oral hearings at the appeal stage. We had two hearings. In both, the presiding judge asked us a question that we had not anticipated at all. And then you have about 20 minutes to come up with a convincing answer while the hearing continues. We managed it, but it made me think a lot about how you should prepare for oral hearings at the UPC. My conclusion is: you should go in with a team, but not too big. In German we say, “Zu viele Köche verderben den Brei” – too many cooks spoil the broth. Two or three people seems ideal. One of them can work quietly on such a surprise question at the side, while the others continue arguing the case. In the end the case went very well for us, so I can speak about it quite calmly now. But in the moment your heart rate definitely goes up. The CJEU's Bosch Siemens Hausgeräte v. Electrolux decision – a real game changer Rolf Claessen:You also mentioned another development that is not even a UPC case, but still very important for European patent litigation. Aloys Hüttermann:Yes. In my view, the most important case of the last twelve months is not a UPC decision but a judgment of the Court of Justice of the EU (CJEU): Bosch Siemens Hausgeräte v. Electrolux. This is going to be a real game changer for European IP law, and I am sure we have not seen the end of its effects yet. One example: someone has recently sued BMW before the Landgericht München I, a German court, for infringement of a US patent based on acts in the US. The argument is that this could be backed by the logic of Bosch Siemens Hausgeräte v. Electrolux. We do not know yet what the court will do with that, but the fact that people are trying this shows how far-reaching the decision might be. Within the UPC we have already seen injunctions being issued for countries outside the UPC territory and even outside the EU, for example including the UK. So you see how these developments start to interact. Rolf Claessen:For listeners who have not followed the case so closely: in very simple terms, the CJEU opened the door for courts in one EU country to rule on patent infringement that took place in other countries as well, right? Aloys Hüttermann:Exactly. Before Bosch Siemens Hausgeräte v. Electrolux we had what was often called the GAT/LuK regime. The basic idea was: if you sue someone in, say, Germany for infringement of a European patent, and you also ask for an injunction for France, and the defendant then challenges the validity of the patent in France, the German court cannot grant you an injunction covering France. The Bosch decision changed that. The legal basis is the Brussels I Recast Regulation (Brussels Ia), which deals with jurisdiction in civil and commercial matters in the EU. It is not specific to IP; it applies to civil cases generally, but it does have some provisions that are relevant for patents. In Bosch, a Swedish court asked the CJEU for guidance on cross-border injunctions. The CJEU more or less overturned its old GAT/LuK case law. Now, in principle, if the defendant is domiciled in a particular Member State, the courts of that state can also grant cross-border relief for other countries, under certain conditions. We will not go into all the details here – that could fill a whole separate IP Fridays episode – but one important concept is the “anchor defendant”. If you sue a group of companies and at least one defendant is domiciled in the forum state, then other group companies in other countries – even outside the EU, for example in Hong Kong – can be drawn into the case and affected by the decision. This is not limited to the UPC, but of course it is highly relevant for UPC litigation. Statistically it increases the chances that at least one defendant will be domiciled in a UPC country, simply because there are many of them. And we have already seen courts like the Landgericht München I grant pan-European injunctions for around 20 countries in a pharmaceutical case. Rolf Claessen:Just to clarify: does it have to be the headquarters of the defendant in that country, or is any registered office enough? Aloys Hüttermann:That is one of the open points. If the headquarters are in Europe, then it is clear that subsidiaries outside Europe can be affected as well. If the group's headquarters are outside Europe and only a subsidiary is here, the situation is less clear and we will have to see what the courts make of it. Does the UPC follow EPO case law? Rolf Claessen:Many patent owners and in-house counsel wonder: does the UPC largely follow the case law of the EPO Boards of Appeal, or is it starting to develop its own distinct line? What is your impression so far – both on substantive issues like novelty and inventive step, and on procedural questions? Aloys Hüttermann:On procedure the UPC is, of course, very different. It has its own procedural rules and they are not the same as at the EPO. If we look at patent validity, however, my impression is that there is “nothing new under the sun” – that was the title of a recent talk I gave and will give again in Hamburg. Substantively, the case law of the UPC and the EPO is very similar. For inventive step, people sometimes say the UPC does not use the classical problem-solution approach but a more “holistic” approach – whatever that is supposed to mean. In practice, in both systems you read and interpret prior-art documents and decide what they really disclose. In my view, the “error bar” that comes from two courts simply reading a document slightly differently is much larger than any systematic difference in legal approach. If you look at other grounds, such as novelty and added matter, the UPC even follows the EPO almost verbatim. The famous “gold standard” for added matter appears all over UPC decisions, even if the EPO case numbers are not always cited. The same is true for novelty. So the rule-based, almost “Hilbertian” EPO approach is very much present at the UPC. There is also a structural reason for that. All patents that the UPC currently deals with have been granted by the EPO. The UPC does not handle patents granted only by national offices. If the UPC wanted to deviate from EPO case law and be more generous, then many patents would never reach the UPC in the first place. The most generous approach you can have is the one used by the granting authority – the EPO. So if the UPC wants to be different, it can only be stricter, not more lenient. And there is little incentive to be systematically stricter, because that would reduce the number of patents that are attractive to enforce before the UPC. Patent owners might simply opt out. Rolf Claessen:We also talked about added matter and a recent case where the Court of Appeal was even stricter than the EPO. That probably gives US patent practitioners a massive headache. They already struggle with added-matter rules in Europe, and now the UPC might be even tougher. Aloys Hüttermann:Yes, especially on added matter. I once spoke with a US practitioner who said, “We hope the UPC will move away from intermediate generalisations.” There is no chance of that. We already have cases where the Court of Appeal confirmed that intermediate generalisations are not allowed, in full alignment with the EPO. You mentioned a recent case where a patent was revoked for added matter, even though it had been granted by the EPO in exactly that form. This shows quite nicely what to expect. If you want to predict how the UPC will handle a revocation action, the best starting point is to ask: “What would the EPO do?” Of course, there will still be cases where the UPC finds an invention to be inventive while the EPO, in parallel opposition proceedings, does not – or vice versa. But those are differences in the appreciation of the facts and the prior art, which you will always have. The underlying legal approach is essentially the same. Rolf Claessen:So you do not see a real example yet where the UPC has taken a totally different route from the EPO on validity? Aloys Hüttermann:No, not really. If I had to estimate how the UPC will decide, I would always start from what I think the EPO would have done. Trends in UPC practice: PIs, equivalents, interim conferences Rolf Claessen:If you look across the different UPC divisions and cases: what trends do you see in practice? For example regarding timelines, preliminary injunctions, how validity attacks are handled, and how UPC cases interact with EPO oppositions or national proceedings? Aloys Hüttermann:If you take the most active divisions – essentially the big four in Germany and the local division in The Hague – they all try to be very careful and diligent in their decisions. But you can already see some differences in practice. For preliminary injunctions there is a clear distinction between the local division in Düsseldorf and most other divisions. Düsseldorf considers one month after knowledge of the infringement as still sufficiently urgent. If you wait longer, it is usually considered too late. In many other divisions, two months is still viewed as fine. Düsseldorf has also been the division that issued most of the ex parte preliminary injunctions so far. Apart from one special outlier where a standing judge from Brussels was temporarily sitting in Milan, Düsseldorf is basically the only one. Other divisions have been much more reluctant. At a conference, Judge Pichlmaier from the Munich division once said that he could hardly imagine a situation where his division would grant an ex parte PI. In his words, the UPC has two types of procedure: one that is fast – the normal main action – and one that is very fast – the inter partes PI procedure. But you do not really have an “ultra-fast” ex parte track, at least not in his division. Another difference relates to amendments and auxiliary requests in PI proceedings. In one recent case in Munich the court said more or less that if you have to amend your patent or rely on auxiliary requests in a PI, you lose. Other divisions have been more flexible and have allowed auxiliary requests. Equivalence is another area where we do not have a unified line yet. So far, only the Hague division has clearly found infringement under the doctrine of equivalents and explicitly used a test taken from Dutch law. Whether that test will be approved by the Court of Appeal is completely open – the first case settled, so the Court of Appeal never ruled on it, and a second one is still very recent. Interestingly, there was another Hague decision a few weeks ago where equivalence was on the table, but the division did not apply that Dutch-law test. We do not know yet why. The Mannheim division has written in one decision that it would be desirable to develop an autonomous pan-European test for equivalence, instead of just importing the German, UK or Dutch criteria. But they did not formulate such a test in that case because it was not necessary for the decision. So we will have to see how that evolves. On timelines, one practical difference is that Düsseldorf usually does not hold an interim conference. That saves them some time. Most other divisions do hold interim conferences. Personally, I like the idea because it can help clarify issues. But you cannot safely read the final outcome from these conferences. I have also seen cases where questions raised at the interim conference did not play any role in the main oral hearing. So they are useful for clarification, but not as a crystal ball. Front-loaded proceedings and typical strategic mistakes Rolf Claessen:If you look at the behaviour of parties so far – both patentees and defendants – what are the most common strategic mistakes you see in UPC litigation? And what would a well-prepared company do differently before the first statement of claim is ever filed? Aloys Hüttermann:You know you do not really want me to answer that question… Rolf Claessen:I do! Aloys Hüttermann:All right. The biggest mistake, of course, is that they do not hire me. That is the main problem. Seriously, it is difficult to judge parties' behaviour from the outside. You rarely know the full picture. There may be national proceedings, licensing discussions, settlement talks, and so on in the background. That can limit what a party can do at the UPC. So instead of criticising, I prefer to say what is a good idea at the UPC. The system is very front-loaded and very fast. If you are sued, you have three months to file your statement of defence and your counterclaim for revocation. In my view, three months are manageable – but only if you use the time wisely and do not waste it on things that are not essential. If you receive a statement of claim, you have to act immediately. You should have a clear strategy within maybe two or three weeks and then implement it. If you change your strategy every few weeks, chances are high that you will fail. Another point is that everything is front-loaded. It is very hard to introduce new documents or new attacks later. Some divisions have been a bit generous in individual cases, but the general line is strict. We have seen, for example, that even if you filed a book in first instance, you may not be allowed to rely on a different chapter from the same book for a new inventive-step attack at the appeal stage. That can be regarded as late-filed, because you could have done it earlier. There is also case law saying that if you first argue inventive step as “D1 plus D2”, and later want to argue “D2 plus D1”, that can already be considered a new, late attack. On the other hand, we had a revocation action where the plaintiff filed about 50 different inventive-step attacks in the initial brief. The division then said: this does not work. Please cut them down or put them in a clear hierarchy. In the end, not all of them were considered. The UPC does not conduct an ex officio examination. It is entitled to manage the case and to tell the parties to limit themselves in the interest of a fair and efficient procedure. Rolf Claessen:I have the feeling that the EPO is also becoming more front-loaded – if you want to rely on documents later, you should file them early. But it sounds like the UPC is even more extreme in that regard. Aloys Hüttermann:Yes, that is true. Protective letters, inspections and the defendant's perspective Rolf Claessen:Suppose someone from a company is listening now and thinks: “We might be exposed at the UPC,” or, “We should maybe use the UPC offensively against competitors.” What would you consider sensible first steps before any concrete dispute arises? And looking three to five years ahead, how central do you expect the UPC to become in global patent litigation compared to the US and China? Aloys Hüttermann:Let me start with the second part. I expect the UPC to become significantly more important. If we have around 200 cases this year, that is a good start, but it is still very small compared to, say, 4,000 to 5,000 patent cases per year in the US and 40,000 or so in China. Even François Bürgin and Klaus Grabinski, in interviews, have said that they are happy with the case load, but the potential is much larger. In my view, it is almost inevitable that we will see four or five times as many UPC cases in the not-too-distant future. As numbers grow, the influence of the UPC will grow as well. Whether, in five or ten years, companies will treat the UPC as their first choice forum – or whether they will usually run it in parallel with US litigation in major disputes – remains to be seen. The UPC would be well equipped for that: the territory it covers is large, Europe is still an important economy, and the UPC procedure is very attractive from a company's perspective. On sensible first steps: if you are worried about being sued, a protective letter can make a lot of sense – especially in divisions like Düsseldorf, where ex parte PIs are possible in principle. A protective letter is not very expensive in terms of court fees. There is also an internal system that ensures the court reads it before deciding on urgent measures. Of course, the content must have a certain quality; a poor protective letter can even backfire. If you are planning to sue someone before the UPC, you should be extremely well prepared when you file. You should already have all important documents and evidence at hand. As we discussed, it is hard to introduce new material later. One tool that is becoming more and more popular is inspection – not just at trade fairs, where we already saw cases very early, but also at company premises. Our firm has already handled such an inspection case. That is something you should keep in mind on both sides: it is a powerful evidence-gathering tool, but also a serious risk if you are on the receiving end. From the defendant's perspective, I do not think the UPC is unfair. If you do your job properly and put a solid revocation counterclaim on the table, then the patentee has only two months to prepare a full reply and all auxiliary requests. And there is a twist that makes life even harder for the patentee than at the EPO. At the EPO the question is mainly: do my auxiliary requests overcome the objections and are they patentable? At the UPC there is an additional layer: do I still have infringement under the amended claims? You may save your patent with an auxiliary request that no longer reads on the defendant's product. That is great for validity, but you have just lost the infringement case. You have kept the patent but lost the battle. And all of this under very tight time limits. That creates considerable pressure on both sides. How to contact Prof. Hüttermann Rolf Claessen:Thank you very much for this really great interview, Aloys. Inside our firm you have a nickname: “the walking encyclopedia of the Unified Patent Court” – because you have written so many books about it and have dealt with the UPC for such a long time. What is the best way for listeners to get in touch with you? Aloys Hüttermann:The easiest way is by email. You can simply write to me, and that is usually the best way to contact me. As you may have noticed, I also like to speak. I am a frequent speaker at conferences. If you happen to be at one of the conferences where I am on the programme – for example, next week in Hamburg – feel free to come up to me and ask me anything in person. But email is probably the most reliable first step. Rolf Claessen:Perfect. Thank you very much, Aloys. Aloys Hüttermann:Thank you. It was a pleasure to be on IP Fridays again. Some of your long-time listeners may remember that a few years ago – when you were not yet part of our firm – we already did an episode on the UPC, back when everything was still very speculative. It is great to be back now that the system is actually in place and working. Rolf Claessen:I am very happy to have you back on the show.

This talk was given by Gil Fronsdal on 2025.11.07 at the IMC Programs in IMC. ******* Equivalence of Ethics and Enlightenment (2025-10-01 00:00:00 -0700) ******* A machine generated transcript of this talk is available. It has not been edited by a human, so errors will exist. Download Transcript: https://www.audiodharma.org/transcripts/24193/download ******* For more talks like this, visit AudioDharma.org ******* If you have enjoyed this talk, please consider supporting AudioDharma with a donation at https://www.audiodharma.org/donate/. ******* This talk is licensed by a Creative Commons Attribution-Noncommercial-No Derivative Works 4.0 License

In this podcast, Nicolas Badre, MD, and Eric Geier, MD, PhD, discuss their article, "Anticholinergic Equivalence in Psychotropic Medications: A Guide for Psychiatrists," which is published in the November-December 2025 issue of the Journal of Clinical Psychopharmacology. Anticholinergic side effects from psychotropic medications are common and can lead to significant adverse events, including cognitive impairment and falls, particularly in vulnerable populations like the elderly. The cumulative anticholinergic burden from multiple medications is a critical concern associated with poorer clinical outcomes. Quantifying this burden is essential for safer prescribing. For their article, they developed a table to provide a practical tool for psychiatrists to quantify and compare the anticholinergic potential of psychotropic medications. doi: 10.1097/JCP.0000000000002073

Send us a textSupport the show⬇️ Subscribe on Patreon or Substack for full episodes ⬇️https://www.patreon.com/HughHendryhttps://hughhendry.substack.comhttps://www.instagram.com/hughhendryofficialhttps://blancbleustbarts.comhttps://www.instagram.com/blancbleuofficial⭐⭐⭐⭐⭐ Leave a five star review and comment on Apple Podcasts!

In this special episode, I take on probably the most controversial and emotionally fraught topic of the moment: the Israel-Hamas conflict.  I think war crimes have been committed on both sides. But that doesn't mean I think the two sides are morally equivalent. Today, I argue that there's a fundamental asymmetry between Israel and Hamas, one that's too often blurred or ignored by the mainstream media. Israel's actions, while sometimes flawed or tragic in consequence, are ultimately rooted in a defensive logic. Hamas, on the other hand, has explicitly genocidal goals. But where does that leave us when we see images of children starving and hear reports that Israel is responsible? Learn more about your ad choices. Visit megaphone.fm/adchoices

How can Problem Strings help build big ideas? In this episode Pam and Kim walk through a Problem String that helps students dig into area and multiplicative equivalence and helps you know how to expertly facilitate.Talking Points:Facilitating the stringTeacher moves during the stringModeling the stringEqual products means total area is the sameWhy we want students to have lots of experience with concepts rather than just direct teaching.Check out our social mediaTwitter: @PWHarrisInstagram: Pam Harris_mathFacebook: Pam Harris, author, mathematics educationLinkedin: Pam Harris Consulting LLC 

How does a Problem String look in front of real students? In this episode Pam and Kim give a play by play for how a Problem String could be facilitated.Talking Points:When to circulate and when to ask for choral responseHelping students communicate thinkingHow and when to engage students in conversationsWhen to anchor strategiesWhen to be intentionally curious to solidify thinkingCheck out our grade level Problem String books!Grade 1 Problem Strings: https://www.mathisfigureoutable.com/nps-1Grade 2 Problem Strings: https://www.mathisfigureoutable.com/NPS-2Grade 3 Problem Strings: https://www.mathisfigureoutable.com/nps-3Grade 4 Problem Strings: https://www.mathisfigureoutable.com/nps-4Grade 5 Problem Strings: https://www.mathisfigureoutable.com/nps-5Check out our social mediaTwitter: @PWHarrisInstagram: Pam Harris_mathFacebook: Pam Harris, author, mathematics educationLinkedin: Pam Harris Consulting LLC 

Many enter marriage believing it will be effortless—a seamless fusion of two souls destined for harmony. But what if this belief is the very thing setting couples up for disappointment? What if the key to a thriving marriage lies not in ease, but in intentional effort and understanding? This week on Win Today, Gabe and Rebekah Lyons join us to dismantle pervasive myths about marriage and intimacy. Drawing from their 28 years together, they offer candid insights into: The biggest lie about intimacy that's sabotaging relationships. Why viewing your spouse as a 'problem to solve' undermines connection. Understanding the anatomy of an affair and how to safeguard your marriage. The concept of marrying your emotional equivalent and its impact on relational dynamics.   In a culture that often glamorizes the idea of effortless romance, Gabe and Rebekah present a refreshing narrative—one that champions resilience, mutual growth, and the beauty of navigating challenges together. Whether you're single, engaged, or have been married for years, this conversation offers invaluable perspectives to cultivate a relationship that not only endures but flourishes. Episode Links Show Notes Buy my NEW BOOK "Healing What You Can't Erase" here! Invite me to speak at your church or event. Connect with me @WINTODAYChris on Instagram, Facebook, Twitter, and YouTube.

Part Two: Gratuitous Goods in a Free EconomyChapter 7 of Abundance, Generosity, and the State: An Inquiry into Economic Principles audiobook.From pp. 219–246 in the print edition.

Justice vs. Moral Equivalence ~ The injustice of blaming the victim - a short interview with Dr. Yaron Brook. Listen to caller's personal dramas four times each week as Dr. Kenner takes your calls and questions on parenting, romance, love, family, marriage, divorce, hobbies, career, mental health - any personal issue! Call anytime, toll free 877-Dr-Kenner. Visit www.drkenner.com for more information about the show.

Dr Nir Lahav PhD, is a Physicist from Bar-Ilan University, now at Cambridge University. His field of study is in Neuroscience and Consciousness studies. He uses network theory, non linear dynamics and chaos. He developed The Relativistic Theory of Consciousness - a new physical theory in order to solve the hard problem of consciousness. He is also developing a model for chaotic neural networks in collaboration with Zachariah Neemeh, his 2022 paper ("A Relativistic Theory of Consciousness") was chosen as one of the most influential articles of the year by Neuroscience News. TIMESTAMPS: (0:00) - Introduction (1:18) - Definition of Consciousness (5:44) - A Relativistic Theory of Consciousness (15:48) - The Equivalence Principle & the Zombie Argument (21:42) - Physicalism vs Reductive Materialism (28:56) - Nir's Exploration of Relativity & Consciousness (37:20) - How to test the Relativistic Theory (future experiments & predictions) (42:08) - Emergentism, Functionalism, & Cognitive Frames of Reference (53:25) - Cognitive Maps & Consciousness (1:00:00) - Animal/AI Consciousness (minimal conditions) (1:03:30) - Free Will & Determinism (1:12:55) - Is the Universe be Conscious? (from a Relativistic Approach) (1:22:18) - From Absolutism to Relativism (1:31:20) - Future work (1:33:10) - Conclusion EPISODE LINKS: - Nir's Website: https://www.lahavnir.com/ - Nir's Paper: https://www.sciencedirect.com/science/article/pii/S0079610723001128 - Nir's "GoFundMe": https://gofund.me/b37c00d5 CONNECT: - Website: https://tevinnaidu.com - Podcast: https://creators.spotify.com/pod/show/mindbodysolution - YouTube: https://youtube.com/mindbodysolution - Twitter: https://twitter.com/drtevinnaidu - Facebook: https://facebook.com/drtevinnaidu - Instagram: https://instagram.com/drtevinnaidu - LinkedIn: https://linkedin.com/in/drtevinnaidu ============================= Disclaimer: The information provided on this channel is for educational purposes only. The content is shared in the spirit of open discourse and does not constitute, nor does it substitute, professional or medical advice. We do not accept any liability for any loss or damage incurred from you acting or not acting as a result of listening/watching any of our contents. You acknowledge that you use the information provided at your own risk. Listeners/viewers are advised to conduct their own research and consult with their own experts in the respective fields.

Greg (@GregBradyTO) speaks with Michael Kerzner, Ontario's solicitor general and MPP for York Centre, about how many in the province are on high alert following bomb threats made to 100s of Jewish institutions and doctors Wednesday. Learn more about your ad choices. Visit megaphone.fm/adchoices

ESGfitness.co.uk/commit Chapters 00:00 Accuracy of Body Analyzer Scales and Body Composition 01:26 The Role of Running in Weight Loss 02:47 Exercise for Mental Health and Mood Improvement 05:24 Focus on Feeling Good First 08:39 Equivalence of Treadmill and Outdoor Walking 09:39 Balancing Steps and Calorie Intake 10:25 Electric Muscle Training and Strength Training 13:17 Estimating Maintenance Calories and Weight Loss 15:03 Managing Fatigue on Medication 17:31 Overcoming Plateaus in Strength Training 18:56 Tracking Food: Before or After Cooking? 20:05 Static vs. Dynamic Stretching

Episode No. 665 features curator Cathleen Chaffee and critic Elisabeth Kirsch.  Chaffee is the curator of "Marisol: A Retrospective," which is at the Buffalo AKG Art Museum (formerly the Albright-Knox Art Gallery) through January 6, 2025. The exhibition presents work Marisol, sometimes remembered as 'the forgotten star of pop art,' made between the 1950s and the early 2000s. It builds on an extraordinary collection of works that Marisol left to the Buffalo AKG Museum upon her death. The museum and DelMonico Books have published a superb catalogue. Amazon and Bookshop offer it for $40-70. Chaffee curated the exhibition with the assistance of Julia Vázquez.  Kirsch is the author of "Handmade Papers, 1980-2005," an essay in the catalogue for "Virginia Jaramillo: Principle of Equivalence," a retrospective now at the Museum of Contemporary Art Chicago. The catalogue was edited, and the exhibition curated, by Erin Dziedzic. At the MCA, where "Jaramillo" is on view through January 5, 2025, its presentation was organized by René Morales and Iris Colburn. The exhibition's middle gallery presents an extensive mini-survey of Jaramillo's paper-constructed works. Amazon and Bookshop offer the catalogue for about $50. Instagram: Cathleen Chaffee, Tyler Green.

Audio, eng_t_rav_2024-07-26_lesson_rb-0332-ishtavut-tzura_n1_p2. Lesson_part :: Daily_lesson 1

Video, eng_t_rav_2024-07-26_lesson_rb-0332-ishtavut-tzura_n1_p2. Lesson_part :: Daily_lesson 1

Audio, eng_t_rav_2024-07-26_lesson_rb-0332-ishtavut-tzura_n1_p2. Lesson_part :: Daily_lesson 1

Video, eng_t_rav_2024-07-26_lesson_rb-0332-ishtavut-tzura_n1_p2. Lesson_part :: Daily_lesson 1

President Biden confronted NBC host Lester Holt, a registered Republican, for making a false equivalence between political rhetoric, emphasizing the dangers of misleading comparisons in media. Subscribe to our Newsletter: https://politicsdoneright.com/newsletter Purchase our Books: As I See It: https://amzn.to/3XpvW5o How To Make America Utopia: https://amzn.to/3VKVFnG It's Worth It: https://amzn.to/3VFByXP Lose Weight And Be Fit Now: https://amzn.to/3xiQK3K Tribulations of an Afro-Latino Caribbean man: https://amzn.to/4c09rbE

Flashback to the Beginning of Season 10!Today, we respond to a listener who asked the important question of what to do and say when someone has just experienced trauma, tragedy, or some other massive thing (TMT).Scripture: Romans 12:15Books Mentioned: This is Hard by Jon Swanson, Walking with God Through Pain and Suffering by Timothy Keller, Hope Is the First Dose by meMy Advice When Speaking to Others after TMT:1. Show up and shut up (I stole this from Jon Swanson)2. Tears over talking3. Empathy over equivalence4. Theology is like salt: it needs to be correctly applied in tiny amounts, and of high quality, or it can hurt more than it helps.Leave a voicemail with your question or comment!Five Ways You Can Support this show:Pray for us!Subscribe, like, and share it with your friends! (We even have a YouTube channel!)Leave reviews and comments wherever you listen to podcasts!You can become a paid partner of the podcast and get special bonus episodes and lots more content by clicking here. Visit one of our affiliate partners and consider using their products (we use them every day):Support and boost your immune system with Armra! Use DRLEEWARREN code at checkout for a discount!Improve your gut health, immune system, and protect your brain with Pique!Other Helpful Links:Click here to access the Hope Is the First Dose playlist of hopeful, healing songs!Be sure to check out my new book, Hope Is the First Dose!Here's a free 5-day Bible study on YouVersion/BibleApp based on my new book!Sign up for my weekly Self-Brain Surgery Newsletter here!All recent episodes with transcripts are available here! (00:01) - Introduction (06:51) - Understanding Romans 12:15 (09:37) - Empathy over Equivalence (12:26) - Theology is Like Salt (16:30) - Practical Tips for Helping Others (19:12) - Actions for the Acute Phase

Accusations of false equivalencies have flown from both sides of the conflict in the Middle East, and this week was no exception with the International Criminal Court's Chief announcing their intention to apply for arrest warrants for top leadership in the Israeli Government and those working for Hamas. Joining Tina Quinn to discuss this latest development is the ABC's Geraldine Doogue, Network Ten's Hugh Riminton and former Middle East Correspondent-turned-Human Rights campaigner, Sophie McNeill. Learn more about your ad choices. Visit megaphone.fm/adchoices

An analysis of latest twist in Philippine politics.

Decency was on the ballot. Remember that? Now, Biden cannot seem to find the words when it's time to condemn anti-Semites or violent protestors.

In the one hundred and forty seventh episode we take another look at the Moral Equivalence Fallacy, starting with Trump comparing January 6 insurrectionists with BLM protestors, Dinesh D'Souza comparing Trump to Navalny, and Matthew Dowd comparing sexual assault to mishandling emails.In Mark's British Politics Corner we look at Shelagh Fogarty comparing silence over ministerial misdeeds to hiding in a fridge, and Rishi Sunak and Sam Armstrong equating pro-Palestinian protestors with war criminals, and .In the Fallacy in the Wild section, we check out examples from Halt and Catch Fire, Criminal Minds, and The I.T. Crowd.Jim and Mark go head to head in Fake News, the game in which Mark has to guess which one of three Trump quotes Jim made up.Then we talk about Trump's insane Presidential Records Act defense in his Florida trial.And finally, we round up some of the other crazy Trump stories from the past week.The full show notes for this episode can be found at https://fallacioustrump.com/ft147 You can contact the guys at pod@fallacioustrump.com, on Twitter @FallaciousTrump, or facebook at facebook.com/groups/fallacioustrumpSupport this podcast at — https://redcircle.com/fallacious-trump/donationsAdvertising Inquiries: https://redcircle.com/brandsPrivacy & Opt-Out: https://redcircle.com/privacy

In this episode, Dr. T. Michael W. Halcomb ( @tmichaelwhalcomb ) looks at “Red Herring, Bandwagon, and Moral Equivalence”. Tune in! ***GlossaHouse resources are available at our website! - https://glossahouse.com/ ✏️ ***Sign up for classes with GlossaHouse U - https://glossahouse.com/pages/classes

Kate Middleton released a video announcing her cancer diagnosis and Bryan plays some audio from it. Democrats are up in arms over NBC hiring former RNC chair Ronna McDaniel. Jill Biden thinks the United States is becoming Nazi Germany because of parental rights bills and schools getting porn out of their libraries. // A checking of the texting. // A deputy in Florida says he prefers when homeowners shoot wannabe robbers who break into residences. Woman sentenced to life in prison for toddler's death. 

We explore the chemical similarities and key differences between brand-name and generic drugs. Learn about:Equivalence in active ingredients, as approved by the FDA.Exceptions: Narrow Therapeutic Index drugs and inactive ingredients.When to consult a doctor before switching brands.Listen to find out if you can save money without sacrificing quality!

The DOL recently issued new guidance on the long-standing contraceptive coverage mandate, and this time they took a different approach. Instead of clarifying prior guidance or issuing novel interpretations of the law, the DOL is giving health plans an alternative way to comply with the contraceptive coverage mandate: either follow the prior guidance issued in 2022… or don't! The DOL introduced the “therapeutic equivalence” approach, whereby plans can comply with the law in a different, potentially less-burdensome way. Join The Phia Group's Kendall Jackson and Jon Jablon as they discuss this “therapeutic equivalence” approach to compliance, what it means for consumers, and what it means for health plans.

This show was recorded 22 February 2024. We encourage you to download the supplemental PDF content for this session by completing the form on this page. Please join us live for future shows and ask your own questions! We hold one RQM+ Live! panel discussion per month and you can sign up at the ⁠Knowledge Center⁠ or ⁠Events⁠ pages at RQMplus.com. Thank you for tuning in.

In this episode, Keone Hon dives deep on Monad, a high-performance Ethereum-compatible layer 1 blockchain focused on delivering high performance and drastically lower transaction costs. We discuss Monad's key technical innovations like MonadBFT consensus, parallel execution, and a custom database enabling 10,000+ TPS. Keone explains how these breakthroughs allow complex 100K gas transactions costing just a fraction of a cent, 100x cheaper than even leading Layer 2s. We also explore how Monad achieves this while maintaining decentralization and a rich developer experience via full EVM equivalence. Tune in to learn why Monad could be a game changer for builders and users looking maximize performance and minimize costs. Key Links Monad: https://www.monad.xyz Socials Keone's Twitter Monad's Twitter Tommy's Twitter Follow Delphi Digital Website: ⁠⁠https://members.delphidigital.io/home⁠⁠ Twitter: ⁠⁠https://twitter.com/Delphi_Digital⁠⁠ Youtube: ⁠⁠https://www.youtube.com/@Delphi_Digital⁠ Disclosures Disclosures: This podcast is strictly informational and educational and is not investment advice or a solicitation to buy or sell any tokens or securities or to make any financial decisions. Do not trade or invest in any project, tokens, or securities based upon this podcast episode. The host and members at Delphi Ventures may personally own tokens or art that are mentioned on the podcast. Our current show features paid sponsorships which may be featured at the start, middle, and/or the end of the episode. These sponsorships are for informational purposes only and are not a solicitation to use any product, service or token. Delphi's transparency page can be viewed ⁠⁠here⁠⁠. As an additional disclosure, Tom Shaughnessy the host is an angel investor in Monad. Keywords Monad, Keone Hon, EVM, Layer 1, high performance, low cost transactions, Ethereum Virtual Machine, EVM compatible, parallel execution, consensus, decentralization, developers, Monad xyz, Tommy Shaughnessy, Delphi Digital, Delphi, Crypto, Cryptocurrency, Ethereum, --- Send in a voice message: https://podcasters.spotify.com/pod/show/the-delphi-podcast/message

Can we equate the struggle of the Jewish people with the struggle of the Palestinians? Can we justify within ourselves that a bus full of fathers, mothers, and children being blown up in Jerusalem is the same as an apartment complex collapsing and killing fathers, mothers, and children in Gaza? Can we equate rape, infanticide, and torture with Israel's retaliatory airstrikes? Are rocks thrown at passing motorists displaying Israeli plates in Judea and Samaria the same as targeted killings of Jihadi terrorist leaders? Former President Barack Obama seems to think so, as do many of his supporters. This entire idea of “moral equivalency”, or simply put - the idea that when Israel acts in self-defense, they must only respond “proportionally”, is a monstrous, ill-informed idea, and if not done away with, could very well be the downfall of the Western World, as we know it.

If you want the quintessential GRE Verbal question, Sentence Equivalence questions might take the cake. Clay Daniel, founder of Clayborne Test Prep and Tutoring, has been tutoring the GRE for decades. In this podcast, Clay walks you through five new Sentence Equivalence problems to help you practice this tricky question type. Achievable is a modern test prep platform for the GRE exam - get a free trial at https://achievable.me

Brent oil price at $74 is $8 too cheap vs our model-derived fair value of $82 for December. Given our optimistic view on demand next year, we see Brent oil trading in the low $80s by March and high $80s by May. We find current oil prices in the lower range of the trading band and see the recent sell-off as a potential buying opportunity—even with thinning liquidity ahead of the holiday period—given our divergence with consensus on fundamental factors affecting oil market dynamics in 2024. Given our outlook on demand, we argue that there is no need for further OPEC cuts were the alliance to accept lower oil prices in the mid-$70s to mid-$80s range. Instead, we believe, the alliance should unwind some of the voluntary reductions in order to gain operational flexibility when demand growth takes a step down in 2025 (and potentially 2026), when most of the post-COVID demand normalization is behind us and decarbonization policies begin to cut into demand for some products. Speaker: Natasha Kaneva, Head of Global Commodities Research  This podcast was recorded on 15 December 2023. This communication is provided for information purposes only. Institutional clients can view the related report at https://www.jpmm.com/research/content/GPS-4582283-0 for more information; please visit www.jpmm.com/research/disclosures for important disclosures. © 2023 JPMorgan Chase & Co. All rights reserved. This material or any portion hereof may not be reprinted, sold or redistributed without the written consent of J.P. Morgan. It is strictly prohibited to use or share without prior written consent from J.P. Morgan any research material received from J.P. Morgan or an authorized third-party (“J.P. Morgan Data”) in any third-party artificial intelligence (“AI”) systems or models when such J.P. Morgan Data is accessible by a third-party. It is permissible to use J.P. Morgan Data for internal business purposes only in an AI system or model that protects the confidentiality of J.P. Morgan Data so as to prevent any and all access to or use of such J.P. Morgan Data by any third-party.

U.S. warship shot down drones over the Red Sea which were fired from Iran-backed Houthis in Yemen. Hawaiian Airlines joins forces with Alaska Airlines and Bryan speculates as to what this might mean for both companies. Rep. Pramila Jayapal's shameful moral equivalence about the Israel-Hamas war. // Thurston County candidate who didn't vote for himself lost by one vote. Hot conservative sports takes about the college football playoff.  // Nicaragua's Miss Universe franchise owner charged with conspiracy. Communist dictators gonna dictate.  

“The Rise in Mental Health Issues with Tara and Lee” “Moral Equivalence and Mental Health” “The WEF and Net Zero for Food” “Nikki Haley comes Pre-Owned”

Moral Equivalence and Mental Health https://www.audacy.com/989wordThe Tara Show Follow us on Social MediaJoin our Live StreamWeekdays - 6am to 10am Facebook: https://www.facebook.com/989wordRumble: https://rumble.com/c/c-2031096X: https://twitter.com/989wordInstagram: https://www.instagram.com/989word/ "Red Meat, Greenville." 12/04/23

This week, Eliana and Chris talk about a family member jumping into the political spotlight and no, its not Mike Johnson. We're looking at the latest narratives to come from the Israel-Hamas conflict, 2024, a celebrity's attempted cancellation and our hosts have a healthy debate around the Speaker's developing strategy. Time Stamps:  3:38 Front Page 50:32 Obsessions  56:45 Reader Mail  58:29 Favorite Items  Show Notes: Wapo: Biden's dismissal of the reported Palestinian death toll  National Review: New York Times: Will Iran Ever Get Around to Killing the Jews Like They Say They Will?  NYT: Penn Donors Are Revolting Over Israel-Hamas, DEI, Trans Rights and Other Issues The Atlantic: The Decolonization Narrative Is Dangerous and False  Power Line: My cousin Dean WaPo: Opinion | Dean Phillips, who's challenging Biden, answers our top 5 questions  NYT: Trump's Verbal Slips Could Weaken His Attacks on Biden's Age WSJ: Why Trump's Drastic Plan to Slash the Government Could Succeed Politico: ‘He Seems to Be Saying His Commitment Is to Minority Rule' Politico: Speaker Johnson taps veteran GOP operative as chief spokesperson  Free Beacon: How New Speaker Mike Johnson Can Silence Liberal Critics and Win Back the Mainstream Media  WaPo: Years into a climate disaster, these people are eating the unthinkable WSJ: ‘Sorry' for Being So Blunt  WFB: Big Yikes, Fam: Taylor Lorenz's New Influencer Book Is Mid. It's Giving Cringe, Does Not Slap, on God. Obsessions: Free Beacon: I Said Hamas Raped and Beheaded. The Yale Daily News Issued a Correction. The Dispatch: Pence Runs Away From the Circus - Chris Stirewalt Favorite Item:  Tablet: The George Kennan Who Wasn't WSJ: The Astrophysicist Who Has a Better Way to Board Airplanes

Yael Bar Tur, Tal Fortgang, and Martin Gurri join Brian C. Anderson to discuss anti-Israel sentiment and the role of information in the Israel–Hamas war.  The audio for this episode is adapted from a recent virtual event (watch here). anti-Israel-sentiment role-of-information-in-Israel–Hamas-war

Yesterday saw a dramatic protest at the US Capitol complex which the political Right equated to the protests on January 6, 2021. We talk about the similarities and where they end. And how any attempts to align such events with the history of Dec 7, 1941 and Sept 11, 2001 fall absurdly short. ____________________________________________________ SUSPENDABLES MERCH: http://The-Suspendables.com Visit http://PatriotCoolers.com/discount/KYLE and use Promo code "KYLE" for 10% off and free shipping over $50. Today's podcast supported by https://CatholicVote.Org If you are interested in supporting the going litigation against the FBI over religious liberties, you can visit https://CatholicVote.Org.

Join Dr. Erez Soref and Dr. Seth Postell as we discuss the error of ‘Bothsidesism' or Moral Equivalence when approaching the Israeli conflict we're facing today. You can join us in this historic effort! Please give now to offer vital assistance to Israelis in crisis, including: https://www.oneforisrael.org/israelis-relief-2023/ Emergency Food: Providing nourishment to those in immediate need. Essential Supplies: Ensuring families have access to critical items. Trauma Support: Offering comfort and care to those traumatized by the events. Funding for Evacuation Accommodations: Helping families find safe shelter during this crisis.

If the Making Sense podcast logo in your player is BLACK, you can SUBSCRIBE to gain access to all full-length episodes at samharris.org/subscribe.

Dennis highlights a column in the City Journal by Juliana Geran Pilon… the efforts of mainstream media, government officials, and elite universities to avoid outright condemnation of Hamas reveal a West at war with itself.  Dennis also analyzes NYT's Bret Stephens' latest column: The Anti-Israel Left Needs to Take a Hard Look at Itself.  Can a husband tell his overweight wife that she should lose weight? Callers weigh in.  Dennis is joined by Ben Shapiro of the Daily Wire… they analyze the ongoing dispute in the Middle East.  Thanks for listening to the Daily Dennis Prager Podcast. To hear the entire three hours of my radio show as a podcast, commercial-free every single day, become a member of Pragertopia. You'll also get access to 15 years' worth of archives, as well as daily show prep. Subscribe today at Pragertopia dot com.See omnystudio.com/listener for privacy information.

Episode No. 609 of The Modern Art Notes Podcast is a summer clips episode featuring artist Virginia Jaramillo. The Kemper Museum of Contemporary Art in Kansas City is presenting "Virginia Jaramillo: Principle of Equivalence," the first retrospective of Jaramillo's work. The exhibition includes 73 paintings and handmade paper works extending back over 70 years. The exhibition was curated by Erin Dziedzic and will be on view through August 26. A catalogue is forthcoming. This episode was recorded on the occasion of  “Virginia Jaramillo: The Curvilinear Paintings, 1969-74” which was at the Menil Collection in 2020. The show was the first solo museum exhibition of Jaramillo's career. Curated by Michelle White, it featured a series of paintings that Jaramillo made featuring the joining of line to color against mostly monochromatic backgrounds. See Episode No. 469 for images.

This talk was given by Gil Fronsdal on 2023.03.26 at the Insight Meditation Center in Redwood City, CA. ******* Video of this talk is available at: https://youtube.com/live/l8lshxaASUY . ******* For more talks like this, visit AudioDharma.org ******* If you have enjoyed this talk, please consider supporting AudioDharma with a donation at https://www.audiodharma.org/donate/. ******* This talk is licensed by a Creative Commons Attribution-Noncommercial-No Derivative Works 4.0 License

This talk was given by Gil Fronsdal on 2023.03.26 at the Insight Meditation Center in Redwood City, CA. ******* Video of this talk is available at: https://youtube.com/live/l8lshxaASUY . ******* For more talks like this, visit AudioDharma.org ******* If you have enjoyed this talk, please consider supporting AudioDharma with a donation at https://www.audiodharma.org/donate/. ******* This talk is licensed by a Creative Commons Attribution-Noncommercial-No Derivative Works 4.0 License

A new peer-reviewed study in the journal Science shows that not only did Exxon scientists suspect climate change driven by the burning of fossil fuels was a growing problem that would lead to crisis if nothing changed, but they were terrifyingly accurate in their modeling and predictions. Alongside this special re-broadcast of Season 1 of Drilled, all about the origins of climate denial and Exxon's role in it, we speak with the study's lead author Geoffrey Supran about its importance. In this episode, the industry's role in creating and then weaponizing false equivalence on climate—the idea that the opinions of a handful of contrarians are equally valid to those of the majority of peer-reviewed studies on the topic.