Podcasts about Deng

Oppermachtig, president en partijleider voor het leven - de eerste sinds Mao. Xi Jinping wist zelfs Donald Trump te imponeren, ‘straight from Central Casting in Hollywood’. Dramatische gebeurtenissen in Beijing - zuivering van de militaire top, aanklachten wegens corruptie en atoomspionage, geruchten over een couppoging - duiden op grote troebelen en een onverwacht machtsvacuüm. Wat weten we, wat betekent dit en welke consequenties heeft dit voor China, voor Xi’s positie en de geopolitieke verhoudingen? *** Deze aflevering is mede mogelijk gemaakt met donaties van luisteraars die we hiervoor hartelijk danken. Word ook vriend van de show! Deze aflevering bevat een advertentie van Oogwereld Heb je belangstelling om in onze podcast te adverteren of ons te sponsoren? Zend ons een mailtje en wij zoeken contact. *** Jaap Jansen en PG Kroeger pluizen de berichten uit. Officiële verklaringen en commentaren in Beijing, maar ook de tsunami aan geruchten over de ongekende ingreep in de Centrale Militaire Commissie. Naast voorzitter Xi Jinping is nog maar één lid in functie. Vicevoorzitter generaal Zhang Youxia en chefstaf generaal Liu Zhenli zijn opgepakt. Ze worden beschuldigd van insubordinatie en 'verraad aan het vertrouwen van de Communistische Partij en haar Centraal Comité'. Vervolgens zijn bijna al hun collega's uit hun functie verwijderd en is in heel China in de topstructuur van de militaire districten en krijgsmachtonderdelen een massale zuivering en ontslaggolf gaande. Die formele beschuldiging wijst erop, dat hier geen sprake is van een verschil van inzicht over militaire kwesties alleen. De ingreep van Xi gaat veel dieper. Dit raakt aan de essentie van het vertrouwen tussen de partij, de politieke leiding en de leiding van het Volksbevrijdingsleger als geheel. En juist die balans van de twee machtscentra van de staat is cruciaal voor de ongebreidelde macht van Xi en die van zijn voorgangers als opperste leider en roerganger. Het conflict gaat dan ook over de continuïteit en fundamenten van die politieke macht, zoals in elke dictatoriale staat. Dit evenwicht is door het onthoofden van de CMC nu in één klap verwoest. De vraag naar die continuïteit van de macht is een vraag naar de betekenis en effecten van het feit dat Xi Jinping leider voor het leven is geworden. Dat roept hoe dan ook de vraag op: en wat dan? De militaire top wil zekerheid over wie hun opperbevelhebber zal zijn, terwijl van een breed gedragen regeling voor de opvolging - zoals Deng Xiaoping die ooit doorvoerde - geen sprake meer is. Wie daar hoe en wanneer over zal gaan is onhelder gehouden. In 2027 zal het volgende partijcongres daar wellicht over beslissen en ook in het Centraal Comité, in de machtige regio's was al enig rumoer merkbaar. De legertop was blijkbaar betrokken bij die onderhuidse signalen en eerdere zuiveringen indiceerden al spanningen en twijfels aan de koers van Xi. Bovendien was sprake van scherp toegenomen spanning over de voorziene aanpak van Taiwan. Het jaar 2027 staat daarin centraal. Dan wil Xi klaar zijn voor een definitieve aanpak van het weerspannige eiland en de hereniging tot één China realiseren. Daar komt nog iets bij dat van grote symbolische betekenis is. Dit najaar herdenkt China de dood van Mao Zedong in1976 en volgend jaar viert het 'honderd jaar Volksbevrijdingsleger'. Dat generaal Zhang insuborinatie pleegde, duidt erop dat hij de koers van Xi ten aanzien van Taiwan wel erg riskant vond en de voorbereidingen voor de invasie niet zonder meer wilde implementeren. Dat zou ertoe leiden dat Xi die grote regime-jubilea zou moeten laten passeren zonder het verwachte succes bij de nationale hereniging. Hij zou dan de al vele jaren bestaande strategie dat China in 2035 op gelijke voet met Amerika moet staan als enige wapenfeit over houden. Zijn gezag als partijleider komt zo in het geding. De vraagt rijst dan of de 74-jarige Xi niet door iemand uit een jongere generatie vervangen moet worden. Zie hier de verbinding tussen de strategische, geopolitieke dilemma's en de onopgeloste opvolgingsvraag. Ook de zeer sombere, soms dystopische speech van Xi bij het recente '75 Jaar Rood China' feest wordt hierdoor ineens veel beter te begrijpen. Heeft in China een dergelijk drama zich al eerder voltrokken? Al tijdens de jonge democratische republiek in de jaren twintig van de vorige eeuw was 'warlordism' een groot probleem. De zwakke democratische instituties werden permanent ondermijnd door regionale militaire bazen die met elkaar burgeroorlogen begonnen en soms ook met Japan samenwerkten tegen de republiek. Mao kon hen met zijn Volksbevrijdingsleger tegen elkaar, Japan en de Kuomintang uitspelen en won zo de overhand. In crisisfasen van Mao's regime - zoals de Grote Sprong Voorwaarts en de Culturele Revolutie - doken onmiddellijk zulke regionale conflicten en warlords weer op. Het leger drukte dit telkens hardhandig te kop in, met een sleutelrol voor maarschalk Ye Jianying. Die was ook cruciaal in de meest opmerkelijke couppoging. In 1971 deed Mao's kroonprins, defensieminister Lin Biao, een greep naar de macht. Dit 'Project 571' mislukte en op de vlucht naar Moskou werd boven Mongolië Lins vliegtuig neergehaald. Na Mao's dood grepen Ye en het leger weer in, namen zijn weduwe Jiang Qing en haar aanhang gevangen en zorgden dat Deng Xiaoping de nieuwe leider werd. Er was weer eenheid in commando over staat en krijgsmacht. Precies die incoherentie is het grote risico dat Xi nu loopt. De onzekerheid over de politieke ontwikkelingen en stabiliteit van China verhoogt de toch al labiele situatie wereldwijd. Dat bijvoorbeeld zeer recent India en de EU, maar ook de EU en Vietnam nadrukkelijk elkaar als strategische partners hebben gevonden heeft hier alles mee te maken. Dat Nederland als het land van ASML in deze troebelen rond Taiwan, China, Japan en Amerika een belangrijke rol speelt, had Den Haag zich kortgeleden nog niet kunnen indenken. De agenda van Rob Jetten en het strategisch inzicht van zijn ministers Tom Berendsen en Sjoerd Sjoerdsma worden al bij aantreden danig op de proef gesteld. *** Verder kijken Volksbevrijdingsleger - Lange Mars Musical (1976) *** Verder luisteren Xi Jinping 453 – 75 jaar Volksrepubliek China, waar is het feestje? 306 - De gevoelige geopolitieke relatie met China24 - Ties Dams over China's nieuwe keizer Xi Jinping China en zijn leiders250 - Nixon in China: de week die de wereld veranderde 225 - Nixon in China: Henry Kissinger's geheime (en hilarische) trip naar Beijing 245 - Oompje neemt de trein – de reis die China naar de 21e eeuw bracht 220 - China's nieuwe culturele revolutie 58 - 70 jaar China, de Volksrepubliek van Mao, Deng en Xi Geopolitieke troebelen 549 - China en Japan op ramkoers 458 - De gedroomde nieuwe wereldorde van Poetin en Xi Lessen voor Jetten, Berendsen en Sjoerdsma 551 – Klem tussen Amerika en China: de koude oorlog rond ASML 558 – Poetins rampjaar, Jettens kans *** Tijdlijn 00:00:00 – Deel 1 00:22:55 – Deel 2 00:45:49 – Deel 3 01:27:10 – EindeSee omnystudio.com/listener for privacy information.

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

Zirian Fatah, President of Kurdish Lobby Australia, has expressed concern over the escalating crisis facing Kurdish communities in Iran and northern Syria. In a letter to the Australian Government, the Lobby highlighted a deadly crackdown in Iran, where thousands of civilians have died amid nationwide protests. Meanwhile, in northern Syria, intensified fighting around Kobani has forced many families to flee, as the city and surrounding areas face a growing humanitarian crisis. Kurdish Lobby Australia is calling on the Australian Government to take urgent action to address these interconnected humanitarian and security threats. - Zirian Fetah, serokê Lobiya Kurd a Australya, li ser krîza ku li hember civakên Kurd li Îran û bakurê Sûriyeyê rû bi rû ne, di xemê de ye. Di nameyeke ji Hukûmeta Australya re, balê dikişîne ser serkutkirina kujer a li Îranê, ku bi hezaran sivîl di nav xwepêşandanên seranserî welêt de mirine. Di heman demê de, li bakurê Sûriyê, şerên dijwar ên li dora Kobaniyê gelek malbat neçar kirine ku aware bibin, ji ber ku bajar û deverên derdorê bi topbaran û krîzeke mirovî ya mezin re rû bi rû dibe. Name banga tedbîrên lezgîn dike da ku van gefên mirovî û ewlehiyê yên bi hev re girêdayî çareser bikin.

Wendi Deng was the third wife of media magnate and billionaire, Rupert Murdoch. The public was quick to call her a gold digger, latching onto the 38-year age difference in their marriage, Wendi's humble upbringing in China, and her controversial involvement in the Murdoch family business. But despite criticism and rumors about her personal life, Wendi leveraged Rupert's connections to build a name for herself as a businesswoman, film producer, and advisor in the tech world. Lena and Alissa discuss ambition, Wendy's celebrity dinner parties & annual pre-Met Gala gathering, and the hypocrisy of the term “gold digger.”  This episode was first published on 10/08/2020. To learn more about listener data and our privacy practices visit: https://www.audacyinc.com/privacy-policy Learn more about your ad choices. Visit https://podcastchoices.com/adchoices

This month Diana is in discussion with UCLA Law student Melissa Deng and Robyn Huey Lao, an allergy mom advocate and pediatric nurse practitioner, about Robyn's journey to shepherd California Senate Bill 68, the Allergen Disclosure for Dining Experiences, to passage.  She explains how she was inspired by her nine-year-old daughter, Addie, and how their love of food inspired this bill. The three talk about building coalitions, the process of introducing bills to the legislature, and the importance of legal research. Melissa also explains her work at UCLA on restaurant training for food allergens. You can find Robyn's website, Addie Tells All, here.You can follow her on Instagram at @AddieTellsAllYou can find the law review article, Fearless Dining: Mandating Universal Allergen Disclosures on Restaurant Menus, that started Robyn thinking here. You can find Melissa Deng on LinkedIn here. Diana Winters is the Deputy Director of the Resnick Center for Food Law & Policy at UCLA Law.

China looms large in our modern world -- and yes, is misunderstood, or not understood at all. Manoj Kewalramani joins Amit Varma in episode 435 of The Seen and the Unseen to put together a primer on China, from ancient times through all the kingdoms through Mao and Deng to Xi and this present moment. (FOR FULL LINKED SHOW NOTES, GO TO SEENUNSEEN.IN.) Also check out: 1. Manoj Kewalramani on LinkedIn, Twitter, Google Scholar, Amazon, Takshashila, Substack and his own website. 2. The Takshashila Institution. 3. The Great Power Show on Spotify, Apple and Substack. 4. Tracking People's Daily -- Manoj Kewalramani's Substack Newsletter. 5. Eye on China -- Anushka Saxena's Substack newsletter (formerly written by Manoj). 6. China in the Changing Geo-economic Landscape -- A Takshashila course on China that begins later this month. 7. The China Dude Is in the House -- Episode 231 of The Seen and the Unseen (w Manoj Kewalramani). 8. What Does China Want? — Episode 143 of The Seen and the Unseen (w Manoj Kewalramani). 9. Chinese Foreign Policy — Episode 81 of The Seen and the Unseen (w Manoj Kewalramani). 10. Foreign Policy is a Big Deal — Episode 170 of The Seen and the Unseen (w Pranay Kotasthane & Manoj Kewalramani). 11. The Dragon and the Elephant — Episode 181 of The Seen and the Unseen (w Hamsini Hariharan & Shibani Mehta). 12. Keeping India Safe — Episode 219 of The Seen and the Unseen (w Sushant Singh). 13. Kanti Bajpai on India vs China -- Episode 234 of The Seen and the Unseen. 14. The History of Indian Cricket -- Mihir Bose. 15. The Invention of China -- Bill Hayton. 16. When Marx Met Confucius. 17. Red Roulette -- Desmond Shum. 18. Deng Xiaoping and the Transformation of China -- Ezra Vogel. 19. China's New Red Guards -- Jude Blanchette. 20. Factional Model-making in China -- Olivia Cheung. 21. America against America -- Wang Huning. 22. The Lying Flat Movement. 23. China Media Project. 24. Probal DasGupta Goes to the Himalayas With Books in His Bag — Episode 412 of The Seen and the Unseen. 25. Wealth and Power -- Orville Schell and John Delury. 26. Rethinking Chinese Politics -- Joseph Fewsmith. 27. The Party's Interests Come First -- Joseph Torigian. 28. Prestige, Manipulation, and Coercion -- Joseph Torigian. 29. John Fairbank, Alastair Iain Johnston and Vijay Gokhale on Amazon. 30. The Long Game -- Vijay Gokhale. 31. China Talk -- Jordan Schneider. 32. Sinica -- Kaiser Kuo. 33. The Third Revolution -- Elizabeth Economy. 34. End of an Era -- Carl Minzner. 35. Hrithik Roshan on Dhurandhar. This episode is sponsored by The Six Percent Club. Join them to go from content idea to launch in just 45 days! Amit Varma runs a course called Life Lessons, which aims to be a launchpad towards learning essential life skills all of you need. For more details, and to sign up, click here. And have you read Amit's newsletter? Subscribe right away to The India Uncut Newsletter! It's free! Also check out Amit's online course, The Art of Clear Writing. Episode art: 'A Maze' by Simahina.

In dieser Episode des Brettspiel News Podcasts ist Erik zu Gast, der mit Wyrmgold im kommenden Jahr sein erstes Spiel veröffentlichen will. Aufgenommen wurde das Gespräch auf der SPIEL in Essen, am Stand des Spielezentrums Herne. Dort spricht Erik über die Entstehung von Believe in me, please und den Weg von der ersten Idee bis zur geplanten Veröffentlichung 2026. Interviewer ist Daniel.Entwickelt wurde der Titel von einem Trio aus Erik sowie Cornel und Hennecke. Ausgangspunkt waren gemeinsame Gespräche über Brettspiele und darüber, welche Art Erlebnis am Tisch entstehen soll. Im Kern versetzt Believe in me, please die Spielenden in die Rolle von Göttern, die um Gläubige ringen. Nach Eriks Schilderung stand das Thema früh fest, bevor einzelne Mechaniken konkret ausformuliert wurden. Die Gottheiten sind dabei für Naturphänomene zuständig, die im Spielverlauf beansprucht und genutzt werden können. Als Tonvorbilder nennt Erik humoristische Einflüsse, unter anderem Terry Pratchett und Monty Python.Auch die Verlagssuche ist Teil des Gesprächs. Erik beschreibt, dass es zunächst sehr unterschiedliche Reaktionen gab, von Zurückhaltung bis deutlichem Interesse. Am Ende habe sich mit Wyrmgold ein Partner gefunden, der die Grundidee verstanden und das Projekt begleiten wollte.Beim Blick auf den Spielablauf geht es vor allem um Interaktion und Entscheidungen unter Zeitdruck am Tisch. Nach Eriks Angaben kombiniert Believe in me, please mehrere bekannte Bausteine wie Tableau Building und Deck Building. Zusätzlich soll ein verdeckter Stichmechanismus enthalten sein. Über ausgespielte Karten werden Naturphänomene beansprucht, die eigene Gottheit weiter ausgebaut und zugleich die Pläne der anderen im Blick behalten. Ein zentraler Anspruch in der Entwicklung sei gewesen, Downtime zu begrenzen, damit die Partien in Bewegung bleiben und Reaktionen auf gegnerische Züge möglich sind.Zur Gestaltung nennt Erik die Illustratorin Maren, die auch an Pagan gearbeitet hat. Für 2026 ist die Veröffentlichung angesetzt, bis dahin wird laut Erik weiter getestet und am Feinschliff gearbeitet.

In this episode, we speak to Martin Lau and Ziqi Deng of FSSA investment managers to find out. We discuss the impact of AI advances on future earnings, whether monetary policy changes are a material development and China's anti-involution policy.

THE AWAKENING OF CHINA'S ECONOMY Colleague Anne Stevenson-Yang, Wild Ride. Returning to China in 1994, the author witnessed a transformation from the destitute, Maoist uniformity of 1985 to a budding export economy. In the earlier era, workers slept on desks and lacked basic goods, but Deng Xiaoping's realization that the state needed hard currency prompted reforms. Deng established Special Economic Zones like Shenzhen to generate foreign capital while attempting to isolate the population from foreign influence, marking the start of China's export boom. NUMBER 5 194R SHANGHAI

SHOW 12-2-2026 THE SHOW BEGIJS WITH DOUBTS ABOUT AI --  a useful invetion that can match the excitement of the first decades of Photography. November 1955 NADAR'S BALLOON AND THE BIRTH OF PHOTOGRAPHY Colleague Anika Burgess, Flashes of Brilliance. In 1863, the photographer Nadar undertook a perilous ascent in a giant balloon to fund experiments for heavier-than-air flight, illustrating the adventurous spirit required of early photographers. This era began with Daguerre's 1839 introduction of the daguerreotype, a process involving highly dangerous chemicals like mercury and iodine to create unique, mirror-like images on copper plates. Pioneers risked their lives using explosive materials to capture reality with unprecedented clarity and permanence. NUMBER 1 PHOTOGRAPHING THE MOON AND SEA Colleague Anika Burgess, Flashes of Brilliance. Early photography expanded scientific understanding, allowing humanity to visualize the inaccessible. James Nasmyth produced realistic images of the moon by photographing plaster models based on telescope observations, aiming to prove its volcanic nature. Simultaneously, Louis Boutan spent a decade perfecting underwater photography, capturing divers in hard-hat helmets. These efforts demonstrated that photography could be a tool for scientific analysis and discovery, revealing details of the natural world previously hidden from the human eye. NUMBER 2 SOCIAL JUSTICE AND NATURE CONSERVATION Colleague Anika Burgess, Flashes of Brilliance. Photography became a powerful agent for social and environmental change. Jacob Riis utilized dangerous flash powder to document the squalid conditions of Manhattan tenements, exposing poverty to the public in How the Other Half Lives. While his methods raised consent issues, they illuminated grim realities. Conversely, Carleton Watkins hauled massive equipment into the wilderness to photograph Yosemite; his majestic images influenced legislation signed by Lincoln to protect the land, proving photography's political impact. NUMBER 3 X-RAYS, SURVEILLANCE, AND MOTION Colleague Anika Burgess, Flashes of Brilliance. The discovery of X-rays in 1895 sparked a "new photography" craze, though the radiation caused severe injuries to early practitioners and subjects. Photography also entered the realm of surveillance; British authorities used hidden cameras to photograph suffragettes, while doctors documented asylum patients without consent. Finally, Eadweard Muybridge's experiments captured horses in motion, settling debates about locomotion and laying the technical groundwork for the future development of motion pictures. NUMBER 4 THE AWAKENING OF CHINA'S ECONOMY Colleague Anne Stevenson-Yang, Wild Ride. Returning to China in 1994, the author witnessed a transformation from the destitute, Maoist uniformity of 1985 to a budding export economy. In the earlier era, workers slept on desks and lacked basic goods, but Deng Xiaoping's realization that the state needed hard currency prompted reforms. Deng established Special Economic Zones like Shenzhen to generate foreign capital while attempting to isolate the population from foreign influence, marking the start of China's export boom. NUMBER 5 RED CAPITALISTS AND SMUGGLERS Colleague Anne Stevenson-Yang, Wild Ride. Following the 1989 Tiananmen crackdown, China reopened to investment in 1992, giving rise to "red capitalists"—often the children of party officials who traded political access for equity. As the central government lost control over local corruption and smuggling rings, it launched "Golden Projects" to digitize and centralize authority over customs and taxes. To avert a banking collapse in 1998, the state created asset management companies to absorb bad loans, effectively rolling over massive debt. NUMBER 6 GHOST CITIES AND THE STIMULUS TRAP Colleague Anne Stevenson-Yang, Wild Ride. China's growth model shifted toward massive infrastructure spending, resulting in "ghost cities" and replica Western towns built to inflate GDP rather than house people. This "Potemkin culture" peaked during the 2008 Olympics, where facades were painted to impress foreigners. To counter the global financial crisis, Beijing flooded the economy with loans, fueling a real estate bubble that consumed more cement in three years than the US did in a century, creating unsustainable debt. NUMBER 7 STAGNATION UNDER SURVEILLANCE Colleague Anne Stevenson-Yang, Wild Ride. The severe lockdowns of the COVID-19 pandemic shattered consumer confidence, leaving citizens insecure and unwilling to spend, which stalled economic recovery. Local governments, cut off from credit and burdened by debt, struggle to provide basic services. Faced with economic stagnation, Xi Jinping has rejected market liberalization in favor of increased surveillance and control, prioritizing regime security over resolving the structural debt crisis or restoring the dynamism of previous decades. NUMBER 8 FAMINE AND FLIGHT TO FREEDOM Colleague Mark Clifford, The Troublemaker. Jimmy Lai was born into a wealthy family that lost everything to the Communist revolution, forcing his father to flee to Hong Kong while his mother endured labor camps. Left behind, Lai survived as a child laborer during a devastating famine where he was perpetually hungry. A chance encounter with a traveler who gave him a chocolate bar inspired him to escape to Hong Kong, the "land of chocolate," stowing away on a boat at age twelve. NUMBER 9 THE FACTORY GUY Colleague Mark Clifford, The Troublemaker. By 1975, Jimmy Lai had risen from a child laborer to a factory owner, purchasing a bankrupt garment facility using stock market profits. Despite being a primary school dropout who learned English from a dictionary, Lai succeeded through relentless work and charm. He capitalized on the boom in American retail sourcing, winning orders from Kmart by producing samples overnight and eventually building Comitex into a leading sweater manufacturer, embodying the Hong Kong dream. NUMBER 10 CONSCIENCE AND CONVERSION Colleague Mark Clifford, The Troublemaker. The 1989 Tiananmen Squaremassacre radicalized Lai, who transitioned from textiles to media, founding Next magazine and Apple Daily to champion democracy. Realizing the brutality of the Chinese Communist Party, he used his wealth to support the student movement and expose regime corruption. As the 1997 handover approached, Lai converted to Catholicism, influenced by his wife and pro-democracy peers, seeking spiritual protection and a moral anchor against the coming political storm. NUMBER 11 PRISON AND LAWFARE Colleague Mark Clifford, The Troublemaker. Following the 2020 National Security Law, authorities raided Apple Daily, froze its assets, and arrested Lai, forcing the newspaper to close. Despite having the means to flee, Lai chose to stay and face imprisonment as a testament to his principles. Now held in solitary confinement, he is subjected to "lawfare"—sham legal proceedings designed to silence him—while he spends his time sketching religious images, remaining a symbol of resistance against Beijing's tyranny. NUMBER 12 FOUNDING OPENAI Colleague Keach Hagey, The Optimist. In 2016, Sam Altman, Greg Brockman, and Ilya Sutskever founded OpenAI as a nonprofit research lab to develop safe artificial general intelligence (AGI). Backed by investors like Elon Musk and Peter Thiel, the organization aimed to be a counterweight to Google's DeepMind, which was driven by profit. The team relied on massive computing power provided by GPUs—originally designed for video games—to train neural networks, recruiting top talent like Sutskever to lead their scientific efforts. NUMBER 13 THE ROOTS OF AMBITION Colleague Keach Hagey, The Optimist. Sam Altman grew up in St. Louis, the son of an idealistic developer and a driven dermatologist mother who instilled ambition and resilience in her children. Altmanattended the progressive John Burroughs School, where his intellect and charisma flourished, allowing him to connect with people on any topic. Though he was a tech enthusiast, his ability to charm others defined him early on, foreshadowing his future as a master persuader in Silicon Valley. NUMBER 14 SILICON VALLEY KINGMAKER Colleague Keach Hagey, The Optimist. At Stanford, Altman co-founded Loopt, a location-sharing app that won him a meeting with Steve Jobs and a spot in the App Store launch. While Loopt was not a commercial success, the experience taught Altman that his true talent lay in investing and spotting future trends rather than coding. He eventually succeeded Paul Graham as president of Y Combinator, becoming a powerful figure in Silicon Valley who could convince skeptics like Peter Thiel to back his visions. NUMBER 15 THE BLIP AND THE FUTURE Colleague Keach Hagey, The Optimist. The viral success of ChatGPT shifted OpenAI's focus from safety to commercialization, despite early internal warnings about the existential risks of AGI. Tensions over safety and Altman's management style led to a "blip" where the nonprofit board fired him, only for him to be quickly reinstated due to employee loyalty. Elon Musk, having lost a power struggle for control of the organization, severed ties, leaving Altman to lead the race toward AGI. NUMBER 16

In this episode of the Evolving Wellness Podcast, I delve into the complexities of vitamin D and melatonin, particularly during the winter months. I discuss the natural decline of vitamin D levels in winter, the critical role of melatonin, and how modern lifestyles disrupt these hormonal balances. The conversation covers the impact of stress, alcohol, and blue light on health, the importance of circadian rhythms, and the potential benefits and drawbacks of vitamin D supplementation. Links: Become a substack subscriber & get early access to podcasts + free courses - https://open.substack.com/pub/sarahkleinerwellness/p/uvbred-light-protocol?r=5eztl9&utm_campaign=post&utm_medium=web&showWelcomeOnShare=trueHoliday Sale - https://www.sarahkleinerwellness.com/SKW-holidayFree Webinar - https://www.sarahkleinerwellness.com/mycircadianapp-free-webinarTimestamps:00:00 Introduction to Vitamin D and Melatonin02:47 Understanding Seasonal Changes in Vitamin D Levels06:13 The Role of Melatonin in Winter Health09:00 Impact of Modern Lifestyle on Vitamin D and Melatonin11:51 The Effects of Stress and Alcohol on Hormonal Balance15:08 Circadian Rhythms and Their Importance17:51 The Dangers of Blue Light and Sleep Disruption21:07 Winter Infections and Their Impact on Vitamin D23:59 Pros and Cons of Vitamin D Supplementation27:13 Understanding Individual Variability in Vitamin D Metabolism29:52 Short-Term Vitamin D Supplementation: When Is It Appropriate?33:14 The Importance of Comprehensive Testing and Monitoring36:01 Conclusion and Future DirectionsReferences & further reading:Article: You've been warned about sunlight but not about supplements - http://sarahkleinerwellness.com/blog/you-ve-been-warned-about-sunlight-but-not-supplementsArticle - Can vitamin D supplements lower your melatonin? https://sarahkleinerwellness.substack.com/p/can-vitamin-d-supplements-lower-yourManson et al. VITAL primary outcomes. N Engl J Med. 2019. PMID: 30415629Sanders et al. Annual high-dose D ↑ falls/fractures. JAMA. 2010. PMID: 20157135Bischoff-Ferrari et al. Monthly high-dose D ↑ falls. JAMA Intern Med. 2016. PMID: 26747333Jackson et al. WHI Ca+D ↑ kidney stones. N Engl J Med. 2006. PMID: 16481635Slominski et al. Melatonin, mitochondria & skin photobiology. Cell Mol Life Sci. 2020.Hamblin MR. Anti-inflammatory photobiomodulation. AIMS Biophys. 2017.Deng et al. Magnesium, vitamin D status & mortality (NHANES). BMC Med. 2013.Holick et al. Translocation of cutaneous vitamin D₃. Endocrinology. 1994.This video is not medical advice & as a supporter to you and your health journey - I encourage you to monitor your labs and work with a professional!________________________________________Get all my free guides and product recommendations to get started on your journey!https://www.sarahkleinerwellness.com/all-free-resourcesCheck out all my courses to understand how to improve your mitochondrial health & experience long lasting health! (Use code PODCAST to save 10%) - https://www.sarahkleinerwellness.com/coursesSign up for my newsletter to get special offers in the future! -https://sarahkleinerwellness.kit.com/profile?_gl=1*1gs7n29*_gcl_aw*R0NMLjE3NjQwOTIxNjcuQ2owS0NRaUF4SlhKQmhEX0FSSXNBSF9KR2poSEZxTTl0blpDSEl4SjYyRHdpa1FuNGc3QXplVll4NVktSFhmSFZZamEwVDAtcU92YXlfQWFBbHp0RUFMd193Y0I.*_gcl_au*MTgwMTYwMTMxMi4xNzYzMzIwODkyLjYyODE4ODI3NC4xNzY1NTA5NjM3LjE3NjU1MDk2MzY.Free Guide to Building your perfect quantum day (start here) -https://www.sarahkleinerwellness.com/opt-in-9d5f6918-77a8-40d7-bedf-93ca2ec8387fMy free product guide with all product recommendations and discount codes:https://www.canva.com/design/DAF7mlgZpJI/xVyE4tiQFEWJmh_Xwx8Kbw/view?utm_content=DAF7mlgZpJI&utm_campaign=designshare&utm_medium=link2&utm_source=uniquelinks&utlId=h0782b52987

Die Europäer einigen sich in der Nacht auf Hilfen für die Ukraine. In den USA werden die Epstein-Akten veröffentlicht. Und Wladimir Putin nimmt den Telefonhörer in die Hand. Das ist die Lage am Freitagmorgen. Die Artikel zum Nachlesen: Mehr Hintergründe hier: EU-Länder beschließen weitere finanzielle Unterstützung für die Ukraine Mehr Hintergründe hier: Der Epstein-Fluch Mehr Hintergründe hier: Wie Trump und Putin Europa angreifen+++ Alle Infos zu unseren Werbepartnern finden Sie hier. Die SPIEGEL-Gruppe ist nicht für den Inhalt dieser Seite verantwortlich. +++ Den SPIEGEL-WhatsApp-Kanal finden Sie hier. Alle SPIEGEL Podcasts finden Sie hier. Mehr Hintergründe zum Thema erhalten Sie mit SPIEGEL+. Entdecken Sie die digitale Welt des SPIEGEL, unter spiegel.de/abonnieren finden Sie das passende Angebot. Informationen zu unserer Datenschutzerklärung.

Masterpiece Audiobooks: Collection of Chinese Classic Novels

Masterpiece Audiobooks: Collection of Chinese Classic Novels

Frank, Lutz; Geuer, Irene www.deutschlandfunk.de, Dlf-Magazin

Een slaande ruzie tussen de nieuwe premier van Japan Sanae Takaichi en het Chinese bewind van Xi Jinping zet een oud conflict op scherp. Dit raakt niet alleen de machtsverhouding in Oost-Azië, maar meteen ook de rol van de Verenigde Staten in de Stille Oceaan en die van Rusland in zijn eigen verre oosten. En omdat het uiteindelijk draait om het eiland Taiwan, raakt het ook de Europese Unie. En bovenal Nederland, als thuisbasis van ASML. Jaap Jansen en PG Kroeger diepen drie vragen uit: -Waarom provoceerde premier Takaichi meteen bij haar aantreden de grote buur? En waarom reageerde Trump, die 'groot respect' voor haar heeft, zo afhoudend? -Waarom sloeg Xi zo fel terug? -Welke diepe historische gevoeligheden, herinneringen en angsten maken deze explosie even begrijpelijk als riskant? *** Deze aflevering is mede mogelijk gemaakt met donaties van luisteraars die we hiervoor hartelijk danken. Word ook vriend van de show! Heb je belangstelling om in onze podcast te adverteren of ons te sponsoren? Zend ons een mailtje en wij zoeken contact. *** Het pacifisme werd Japan na 1945 opgelegd door president Harry Truman als prijs voor de terugkeer onder de 'fatsoenlijke naties'. Maar in 1972 kwam de 'Nixon Shokku'. De opening naar China door Richard Nixon leek Japan in de kou te zetten en dwong tot herijking van de geopolitieke strategie. Premier Shinzo Abe zette de deur open naar 'zelfverdediging' als agressievere houding en stelde: “Een noodsituatie rond Taiwan is een noodtoestand voor Japan." Dat zijn protegee Takaichi dit herhaalde toen zij Taiwan bezocht, alarmeerde Beijing. Haar coalitie werd direct vanuit China onder druk gezet. Maar dit gaf haar populariteit alleen maar een impuls. Het lijkt erop dat het Chinese bewind hier ook een onverwachte kans zag. Een overleg met een hoge ambtenaar uit Tokyo werd theatraal in scène gezet om hevige nationalistische en historische affecten op te jagen. Zowel militair als cultureel werd Japan in de ban gedaan. Popconcerten van JO1 werden geschrapt, toerisme opgeschort. Een herhaling van massale anti-Japan demonstraties van 2010 dreigde. De Japanse premier probeerde meteen te sussen. Xi Jinping kan deze opwinding goed gebruiken. Hij laat het volk stoom afblazen nu hij zijn nieuwe vijfjarenplan inluidt waarin hightech prioriteit heeft maar het platteland en de middenklasse moeten inleveren. En door Japan aan te pakken terwijl Trump hem schijnbaar bijvalt, laat hij Taiwan voelen dat het eiland nog verder in het isolement gedreven wordt. Deze harde aanpak is in China niet zonder reden populair. Japan overtrof na 1870 de grote buur als nieuwe, moderne wereldmacht en veroverde Taiwan en Korea. Sleutelfiguur in deze razendsnelle ontwikkeling was keizer Meiji die zijn land opengooide naar het Westen als een soort Thorbecke of Deng Xiaoping van zijn tijd. De gruwelen van Japanse agressie in China na 1930 en Mao's militaire triomf over Japan drukken een zwaar stempel op de relaties. Ten diepste is China nog steeds bevreesd voor een ambitieus Japan. Dat premier Takaichi zich als een soort beschermvrouwe van Taiwan zou profileren raakt een open zenuw. Maar tegelijkertijd kan Xi dat eiland laten nu voelen hoe het alleen staat. Het kan zich maar beter in de open armen van China storten. Zijn droom van een 'vreedzame hereniging' naar het model van Dengs greep naar Hong Kong kan zo dichterbij komen. Hij zou dan de voltooier zijn van de nationale eenheid en als heerser voorgoed de gelijke worden van Mao en Deng. *** Verder luisteren 458 - De gedroomde nieuwe wereldorde van Poetin en Xi https://art19.com/shows/betrouwbare-bronnen/episodes/7e62cdac-bdb9-450c-af23-a7f974ec3e42 453 – 75 jaar Volksrepubliek China, waar is het feestje? https://art19.com/shows/betrouwbare-bronnen/episodes/2268a339-e0ca-4d2a-85bd-2ec5c4b6a1ca 24 - Ties Dams over China's nieuwe keizer Xi Jinping https://art19.com/shows/betrouwbare-bronnen/episodes/796c8734-7866-4295-b672-335e345da39e 220 - China's nieuwe culturele revolutie https://art19.com/shows/betrouwbare-bronnen/episodes/3d52b1c2-d383-4e2c-991b-5531b6de78ae 245 - Oompje neemt de trein – de reis die China naar de 21e eeuw bracht https://art19.com/shows/betrouwbare-bronnen/episodes/8041cd16-d577-45e1-83a9-efd7676c226a 250 - Nixon in China: de week die de wereld veranderde https://art19.com/shows/betrouwbare-bronnen/episodes/bee983d6-1372-470a-8ce9-27ea6a2d3020 225 - Nixon in China: Henry Kissinger's geheime (en hilarische) trip naar Beijing https://art19.com/shows/betrouwbare-bronnen/episodes/cff20ade-b4b1-47a8-b554-0fccc620e096 447 - Als Trump wint staat Europa er alleen voor https://art19.com/shows/betrouwbare-bronnen/episodes/eee9ebfb-042b-4753-b70d-a48e915b5beb 488 - Het Congres van Wenen (1814-1815) als briljant machtsspel https://art19.com/shows/betrouwbare-bronnen/episodes/1423134d-c671-4a71-805a-1d21ab9f7de6 *** Tijdlijn 00:00:00 – Deel 1 00:36:16 – Deel 2 00:54:06 – Deel 3 01:20:19 – EindeSee omnystudio.com/listener for privacy information.

Ein neuer Sonntag, eine neue Folge Edeltalk für euch auf die Ohren! Heute sprechen wir über Sexualität, Fetische und darüber, ob wir vielleicht alle keine Pornos mehr schauen sollten – und warum im Büro plötzlich alle Dokumente zusammenkleben. Viel Spaß beim Zuhören! Wenn du Unterstützung brauchst oder mit jemandem sprechen willst, findest du hier anonyme und professionelle Hilfe: Nummer gegen Kummer: https://www.nummergegenkummer.de Therapiesuche (Sexualtherapie): https://www.profamilia.de/angebote-vor-ort

The common story of modern China's development is that it has two ages: Mao, and reform. The truth is there are at least six internally coherent economic eras within the country's journey from basket case to superpower. Each with their own rules and obsessions. That's certainly the view of Philip Pilkington, who has been crunching the deep data on the Chinese economy, in a new paper for Eurasia Magazine. This week, in an hour long special, Andrew Collingwood quizzes him on the particularities of these periods: from the black-and-white-cats of Deng, to the red-in-tooth-and-claw market mercantilism of Hu Jintao, up to Xi's property sinking funds and robot army. As Philip argues, most US Republicans still imagine that the central danger of China is that it trades unfairly - in truth, the country has moved on from that point on the global value chain.Remember you can get special paywalled premium episodes of Multipolarity every month on Patreon: https://patreon.com/multipolarity or by becoming a member on our YouTube Channel (just click Join).

After Dark with Hosts Rob & Andrew – Simon Deng recounts his childhood enslavement and warns Americans about rising extremism he associates with radical Islamist movements. Speaking from his experiences in Sudan and the United States, he urges vigilance as he highlights ongoing human rights abuses, cultural pressures, and political shifts he believes threaten freedom in communities across the country today and beyond...

After Dark with Hosts Rob & Andrew – Simon Deng recounts his childhood enslavement and warns Americans about rising extremism he associates with radical Islamist movements. Speaking from his experiences in Sudan and the United States, he urges vigilance as he highlights ongoing human rights abuses, cultural pressures, and political shifts he believes threaten freedom in communities across the country today and beyond...

Prekybos ir pramogų centrai jau ruošiasi Kalėdoms ir šventiniams išpardavimams. Praeitų metų duomenys rodo, kad vienas iš keturių pirkėjų šventinius pirkinius dengė kreditu ar skolintais pinigais. Ar verta dovanų ieškoti skolon ir ar „paskutinės minutės“ pasiūlymai – gera proga sutaupyti?Ištrauka iš tinlalaidės „Prieš atsisveikinant“: ką daryti, jei būdami šalia sunkiai sergančių, mirties link artėjančių žmonių nerandame tinkamų žodžių prabilti sunkiomis temomis?Išmanieji įrenginiai ir įvairios programėlės tapo neatsiejama daugelio sportuojančiųjų kasdienybės dalimi. Jie skaičiuoja žingsnius, pulsą, miegą, treniruočių krūvį ir net įspėja, kada reikia pailsėti. Tačiau ar visada šie įrenginiai padeda, o skaičiai nusako realią situaciją? Galiausiai, ar išmanusis laikrodis gali žinoti žmogaus savijautą geriau už jį patį?KTU mechanikos inžinerijos doktorantė stažavosi NASA tyrimų centre, kur kūrė sistemą, padedančią robotams dirbti tiksliau.Ved. Ignas Andriukevičius

More Chapters AI Contribution: Courtesy of Google NotebookLM

More Chapters AI Contribution: Courtesy of Google NotebookLM

‘There's basically a total failure of governance. Nothing is working. There is very serious political instability that has actually manifested itself in violence all across the country.'James Copnall speaks to senior South Sudanese politician Nhial Deng Nhial, about the political instability that's threatening the future of the world's youngest country.Nhial, who has served in numerous important roles in, including as South Sudan's Foreign Minister, made headlines last month when he openly criticised the country's president, Salva Kiir, going from former ally to a voice of opposition. South Sudan gained independence from Sudan in 2011 after decades of struggle led by the Sudan People's Liberation Movement, or SPLM. But just two years into independence, a power struggle between President Salva Kiir and his vice-president Riek Machar led to a civil war in which 400,000 lost their lives. The civil war was brought to an end in 2018 through a peace deal that saw the creation of a unity government that was meant to pave the way for elections in 2022. However, these never happened, and following renewed clashes between the army and opposition forces earlier this year. The United Nations is deeply concerned about a possible return to outright war. Amid these growing tensions in the country, even once-staunch allies of the President are now speaking out - including Nhial Deng Nhial who suspended his membership of the ruling SPLM, and accused the government of corruption, fuelling insecurity and refusing to hold long-delayed elections. President Salva is no stranger to public criticism - but there was a sense of shock that the latest to take aim at South Sudan's leader was Nhial Deng Nhial, a prominent figure in the country, and the party, for decades. The Interview brings you conversations with people shaping our world, from all over the world. The best interviews from the BBC. You can listen on the BBC World Service, Mondays and Wednesdays at 0700 GMT. Or you can listen to The Interview as a podcast, out twice a week on BBC Sounds or wherever you get your podcasts.Presenter: James Copnall Producer: Ben Cooper Editor: Justine LangGet in touch with us on email TheInterview@bbc.co.uk and use the hashtag #TheInterviewBBC on social media.(Image: Nhial Deng Nhial Credit: HANNAH MCNEISH/AFP/GettyImages)

More Chapters AI Contribution: Courtesy of Google NotebookLM

More Chapters AI Contribution: Courtesy of Google NotebookLM

More Chapters AI Contribution: Courtesy of Google NotebookLM

Why do leaders with vast expert bureaucracies at their fingertips make devastating foreign policy decisions? Tyler Jost, professor at Brown, joins ChinaTalk to discuss his first book, Bureaucracies at War, a fascinating analysis of miscalculation in international conflicts. As we travel from Mao's role in border conflicts, to Deng's blunder in Vietnam, to LBJ's own Vietnam error, a tragic pattern emerges — leaders gradually isolating themselves from their own information gathering systems with catastrophic consequences. Today our conversation covers… How Mao's early success undermined his long-term decision-making, The role of succession pressures in both Deng's and LBJ's actions in Vietnam, The bureaucratic mechanisms that lead to echo chambers, and how China's siloed institutions affect Xi's governance, The lingering question of succession in China, What we can learn from the institutional failures behind Vietnam and Iraq. Learn more about your ad choices. Visit megaphone.fm/adchoices

Why do leaders with vast expert bureaucracies at their fingertips make devastating foreign policy decisions? Tyler Jost, professor at Brown, joins ChinaTalk to discuss his first book, Bureaucracies at War, a fascinating analysis of miscalculation in international conflicts. As we travel from Mao's role in border conflicts, to Deng's blunder in Vietnam, to LBJ's own Vietnam error, a tragic pattern emerges — leaders gradually isolating themselves from their own information gathering systems with catastrophic consequences. Today our conversation covers… How Mao's early success undermined his long-term decision-making, The role of succession pressures in both Deng's and LBJ's actions in Vietnam, The bureaucratic mechanisms that lead to echo chambers, and how China's siloed institutions affect Xi's governance, The lingering question of succession in China, What we can learn from the institutional failures behind Vietnam and Iraq. Learn more about your ad choices. Visit megaphone.fm/adchoices

More Chapters AI Contribution: Courtesy of Google NotebookLM

In this episode we're talking to Deng. Deng has been a creative consultant in the advertising industry for over 20 years. Deng tells us stories about her childhood home, which turned out to be worst kind of energy magnet ever, how she had her third eye closed by a feng shui master up in the mountains of China, an angry tree spirit in her former workplace, and a doppelganger posing as officemate that their poor kuya guard up and quit. Grimcast Manila is a podcast of creepy stories from around the metro. Hosted by self-confessed scaredy-cat Antonette, the show invites a series of guests to come tell their ghost stories and other creepy tales. From things that go bump in the night to ghosts that seem to want to tag along, Grimcast features tales of real first-hand experiences.  Grimcast Manila is produced, recorded, and distributed by Big Baby Studios.

Jessica reports LIVE from Jakarta while Spencer analysis every detail from GymCastic headquarters on the first day of event finals! World Championships Headquarters Get for all Jakarta Worlds Videos, Interviews, Podcasts, Fantasy, Guides Extended Episode + Live Q&A (Members) +30 extra minutes of analysis, behind-the-scenes secret stories, and answering your questions. Here's how to ask questions live. Can't make it live? Add Club bonus episodes to your favorite podcast player (instructions here). Tip: After logging in, refresh this page and the extended player will appear below.  Headlines IOC stops Olympic talks with Indonesia over Israeli athlete ban How to Report Exploitative Photography during a FIG meet  Contact the FIG and LOC safeguarding officers on site. They are listed in the work plan, which is accessible on the event page (e.g. Jakarta: https://live.gymnastics.sport/event_detail.php?idevent=17810 They can also be reached by phone or WhatsApp. Anonymous reports can be filed directly to the Gymnastics Ethics Foundation   FIG Safeguarding page Chapters 00:00 – Intro & Sponsors — Gymnastics Medicine, Club Gym Nerd 02:00 – Welcome from Jakarta: Jessica & Spencer on Day 1 of Event Finals 03:40 – Headlines: IOC vs. Indonesia, Fujitsu robots & FIG ethics 08:10 – Are the medals light or heavy? 08:35 – Women's Vault Final 09:00 – Melnikova, Fontaine & Josc medal recap 09:45 – Deng's vault crash & DNS rule explained 12:30 – Antwerp flashback & Voinea precedent 14:15 – Valen's “no-pike” Rudi & judging notes 15:40 – Kalmykova, Schönemaier & Fontaine highlights 20:05 – Melnikova's Cheng vs. form deductions 21:30 – Vault wrap-up 22:20 – Women's Uneven Bars Final 22:45 – Hit-a-thon! Skye Blakely sticks 24:20 – Melnikova & McDonald clean hits 26:10 – Yang's no-release issue 27:30 – Zoya's one-leg heroic routine 29:20 – Bars recap 30:00 – Men's Floor Final 30:25 – Jake Jarman's triple-double clinic 32:05 – Luke Whitlock & Yulo analysis 34:10 – Minami's honest fall 35:25 – Milad's Shushunova & artistry talk 37:05 – Floor medal recap 38:00 – Pommel Horse 38:20 – Highlights & scoring notes 41:00 – PH results 41:40 – Rings Final 42:00 – Whittenburg, World Champion at 31 43:20 – Adem celebration & medal reaction 46:00 – Nelson's style points 48:20 – Awards of the Day & BTS Teaser 48:40 – Best routines, surprises & Club Gym Nerd info 52:00 – Live Q&A & upcoming finals preview 54:35 – Show Close 55:00 – Tomorrow's coverage preview & sign-off from Jakarta How Do I Watch the Competition? All sessions of the competition will be streamed on Eurovision Sport. Follow along here! Gymnastics Indonesia's YouTube channel will stream all qualification sessions Live scores from the FIG and Swiss Timing Check out NBC's behind-the-scenes mini-doc on the US Women's World Trials US viewers check out Peacock and NBC broadcast schedlue here. GymCastic Updates Subscribe to our YouTube Channel Coming Up 6 days of LIVE podcasts at World Championships in Jakarta Club members get extended coverage and can join us live to ask questions immediately after the meet Play our World Championships Fantasy Game! Win a Club Gym Nerd Scholarship: Go to our Forum > Show Stuff > GymCastic Scholarship We are matching every new sponsorship If you would like access to the club content, but aren't currently in a position to purchase a membership, all you need to do is fill out the form that's linked in our message board If you would also like to sponsor a scholarship, please email editor@gymcastic.com. Thank you! Support Our Work Club Gym Nerd: Join Here Become a Sponsor: GymCastic is matching all donations Nearly 50 scholarships have been awarded so far Learn More Headstand Game: Play Now Forum: Start Chatting Merch: Shop Now Thank you to our Sponsors Gymnastics Medicine Beam Queen Bootcamp's Overcoming Fear Workshop Huel Daily Greens Ready to Drink - Get 15% off your purchase for New Customers with our exclusive code GYMCASTIC at huel.com/GYMCASTIC. Use our code and fill out the post checkout survey to help support the show! Resources Jakarta schedule & times: See our live podcast times on the Worlds HQ schedule Guides: Download the quick-reference guide on the Jakarta Headquarters page The Balance Beam Situation: Spencer's GIF Code of Points Gymnastics History and Code of Points Archive from Uncle Tim Kensley's men's gymnastics site Neutral Deductions Unlock the Extended Episode Join Club Gym Nerd → Choose a plan Complete checkout — your site account is created. Log in here → /my-account/ Return to this page and refresh. The extended player appears automatically.

Lääth yic jäl Nhial Deng Nhial SPLM yic ke ci Atem Yaak Atem kuanyic ke lɛ̈ɛ̈thic.

Fresh out of the studio, Yuying Deng, Co-founder and CEO of Esevel, shares her transformative journey from corporate lawyer to healthcare operator to tech entrepreneur with our guest host Yana Fry from Yana TV. Yuying discusses how the pandemic's sudden shift to remote work in April 2020 revealed critical gaps in IT infrastructure for distributed teams, inspiring her to launch Esevel—a platform now serving companies across 88 countries. Yuying challenges the traditional HQ-centric worldview, advocating that "HQ should be a mindset, not a location," and shares how Esevel deliberately builds leadership opportunities for talented professionals regardless of whether they're based in Manila, Singapore, or São Paulo. Last but not least, Yuying shares what great would look like for Esevel's future: becoming the indispensable tool companies think of first when scaling global teams, while proving that talent and performance matter more than location."Many companies that say they do distributed and remote work actually still have a very HQ-centric worldview. That means leadership is in HQ, strategy is formed in HQ, and high-impact jobs are in the HQ as well. So when they hire remote and distributed teams. For example, in the Philippines, Brazil, and India they use these more as back-office functions. So you have very talented people who join them there, thinking that they could rise in a global company. But very soon they find that they hit a glass ceiling and are no longer able to advance, and so they move on to another firm. I think that's a massive waste of talent, especially if you're talking about here in Asia. This is the world's fastest-growing region. People are ambitious, people are bright, and they are able to take on leadership positions if they're given the opportunity to. This is one thing that we have really tried to reverse at Esevel. You do not have to be at HQ in order to rise into a leadership position. As long as you perform your job and perform it well, we look at performance more than location. So I think that is one thing that has to shift: HQ shouldn't be a location. HQ should actually be a mindset. And I think that's something that a lot of remote companies or distributed work companies have correct when it comes to that." - Deng YuyingEpisode Highlights:[00:00] Quote of the Day by Deng Yuying[02:00] Introducing Yuying Deng, CEO of Esevel[02:25] From lawyer to operator to founder[03:32] MBA at INSEAD shaped entrepreneurial journey[03:53] Built community care division for Orange Valley[04:24] Family business dynamics and PE exit lessons[05:44] Esevel: IT operations platform for distributed teams[07:56] Company DNA shaped by pandemic remote work[08:38] Importance of staying close to customer problems[10:16] Managing operations across 88 countries globally[12:39] Failure is a feature, not a bug[14:33] Operational complexity and doing boring work well[16:35] Future of hybrid and remote work[19:48] HQ should be a mindset not location[21:25] Characteristics needed for remote work success[22:40] Growth opportunities regardless of employee location[24:58] Founding a company is like raising child[26:52] No perfect time for major life decisions[29:31] Ethical principles learned from parents[30:33] Vision for Esevel and family independence[32:28] Partnership requires mutual support for success[35:48] Rising through adversity with determination[36:34] Legacy focused on happy, independent childrenProfile: Yuying Deng, CEO of Esevel: https://esevel.comLinkedIn Profile: https://www.linkedin.com/in/yuyingdeng/Guest Host: Yana Fry from Yana TV:  https://www.youtube.com/@yanatvsgLinkedIn Profile: https://www.linkedin.com/in/yanafry/Podcast Information: Bernard Leong hosts and produces the show. The proper credits for the intro and end music are "Energetic Sports Drive." G. Thomas Craig mixed and edited the episode in both video and audio format.

Host Anne Thompson explores the diverse and innovative methods being employed in Australia to build and retain the technical workforce in the mining industry.  From attracting new talent, to supporting early career geoscientists, to training professionals from other industries - this episode might just change your perspective on what is possible.First up, Brendan Howard, General Manager for Technical Capability at Rio Tinto and a founder of the MiEX (Minerals Industry Experience) Program, addresses the mining sector's shrinking and less diverse talent pipeline. The success of MiEX, a collaborative industry-led program, highlights the value of early engagement with undergraduate students. The pilot in January 2025 drew over 500 applicants for 69 places with 90% of participants expressing intent to pursue mining careers.  New conversations are underway in North America to bring this exciting opportunity to more students.Richard Lilly, the University of Adelaide presents NExUS (National Exploration Undercover School) based in South Australia, which provides hands-on experience and essential training to geoscience students, bridging the gap between academic knowledge and industry requirements. NExUS combines technical learning with professional networking, using the South Australian Drill Core Library and nearby field sites to expose students to geophysics, core logging, and exploration through cover. Lilly is now stepping down after 10 years as Director, with the hope that this successful model will continue and be expanded to other jurisdictions to better prepare geoscientists for evolving exploration challenges.Lastly, Deng Ngang Deng, from Target Mining Resources, shares his inspiring journey from South Sudan to professional geoscientist in Western Australia.  Deng was involved in establishing the Pan Africa Resource Reporting Code (PARC) and an invited plenary speaker at SEG 2024 in Namibia. At home in Australia, he has developed an innovative approach to training and employing geoscientists and other technical mining workers.  Initiated during the pandemic, the company continues to expand and add to their portfolio of skills they train for, successfully filling a gap in mining recruitment.Many thanks to Avo Media for production support.  Theme music is Confluence by Eastwindseastwindsmusic.com 

Hur mår den svenska järnvägen? Är den sjuk eller frisk? Vetenskapsradion tar tempen på det svenska järnvägsnätet. Lyssna på alla avsnitt i Sveriges Radio Play. Programmet sändes första gången 13/5 – 2025.Det var på Valborgsmässoafton 1855 i Skaveryd, i närheten av Alingsås, som man satte spaden i jorden för att påbörja Sveriges första järnvägsbygge, sedan följde en intensiv utbyggnad och under första hälften av 1900-talet så var järnvägen utbyggd ungefär så som vi känner den i dag.Och faktum är att det till stora delar är samma järnväg som vi fortfarande åker på som man gjorde för 100 år sedan. Naturligtvis är en hel material utbytt men inte tillräckligt tycker Sebastian Stichel som är järnvägsforskare på Kungliga Tekniska Högskolan:”Egentligen skulle man behöva stänga av, sträcka för sträcka, och byta ut alla komponenter och börja med en modern ny järnväg”Vi besöker Järnvägsmuseet i Gävle för att få reda den svenska järnvägens historia, vi får träffa en rallare som var med och byggde Inlandsbanan mellan mellan Dorotea-Storuman och Sorsele och vi får veta hur man bygger en järnväg.Medverkande: Henrik Reuterdahl, Järnvägsmuseet i Gävle, och Sebastian Stichel, järnvägsforskare på Kungliga Tekniska Högskolan.Reporter: Joacim Lindwalljoacim.lindwall@sr.seProducent: Lars Broströmlars.brostrom@sr.se

Hur tänker en narcissist och varför beter den sig som den gör? I det här avsnittet berättar psykiatern Peder Björling, som även är legitimerad psykoterapeut, om narcissism på djupet och vad det egentligen är. Peder Björling är också aktuell med boken "Störst, bäst och skörast : en bok om narcissism" som gavs ut i maj.Jag har blivit nominerad till Årets Programledare för Podd i Guldörat 2025. Jag skulle bli så glad om ni vill rösta på mig: https://www.guldorat.se/rosta-2025Hela säsongen av Älskade Psykopat finns på podplay.se och i podplay-appen: https://www.podplay.com/sv-se/podcasts/alskade-psykopat-294350Följ @alskadepsykopat på Instagram.

On this timely and informative episode of The Member Engagement Show, Dorothy Deng, Partner at Whiteford, Taylor & Preston discusses AI and copyright law, especially as it relates to association management. There are many legal considerations for associations using AI and this is an excellent primer on the copyright fundamentals to stay safe and within the law. Topics covered include: Fundamental copyright considerations and how they that relate to AI use. The embedded principles of copyright law - human authorship and fair use. Does writing AI prompts afford you ownership of the AI output? Why you're on (a little) safer ground if your prompt includes pre-existing human created copyrighted material. Why those “This material was AI generated” labels might prevent you from having a copyright claim. Does the percentage of human vs. AI involvement in a work matter to the copyright office? How to protect yourself from unwittingly infringing on other people's copyrights. How associations should keep copyright in mind when using AI. The licensing opportunities for associations as LLMs seek higher quality data.   Some Helpful Links: The U.S. Copyright Office's Report on AI  

JBS Strategic Analyst David Harris meets Simon Deng, a South Sudanese freedom fighter, humanitarian activist, Christian, former slave and vocal advocate for Israel and the release of Hamas hostages.

In this episode, Jacob sits down with Peter Deng, General Partner at Felicis and former Product Leader at OpenAI, Facebook, and Uber. Peter shares his insider perspective on building ChatGPT Enterprise in just seven weeks and leading voice mode development at OpenAI. The conversation covers everything from why traditional SaaS pricing models are broken for AI products to how evals became the new product specs, the "AI under your fingernails" test for founding teams, and why current agents are massively overhyped.They also explore how consumer AI will fragment across multiple winners rather than consolidate into a single super app, the coming integration between ChatGPT and apps like Uber, and why voice AI will unlock entirely new categories of applications. Plus, insights on the changing dynamics between foundation models and startups, and what it really takes to build defensible AI companies. It's a comprehensive look at AI product strategy from someone who's been at the center of the industry's biggest breakthroughs. (0:00) Intro(1:17) AI Business Models and Pricing Strategies(7:48) Product Development in AI Companies(18:36) The Role of Product Managers in AI(23:06) Voice Interaction and AI(26:43) AI in Education(30:39) Consumer and Enterprise Adoption of AI(33:36) The Impact of AI on Salaries and HR(40:37) The Role of Unique Data in AI Development(49:03) Challenges and Strategies for AI Companies(52:58) The Future of AI and Its Impact on Society(57:31) Reflections on OpenAI(58:38) Quickfire With your co-hosts: @jacobeffron - Partner at Redpoint, Former PM Flatiron Health @patrickachase - Partner at Redpoint, Former ML Engineer LinkedIn @ericabrescia - Former COO Github, Founder Bitnami (acq'd by VMWare) @jordan_segall - Partner at Redpoint

TSN Golf Analyst Mark Zecchino joined OverDrive to discuss the headlines around the Tour Championship and CPKC Women's Open, Aphrodite Deng's incredible start, how her ball striking has contributed, Scottie Scheffler's greatness and his comparison to Tiger Woods, Brooke Henderson's inconsistency issues and more.

Matt acknowledges the tragic passing over the weekend of Omaha Basketball player Deng Mayar.

Consider DONATING to help us continue and expand our media efforts. If you cannot at this time, please share this video with someone who might benefit from it. We thank you for your support! https://tinyurl.com/HereIAmWithShaiDavidai NEW SUPPORT ME ON PATREON! https://www.patreon.com/ShaiDavidai --------- Guest: Simon Deng Wikipedia: https://en.wikipedia.org/wiki/Simon_Deng Website: https://africanjewishalliance.org/ IG: https://www.instagram.com/africanjewishalliance/?hl=en In Part 2 of 2 of "Here I Am with Shai Davidai," host Shai continues his powerful conversation with Simon Deng, a former Sudanese slave and human rights activist. Simon shares his journey from Sudan to the United States, his advocacy through the Sudan Freedom Walk, and his efforts to raise awareness about genocide and human rights abuses. The episode explores Simon's solidarity with victims worldwide, including his recent walk in Israel to support hostage families, and his reflections on hope, faith, and the universal fight against evil. Don't miss this moving conclusion to Simon's inspiring story.

Consider DONATING to help us continue and expand our media efforts. If you cannot at this time, please share this video with someone who might benefit from it. We thank you for your support! https://tinyurl.com/HereIAmWithShaiDavidai NEW SUPPORT ME ON PATREON! https://www.patreon.com/ShaiDavidai --------- Guest: Simon Deng Wikipedia: https://en.wikipedia.org/wiki/Simon_Deng Website: https://africanjewishalliance.org/ IG: https://www.instagram.com/africanjewishalliance/?hl=en In this powerful episode (Part 1 of 2) of "Here I Am," host Shai Davidai sits down with Simon Deng, a former South Sudanese slave and renowned human rights activist. Simon shares his harrowing childhood experiences—growing up in a war-torn village, witnessing violence, and ultimately being kidnapped and forced into slavery at the age of nine. He recounts the trauma of being separated from his family, the brutal realities of life as a child slave, and the resilience that helped him survive. Simon also reflects on his journey to freedom and his mission to be a voice for the voiceless, using his story to advocate for human rights and justice. This deeply moving conversation sheds light on the ongoing struggles faced by many and the enduring power of hope and activism.

We honor the absolute rockstar exit Ozzy made from this world. What ever happened to ant farms? Drinking water is for cucks. Plus, Annabelle was no where near the scene of the crime. Cody will help your zoo. Plus so much more on a Wednesdee!

Adoring fans from around the world converged in Thailand this week to celebrate the first birthday of Moo Deng, the baby pygmy hippo who became a social media sensation. John Yang reports. PBS News is supported by - https://www.pbs.org/newshour/about/funders

Leah found a stash of off-brand Labubus, HBD to our shiny, sassy queen Moo Deng and if you missed your chance at Como Zoo there's another stinky corpse flower about to bloom See Privacy Policy at https://art19.com/privacy and California Privacy Notice at https://art19.com/privacy#do-not-sell-my-info.

Joseph Torigian's The Party's Interest Comes First: The Life of Xi Zhongxun, Father of Xi Jinping is a monumental scholarly achievement — easily a contender for one of the best China books of the decade. Joseph's goal, in his own words, was to “shine as much light into the darkness of the past as possible” to understand the nature of authoritarian politics, and he succeeds beyond my wildest expectations. This biography gives me a feel for Chinese politics that I honestly thought I'd never have. It does an incredible job of digging deep to shed light on some of the most consequential moments in CCP history, as well as conveying what it was like to live as a senior official under Mao and Deng. Reading it was a powerful experience at both an intellectual and human level. We get memorable vignettes, like 15-year-old Xi Zhongxun attempting to assassinate a teacher, or General Peng Dehuai using his shoe to silence Xi Zhongxun's snoring in their shared bunk. In this interview, we discuss: What we can learn about authoritarianism, the CCP, and China's future from studying Xi's father, Torigian's methodology for uncovering hidden Party history, How the Party became an existential source of meaning, and how it weaponized suffering to paradoxically deepen political loyalty, The arc of Xi Zhongxun's life — from a young revolutionary to key advocate of reform — and his role during Tiananmen, The interplay of family, love, and career under the all-encompassing shadow of the Party, The role of “Surrogate fathers” and patronage in navigating political ascent, How literature shaped China's early revolutionaries, and even impacted the Party as we know it today. Co-hosting today is Jon Sine, former ChinaTalk intern. Outro music: The Temptations - Papa Was A Rolling Stone (YouTube Link) Learn more about your ad choices. Visit megaphone.fm/adchoices

Peter Deng has led product teams at OpenAI, Instagram, Uber, Facebook, Airtable, and Oculus and helped build products used by billions—including Facebook's News Feed, the standalone Messenger app, Instagram filters, Uber Reserve, ChatGPT, and more. Currently he's investing in early-stage founders at Felicis. In this episode, Peter dives into his most valuable lessons from building and scaling some of tech's most iconic products and companies.What you'll learn:1. Peter's one‑sentence test for hiring superstars2. Why your product (probably) doesn't matter3. Why you don't need a tech breakthrough to build a huge business4. The five PM archetypes, and how to build a team of Avengers5. Counterintuitive lessons on growing products from 0 to 1, and 1 to 1006. The importance of data flywheels and workflows—Brought to you by:Paragon—Ship every SaaS integration your customers wantPragmatic Institute—Industry‑recognized product, marketing, and AI training and certificationsContentsquare—Create better digital experiences—Where to find Peter Deng:• X: https://x.com/pxd• LinkedIn: https://www.linkedin.com/in/peterxdeng/—In this episode, we cover:(00:00) Introduction to Peter Deng(05:41) AI and AGI insights(11:35) The future of education with AI(16:53) The power of language in leadership(21:01) Building iconic products(36:44) Scaling from zero to 100(41:56) Balancing short- and long-term goals(47:12) Creating a healthy tension in teams(50:02) The five archetypes of product managers(55:39) Primary and secondary archetypes(58:47) Hiring for growth mindset and autonomy(01:15:52) Effective management and communication strategies(01:19:23) Presentation advice and self-advocacy(01:25:50) Balancing craft and practicality in product management(01:30:40) The importance of empathy in design thinking(01:35:45) Career decisions and learning opportunities(01:42:05) Lessons from product failures(01:45:42) Lightning round and final thoughts—Referenced:• OpenAI: https://openai.com/• Artificial general intelligence (AGI): https://en.wikipedia.org/wiki/Artificial_general_intelligence• Head of ChatGPT answers philosophical questions about AI at SXSW 2024 with SignalFire's Josh Constine: https://www.youtube.com/watch?v=mgbgI0R6XCw• Professors Are Using A.I., Too. Now What?: https://www.npr.org/2025/05/21/1252663599/kashmir-hill-ai#:~:text=Now%20What• Herbert H. Clark: https://web.stanford.edu/~clark/• Russian speakers get the blues: https://www.newscientist.com/article/dn11759-russian-speakers-get-the-blues/• Ilya Sutskever (OpenAI Chief Scientist)—Building AGI, Alignment, Future Models, Spies, Microsoft, Taiwan, & Enlightenment: https://www.dwarkesh.com/p/ilya-sutskever• Anthropic's CPO on what comes next | Mike Krieger (co-founder of Instagram): https://www.lennysnewsletter.com/p/anthropics-cpo-heres-what-comes-next• Kevin Systrom on LinkedIn: https://www.linkedin.com/in/kevinsystrom/• Building a magical AI code editor used by over 1 million developers in four months: The untold story of Windsurf | Varun Mohan (co-founder and CEO): https://www.lennysnewsletter.com/p/the-untold-story-of-windsurf-varun-mohan• Microsoft CPO: If you aren't prototyping with AI, you're doing it wrong | Aparna Chennapragada: https://www.lennysnewsletter.com/p/microsoft-cpo-on-ai• The rise of Cursor: The $300M ARR AI tool that engineers can't stop using | Michael Truell (co-founder and CEO): https://www.lennysnewsletter.com/p/the-rise-of-cursor-michael-truell• Building Lovable: $10M ARR in 60 days with 15 people | Anton Osika (CEO and co-founder): https://www.lennysnewsletter.com/p/building-lovable-anton-osika• Granola: https://www.granola.ai/• Inside Bolt: From near-death to ~$40m ARR in 5 months—one of the fastest-growing products in history | Eric Simons (founder and CEO of StackBlitz): https://www.lennysnewsletter.com/p/inside-bolt-eric-simons• OpenAI's CPO on how AI changes must-have skills, moats, coding, startup playbooks, more | Kevin Weil (CPO at OpenAI, ex-Instagram, Twitter): https://www.lennysnewsletter.com/p/kevin-weil-open-ai• Fidji Simo on LinkedIn: https://www.linkedin.com/in/fidjisimo/• Airtable: https://www.airtable.com/• George Lee on LinkedIn: https://www.linkedin.com/in/geolee/• Andrew Chen on LinkedIn: https://www.linkedin.com/in/andrewchen/• Lauryn Motamedi on LinkedIn: https://www.linkedin.com/in/laurynmotamedi/• Twilio: https://www.twilio.com/• Nick Turley on LinkedIn: https://www.linkedin.com/in/nicholasturley/• Ian Silber on LinkedIn: https://www.linkedin.com/in/iansilber/• Thomas Dimson on LinkedIn: https://www.linkedin.com/in/thomasdimson/• Joey Flynn on LinkedIn: https://www.linkedin.com/in/joey-flynn-8291586b/• Ryan O'Rourke's website: https://www.rourkery.com/• Joanne Jang on LinkedIn: https://www.linkedin.com/in/jangjoanne/• Behind the founder: Marc Benioff: https://www.lennysnewsletter.com/p/behind-the-founder-marc-benioff• Jill Hazelbaker on LinkedIn: https://www.linkedin.com/in/jill-hazelbaker-3aa32422/• Guy Kawasaki's website: https://guykawasaki.com/• Eric Antonow on LinkedIn: https://www.linkedin.com/in/antonow/• Sachin Kansal on LinkedIn: https://www.linkedin.com/in/sachinkansal/• IDEO design thinking: https://designthinking.ideo.com/• The 7 Steps of the Design Thinking Process: https://www.ideou.com/blogs/inspiration/design-thinking-process• Linear's secret to building beloved B2B products | Nan Yu (Head of Product): https://www.lennysnewsletter.com/p/linears-secret-to-building-beloved-b2b-products-nan-yu• Jeff Bezos's quote: https://news.ycombinator.com/item?id=27778175• Friendster: https://en.wikipedia.org/wiki/Friendster• Myspace: https://en.wikipedia.org/wiki/Myspace• How LinkedIn became interesting: The inside story | Tomer Cohen (CPO at LinkedIn): https://www.lennysnewsletter.com/p/how-linkedin-became-interesting-tomer-cohen• “Smile” by Jay-Z: https://www.youtube.com/watch?v=SSumXG5_rs8&list=RDSSumXG5_rs8&start_radio=1• The Wire on HBO: https://www.hbo.com/the-wire• Felicis: https://www.felicis.com/—Recommended books:• Sapiens: A Brief History of Humankind: https://www.amazon.com/Sapiens-Humankind-Yuval-Noah-Harari/dp/0062316095• The Design of Everyday Things: https://www.amazon.com/Design-Everyday-Things-Revised-Expanded/dp/0465050654• The Silk Roads: A New History of the World: https://www.amazon.com/Silk-Roads-New-History-World/dp/1101912375—Production and marketing by https://penname.co/. For inquiries about sponsoring the podcast, email podcast@lennyrachitsky.com.—Lenny may be an investor in the companies discussed. This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit www.lennysnewsletter.com/subscribe

From the Taiping and Boxer Rebellions, to the Chinese Revolution and Civil War, through the Long March and the rise of Mao Zedong, to the Great Leap Forward and Cultural Revolution, all the way to Deng's Reform and China today, Professor of East Asian and Global History Dr. Ken Hammond walk us through 200 years of Chinese history to highlight in detail how modern China was forged through centuries of class struggle, resistance, rebellion, and revolution.  After listening to this mega-episode you will have a profound, and deeply inspired, understanding of the rich modern history of China, and be much better able to understand its present and future.   This series originally aired on Guerrilla History in the Spring of 2024 Support Guerrilla History HERE Learn More, Follow, and Support Rev Left Radio HERE