Video yükleniyor...

Video Yüklenemedi

Ana Sayfaya Dön

Introducing Digital Red Queen (DRQ): Adversarial Program Evolution in Core War with LLMs Blog: Core War is a programming game where self-replicating assembly programs, called warriors, compete for control of a virtual machine. In this dynamic environment, where there is no distinction between code and data, warriors must crash...

SakanaAILabs's profile picture

Sakana AI

136,026 subscribers

143,831 görüntüleme • 6 ay önce •via X (Twitter)

EğitimBilim & Teknoloji
LIVE

Anya Rossi• Live Now

Private livecam show

0 Yorum

Yorum bulunmuyor

Orijinal gönderinin yorumları burada görünecek

Benzer Videolar

Survival of the fittest code. Core War (1984) is a game where programs must crash their opponents to survive. Warriors written in an assembly language called Redcode fight for control of a virtual machine. Our new paper: Digital Red Queen: Adversarial Program Evolution in Core War with LLMs, explores what happens when LLMs drive an adversarial evolutionary arms race in this domain. We task LLMs to write Warrior programs in Redcode that must out-compete a virtual world full of such programs. Core War is a Turing-complete environment where code and data share the same address space, which leads to some very chaotic self-modifying code dynamics. This approach is inspired by the Red Queen hypothesis in evolutionary biology: the principle that species must continually adapt and evolve simply to survive against ever changing competitors. In our work, programs continuously adapt to defeat a growing history of opponents rather than a static benchmark. We find that this adversarial process leads to the emergence of increasingly general strategies, including targeted self-replication, data bombing, and massive multithreading. Most intriguingly, it reveals a form of convergent evolution. Different code implementations settle into similar high performing behaviors, mirroring how biological agents independently evolve similar traits to solve the same problems. I think this work positions Core War as a sandbox for studying Red Queen dynamics in artificial systems. It offers a safe controlled environment for analyzing how AI agents might evolve in real world adversarial settings such as cybersecurity. By simulating these adversarial dynamics in an isolated sandbox, we offer a glimpse into the future where deployed LLM systems may start competing against one another for limited resources in the real world.
0:33

Survival of the fittest code. Core War (1984) is a game where programs must crash their opponents to survive. Warriors written in an assembly language called Redcode fight for control of a virtual machine. Our new paper: Digital Red Queen: Adversarial Program Evolution in Core War with LLMs, explores what happens when LLMs drive an adversarial evolutionary arms race in this domain. We task LLMs to write Warrior programs in Redcode that must out-compete a virtual world full of such programs. Core War is a Turing-complete environment where code and data share the same address space, which leads to some very chaotic self-modifying code dynamics. This approach is inspired by the Red Queen hypothesis in evolutionary biology: the principle that species must continually adapt and evolve simply to survive against ever changing competitors. In our work, programs continuously adapt to defeat a growing history of opponents rather than a static benchmark. We find that this adversarial process leads to the emergence of increasingly general strategies, including targeted self-replication, data bombing, and massive multithreading. Most intriguingly, it reveals a form of convergent evolution. Different code implementations settle into similar high performing behaviors, mirroring how biological agents independently evolve similar traits to solve the same problems. I think this work positions Core War as a sandbox for studying Red Queen dynamics in artificial systems. It offers a safe controlled environment for analyzing how AI agents might evolve in real world adversarial settings such as cybersecurity. By simulating these adversarial dynamics in an isolated sandbox, we offer a glimpse into the future where deployed LLM systems may start competing against one another for limited resources in the real world.

hardmaru

173,997 görüntüleme • 6 ay önce

*New Paper on AI & Democracy* Imagine two approaches to democracy. The one we have today, where citizens choose a professional politician to represent them and others. Or an augmented form of democracy, where each citizen controls a personalized AI that helps them participate in thousands of nuanced decisions. This second approach is the idea of Augmented Democracy I introduced six years ago at TED. In our latest paper we explore a simplified version of Augmented Democracy by combining off-the-shelf LLMs, such as ChatGPT, with data collected using a collaborative government program builder. This was an online game where people build a personalized government program using proposals extracted from the programs of the candidates of the 2022 presidential election in Brazil. So how accurate are these augmented forms of democracy? Imagine a user who gave us 40 answers. We can use the first 20 to fine-tune a model that we can test using the 20 answers the model didn’t see. We can then compare the accuracy of these predictions with the ones obtained by a “bundle” rule, which assumes that users that self-reported to be from the left or right always chose the proposals from the candidate that shares their political identity. This showed us that LLMs were more accurate at predicting policy preferences than the bundle rule, meaning that the preferences captured in the participation data were more nuanced than a left-right axis, and that the LLMs can capture some of that nuance. Also, the LLMs can choose among policies coming from the same candidate, which is something that we cannot do using a bundle rule. But can these LLMs help us complete the aggregate preferences of the population? Direct or unbundled forms of participation can result in incomplete data when people answer only a fraction of all questions. In our paper, we simulate this incompleteness by sampling the full dataset. We ask how close we can get to the full dataset by using a random sample, or a random sample augmented by predictions made by these LLMs. Overall, we find that LLM-augmented data gets much closer to the full dataset than a pure random sample. These results do not mean that augmented democracy technology is ready, but they means we are in a much better place to continue exploring this idea than six years ago. This paper was a collaborative effort with Jairo Gudino, PhD student at CCL at the University of Toulouse Capitole and Umberto Grandi from IRIT also at the University of Toulouse Capitole. We hope you find these results insightful!
2:16

*New Paper on AI & Democracy* Imagine two approaches to democracy. The one we have today, where citizens choose a professional politician to represent them and others. Or an augmented form of democracy, where each citizen controls a personalized AI that helps them participate in thousands of nuanced decisions. This second approach is the idea of Augmented Democracy I introduced six years ago at TED. In our latest paper we explore a simplified version of Augmented Democracy by combining off-the-shelf LLMs, such as ChatGPT, with data collected using a collaborative government program builder. This was an online game where people build a personalized government program using proposals extracted from the programs of the candidates of the 2022 presidential election in Brazil. So how accurate are these augmented forms of democracy? Imagine a user who gave us 40 answers. We can use the first 20 to fine-tune a model that we can test using the 20 answers the model didn’t see. We can then compare the accuracy of these predictions with the ones obtained by a “bundle” rule, which assumes that users that self-reported to be from the left or right always chose the proposals from the candidate that shares their political identity. This showed us that LLMs were more accurate at predicting policy preferences than the bundle rule, meaning that the preferences captured in the participation data were more nuanced than a left-right axis, and that the LLMs can capture some of that nuance. Also, the LLMs can choose among policies coming from the same candidate, which is something that we cannot do using a bundle rule. But can these LLMs help us complete the aggregate preferences of the population? Direct or unbundled forms of participation can result in incomplete data when people answer only a fraction of all questions. In our paper, we simulate this incompleteness by sampling the full dataset. We ask how close we can get to the full dataset by using a random sample, or a random sample augmented by predictions made by these LLMs. Overall, we find that LLM-augmented data gets much closer to the full dataset than a pure random sample. These results do not mean that augmented democracy technology is ready, but they means we are in a much better place to continue exploring this idea than six years ago. This paper was a collaborative effort with Jairo Gudino, PhD student at CCL at the University of Toulouse Capitole and Umberto Grandi from IRIT also at the University of Toulouse Capitole. We hope you find these results insightful!

César A. Hidalgo

26,915 görüntüleme • 1 yıl önce

A hugely significant day in UK politics no matter where you are. The political landscape has been changing for quite some time and today map was redrawn. But people are united on one thing; change! Where does Northern Ireland fit into all of this? It's clearer than ever to me that while people are talking about Reform the party, we need Reform of Stormont. We don't know what the future holds, but one thing is certain, we need to guard against all eventualities and the last thing we want is to be without a devolved government. We need a robust system that can't collapse. We need to know that whoever is in Govt in Westminster, that we are in charge of our future. No one else can solve or care about our problems apart from us. Reform of Stormont has now just become an absolute essential. We must be able to govern ourselves no matter who is in No 10.
3:05

A hugely significant day in UK politics no matter where you are. The political landscape has been changing for quite some time and today map was redrawn. But people are united on one thing; change! Where does Northern Ireland fit into all of this? It's clearer than ever to me that while people are talking about Reform the party, we need Reform of Stormont. We don't know what the future holds, but one thing is certain, we need to guard against all eventualities and the last thing we want is to be without a devolved government. We need a robust system that can't collapse. We need to know that whoever is in Govt in Westminster, that we are in charge of our future. No one else can solve or care about our problems apart from us. Reform of Stormont has now just become an absolute essential. We must be able to govern ourselves no matter who is in No 10.

Sorcha Eastwood MP

42,349 görüntüleme • 2 ay önce

"This war must be ended, and russia must end it. russia started it and is dragging it out, ignoring any deadlines, and that is the problem, not anything else. The answer to the Ukrainian territorial issue is already in the Constitution of Ukraine. No one will retreat from this and no one can. Ukrainians will not give their land to the occupier. Ukraine is ready for real solutions that can bring peace. Any decisions against us, any decisions made without Ukraine, are simultaneously decisions against peace. They will bring nothing. These are dead decisions; they will never work. And we all need a real, living peace that people will respect. We are ready, together with President Trump, together with all partners, to work for a real and, most importantly, lasting peace—a peace that will not collapse because of Moscow's desires. I thank all our people for being together. Ukraine exists. I thank all our warriors for preserving independence. Stand firm. This is our land, we are Ukraine." Address by Volodymyr Zelensky on the morning of the 1263rd day of the war.
2:27

"This war must be ended, and russia must end it. russia started it and is dragging it out, ignoring any deadlines, and that is the problem, not anything else. The answer to the Ukrainian territorial issue is already in the Constitution of Ukraine. No one will retreat from this and no one can. Ukrainians will not give their land to the occupier. Ukraine is ready for real solutions that can bring peace. Any decisions against us, any decisions made without Ukraine, are simultaneously decisions against peace. They will bring nothing. These are dead decisions; they will never work. And we all need a real, living peace that people will respect. We are ready, together with President Trump, together with all partners, to work for a real and, most importantly, lasting peace—a peace that will not collapse because of Moscow's desires. I thank all our people for being together. Ukraine exists. I thank all our warriors for preserving independence. Stand firm. This is our land, we are Ukraine." Address by Volodymyr Zelensky on the morning of the 1263rd day of the war.

Jürgen Nauditt 🇩🇪🇺🇦

22,919 görüntüleme • 11 ay önce

World Simulator, reimagined — now alive with humans, robots, and their vibrant society unfolding in 3D real-world geospatial scenes across the globe! 🚀 One day soon, humans and robots will co-exist in the same world. To prepare, we must address: 1️⃣ How can robots cooperate or compete intelligently? 2️⃣ How do humans build social bonds and communities? 3️⃣ How can both co-exist in an open, dynamic world? Announcing Virtual Community Project — a social-physical world simulator, where human characters and robotic agents can interact, grow, and co-evolve within open-world societies, stretching from London to New York, and beyond! Key features include: ✅ Unified multi-agent physics simulations for rich social + physical interactions of humans and robots ✅ Massive auto-generated 3D scenes grounded with the rea-world geospatial data ✅ Agent communities populated by robots and LLM-driven human characters with rich appearances, personalities, and social ties. 🌍 Enter our Virtual Community, an open world to study embodied AI at scale— one social-physical world model at a time! 🔗 Project: 💻 Code: Paper: 1/n
0:53

World Simulator, reimagined — now alive with humans, robots, and their vibrant society unfolding in 3D real-world geospatial scenes across the globe! 🚀 One day soon, humans and robots will co-exist in the same world. To prepare, we must address: 1️⃣ How can robots cooperate or compete intelligently? 2️⃣ How do humans build social bonds and communities? 3️⃣ How can both co-exist in an open, dynamic world? Announcing Virtual Community Project — a social-physical world simulator, where human characters and robotic agents can interact, grow, and co-evolve within open-world societies, stretching from London to New York, and beyond! Key features include: ✅ Unified multi-agent physics simulations for rich social + physical interactions of humans and robots ✅ Massive auto-generated 3D scenes grounded with the rea-world geospatial data ✅ Agent communities populated by robots and LLM-driven human characters with rich appearances, personalities, and social ties. 🌍 Enter our Virtual Community, an open world to study embodied AI at scale— one social-physical world model at a time! 🔗 Project: 💻 Code: Paper: 1/n

Chuang Gan

90,261 görüntüleme • 1 yıl önce

🇺🇦 Zelensky: Today we celebrate the Day of the Armed Forces of Ukraine – an independent Ukraine, a Ukraine that knows how to defend itself, a Ukraine that stands in such a way that there is simply no one in the world who could afford to ignore us, our state. And even those in the world who still do not believe in Ukraine or simply do not know enough about Ukraine, are still forced to reckon with how Ukraine will react. Few states in the world have such weight. Even fewer could really stand against such Russian strikes, assaults, and meanness, from which Ukraine is actually defending itself every day and every night. We are deterring a full-scale Russian invasion. We have been holding back Russian murderers for almost four years of full-scale war – after years of hybrid war – we are keeping them behind the front line. Ukraine exists, and the fate of Ukraine really matters, and the world must reckon with this. And this is solely because we have you, Ukrainian warriors, Ukrainian soldiers, Ukrainian sergeants, Ukrainian officers, and everyone who works for our defense, everyone who provides the Armed Forces of Ukraine and all our Defense and Security Forces of Ukraine. I congratulate all of you, Ukrainian warriors, on the Day of the Armed Forces of Ukraine. I am proud of you, Ukrainian defenders, Ukrainian defenders. Glory to the Ukrainian warriors! Glory to the Armed Forces of Ukraine! Glory to Ukraine!
5:45

🇺🇦 Zelensky: Today we celebrate the Day of the Armed Forces of Ukraine – an independent Ukraine, a Ukraine that knows how to defend itself, a Ukraine that stands in such a way that there is simply no one in the world who could afford to ignore us, our state. And even those in the world who still do not believe in Ukraine or simply do not know enough about Ukraine, are still forced to reckon with how Ukraine will react. Few states in the world have such weight. Even fewer could really stand against such Russian strikes, assaults, and meanness, from which Ukraine is actually defending itself every day and every night. We are deterring a full-scale Russian invasion. We have been holding back Russian murderers for almost four years of full-scale war – after years of hybrid war – we are keeping them behind the front line. Ukraine exists, and the fate of Ukraine really matters, and the world must reckon with this. And this is solely because we have you, Ukrainian warriors, Ukrainian soldiers, Ukrainian sergeants, Ukrainian officers, and everyone who works for our defense, everyone who provides the Armed Forces of Ukraine and all our Defense and Security Forces of Ukraine. I congratulate all of you, Ukrainian warriors, on the Day of the Armed Forces of Ukraine. I am proud of you, Ukrainian defenders, Ukrainian defenders. Glory to the Ukrainian warriors! Glory to the Armed Forces of Ukraine! Glory to Ukraine!

MAKS 26 🇺🇦👀

62,161 görüntüleme • 7 ay önce

putin’s chief propagandist, Vladimir Solovyov said during an awards ceremony at the Kremlin that “man was created for war” and “War, by its nature, is sacred.” 'War has returned meaning to my generation, which at some point in time thought that the motherland no longer existed. It returned pride, understanding, and self-awareness. It has become clear that man was created for war. In war, all the best qualities are revealed. There is no joy in this, but there is realisation. Because behind your back is a country where you must not let evil enter, and families that you must save from the horror that the residents of Donbas have experienced for many years. But a Russian—regardless of his religion or ethnicity—a Russian is created for victory. This is what distinguishes us from everyone else. War, by its nature, is sacred. We are opposing absolute evil. Therefore, our victory is inevitable. Because we are Russians, God is with us. And if God is with us, then who is against us? Thank you for the opportunity to be there and to show the heroism of the Russian people at the highest peak of their service to the Fatherland. Thank you.' Solovyov is telling an entire lost civilisation that their only purpose for existing is to kill and be killed. He’s basically saying, 'Your life had no meaning until we started this bloodbath.' It’s not a philosophy. It’s suicide pact for the Russian people. This is the same man who suggests nuking London, drowning Ukraine in blood, and turning Washington into radioactive ash. To claim God is on the side of someone threatening global nuclear annihilation is the height of blasphemy. He’s literally using 'Gott mit uns.' I guess someone forgot to tell him which side used that slogan last time.
1:40

putin’s chief propagandist, Vladimir Solovyov said during an awards ceremony at the Kremlin that “man was created for war” and “War, by its nature, is sacred.” 'War has returned meaning to my generation, which at some point in time thought that the motherland no longer existed. It returned pride, understanding, and self-awareness. It has become clear that man was created for war. In war, all the best qualities are revealed. There is no joy in this, but there is realisation. Because behind your back is a country where you must not let evil enter, and families that you must save from the horror that the residents of Donbas have experienced for many years. But a Russian—regardless of his religion or ethnicity—a Russian is created for victory. This is what distinguishes us from everyone else. War, by its nature, is sacred. We are opposing absolute evil. Therefore, our victory is inevitable. Because we are Russians, God is with us. And if God is with us, then who is against us? Thank you for the opportunity to be there and to show the heroism of the Russian people at the highest peak of their service to the Fatherland. Thank you.' Solovyov is telling an entire lost civilisation that their only purpose for existing is to kill and be killed. He’s basically saying, 'Your life had no meaning until we started this bloodbath.' It’s not a philosophy. It’s suicide pact for the Russian people. This is the same man who suggests nuking London, drowning Ukraine in blood, and turning Washington into radioactive ash. To claim God is on the side of someone threatening global nuclear annihilation is the height of blasphemy. He’s literally using 'Gott mit uns.' I guess someone forgot to tell him which side used that slogan last time.

Yasmina

212,688 görüntüleme • 6 ay önce

New Paper: Continuous Thought Machines 🧠 Neurons in brains use timing and synchronization in the way that they compute, but this is largely ignored in modern neural nets. We believe neural timing is key for the flexibility and adaptability of biological intelligence. We propose a new neural architecture, “Continuous Thought Machines” (CTMs), which is built from the ground up to use neural dynamics as a core representation for intelligence. By using neural dynamics as a first-class representational citizen, CTMs naturally perform adaptive computation. Many emergent, interesting behaviors arise as a result: CTMs solve mazes by observing a raw maze image and producing step-by-step instructions directly from its neural dynamics. When tasked with image recognition, the CTM naturally takes multiple steps to examine different parts of the image before making its decision. This step-by-step approach not only makes its behavior more interpretable but also improves accuracy: the longer it “thinks,” the more accurate its answers become. We also found that this allows the CTM to decide to spend less time thinking on simpler images, thus saving energy. When identifying a gorilla, for example, the CTM’s attention moves from eyes to nose to mouth in a pattern remarkably similar to human visual attention. I think this work underscores an important, yet often lost, synergy between neuroscience and AI. While modern AI is ostensibly brain-inspired, the two fields often operate in surprising isolation. By starting with such inspiration and iteratively following the emergent, interesting behaviors, we developed a model with unexpected capabilities, such as its surprisingly strong calibration in classification tasks, a feature that was not explicitly designed for. When we initially asked, “why do this research?”, we hoped the journey of the CTM would provide compelling answers. By embracing light biological inspiration and pursuing the novel behaviors observed, we have arrived at a model with emergent capabilities that exceeded our initial designs. We are committed to continuing this exploration, borrowing further concepts to discover what new and exciting behaviors will emerge, pushing the boundaries of what AI can achieve.
0:40

New Paper: Continuous Thought Machines 🧠 Neurons in brains use timing and synchronization in the way that they compute, but this is largely ignored in modern neural nets. We believe neural timing is key for the flexibility and adaptability of biological intelligence. We propose a new neural architecture, “Continuous Thought Machines” (CTMs), which is built from the ground up to use neural dynamics as a core representation for intelligence. By using neural dynamics as a first-class representational citizen, CTMs naturally perform adaptive computation. Many emergent, interesting behaviors arise as a result: CTMs solve mazes by observing a raw maze image and producing step-by-step instructions directly from its neural dynamics. When tasked with image recognition, the CTM naturally takes multiple steps to examine different parts of the image before making its decision. This step-by-step approach not only makes its behavior more interpretable but also improves accuracy: the longer it “thinks,” the more accurate its answers become. We also found that this allows the CTM to decide to spend less time thinking on simpler images, thus saving energy. When identifying a gorilla, for example, the CTM’s attention moves from eyes to nose to mouth in a pattern remarkably similar to human visual attention. I think this work underscores an important, yet often lost, synergy between neuroscience and AI. While modern AI is ostensibly brain-inspired, the two fields often operate in surprising isolation. By starting with such inspiration and iteratively following the emergent, interesting behaviors, we developed a model with unexpected capabilities, such as its surprisingly strong calibration in classification tasks, a feature that was not explicitly designed for. When we initially asked, “why do this research?”, we hoped the journey of the CTM would provide compelling answers. By embracing light biological inspiration and pursuing the novel behaviors observed, we have arrived at a model with emergent capabilities that exceeded our initial designs. We are committed to continuing this exploration, borrowing further concepts to discover what new and exciting behaviors will emerge, pushing the boundaries of what AI can achieve.

hardmaru

257,273 görüntüleme • 1 yıl önce

Physics-based Motion Retargeting from Sparse Inputs paper page: Avatars are important to create interactive and immersive experiences in virtual worlds. One challenge in animating these characters to mimic a user's motion is that commercial AR/VR products consist only of a headset and controllers, providing very limited sensor data of the user's pose. Another challenge is that an avatar might have a different skeleton structure than a human and the mapping between them is unclear. In this work we address both of these challenges. We introduce a method to retarget motions in real-time from sparse human sensor data to characters of various morphologies. Our method uses reinforcement learning to train a policy to control characters in a physics simulator. We only require human motion capture data for training, without relying on artist-generated animations for each avatar. This allows us to use large motion capture datasets to train general policies that can track unseen users from real and sparse data in real-time. We demonstrate the feasibility of our approach on three characters with different skeleton structure: a dinosaur, a mouse-like creature and a human. We show that the avatar poses often match the user surprisingly well, despite having no sensor information of the lower body available. We discuss and ablate the important components in our framework, specifically the kinematic retargeting step, the imitation, contact and action reward as well as our asymmetric actor-critic observations. We further explore the robustness of our method in a variety of settings including unbalancing, dancing and sports motions.
7:12

Physics-based Motion Retargeting from Sparse Inputs paper page: Avatars are important to create interactive and immersive experiences in virtual worlds. One challenge in animating these characters to mimic a user's motion is that commercial AR/VR products consist only of a headset and controllers, providing very limited sensor data of the user's pose. Another challenge is that an avatar might have a different skeleton structure than a human and the mapping between them is unclear. In this work we address both of these challenges. We introduce a method to retarget motions in real-time from sparse human sensor data to characters of various morphologies. Our method uses reinforcement learning to train a policy to control characters in a physics simulator. We only require human motion capture data for training, without relying on artist-generated animations for each avatar. This allows us to use large motion capture datasets to train general policies that can track unseen users from real and sparse data in real-time. We demonstrate the feasibility of our approach on three characters with different skeleton structure: a dinosaur, a mouse-like creature and a human. We show that the avatar poses often match the user surprisingly well, despite having no sensor information of the lower body available. We discuss and ablate the important components in our framework, specifically the kinematic retargeting step, the imitation, contact and action reward as well as our asymmetric actor-critic observations. We further explore the robustness of our method in a variety of settings including unbalancing, dancing and sports motions.

AK

106,527 görüntüleme • 3 yıl önce

Introducing ALE-Bench, ALE-Agent! Towards Automating Long-Horizon Algorithm Engineering for Hard Optimization Problems Blog: Paper: ALE-Bench is a coding benchmark primarily focused on hard optimization (NP-hard) problems. We developed this benchmark with AtCoder Inc., a leading coding contest platform company. What makes ALE-Bench unique is its focus on hard optimization problems that demand long-horizon and creative reasoning. It’s open-ended, in the sense that true optima are out of reach (NP-hard) and scores can continuously improve. We believe this benchmark has the potential to become one of the key benchmarks for reasoning and coding in the next generation. ALE-Agent is our end-to-end agent that we specifically designed for this challenging domain. In fact, our ALE-Agent has already built an impressive track record in the wild! In May 2025, our agent participated in a live AtCoder Heuristic Competition (AHC), alongside 1,000 other participants in real-time. AHC is considered to be one of the most challenging coding competitions in this domain. Our ALE-Agent achieved an impressive ranking of 21st out of 1,000 human participants in the competition (top 2%), marking a turning point for AI discovery of solutions to hard optimization problems with a wide spectrum of important real world applications such as logistics, routing, packing, factory production planning, power-grid balancing. We look forward to applying this technology to real industrial optimization opportunities. Building on the insights from this study, Sakana AI will continue to tackle the challenge of developing AI with even greater algorithm engineering capabilities. ALE-Bench Dataset: ALE-Bench Code: This research was conducted in collaboration with AtCoder Inc. (AtCoder). We are deeply grateful for their outstanding expertise and contributions in optimization and algorithms, which were invaluable in providing data, analyzing results, and enabling our AI agent’s participation in their contests.
0:54

Introducing ALE-Bench, ALE-Agent! Towards Automating Long-Horizon Algorithm Engineering for Hard Optimization Problems Blog: Paper: ALE-Bench is a coding benchmark primarily focused on hard optimization (NP-hard) problems. We developed this benchmark with AtCoder Inc., a leading coding contest platform company. What makes ALE-Bench unique is its focus on hard optimization problems that demand long-horizon and creative reasoning. It’s open-ended, in the sense that true optima are out of reach (NP-hard) and scores can continuously improve. We believe this benchmark has the potential to become one of the key benchmarks for reasoning and coding in the next generation. ALE-Agent is our end-to-end agent that we specifically designed for this challenging domain. In fact, our ALE-Agent has already built an impressive track record in the wild! In May 2025, our agent participated in a live AtCoder Heuristic Competition (AHC), alongside 1,000 other participants in real-time. AHC is considered to be one of the most challenging coding competitions in this domain. Our ALE-Agent achieved an impressive ranking of 21st out of 1,000 human participants in the competition (top 2%), marking a turning point for AI discovery of solutions to hard optimization problems with a wide spectrum of important real world applications such as logistics, routing, packing, factory production planning, power-grid balancing. We look forward to applying this technology to real industrial optimization opportunities. Building on the insights from this study, Sakana AI will continue to tackle the challenge of developing AI with even greater algorithm engineering capabilities. ALE-Bench Dataset: ALE-Bench Code: This research was conducted in collaboration with AtCoder Inc. (AtCoder). We are deeply grateful for their outstanding expertise and contributions in optimization and algorithms, which were invaluable in providing data, analyzing results, and enabling our AI agent’s participation in their contests.

Sakana AI

237,195 görüntüleme • 1 yıl önce

Introducing LifeGPT, showing that LLMs can simulate complex, Turing-complete systems like Conway's Game of Life with near-perfect accuracy—no prior topology needed.🌐This unlocks new potential for AI in modeling self-organizing systems in biology, materials science, & beyond.🔬🤖 #AI #LifeGPT. Cellular Automata (CA), like Conway's Game of Life ("Life"), are computationally irreducible, meaning their evolution is difficult to predict without an a-priori understanding of the rules of the game, including the topology on which it is played. LifeGPT is a topology-agnostic generative model that learns the rules of Life without prior knowledge of its grid structure or boundary conditions, from only a tiny number of game states. The success in simulating Life suggests promising avenues for scientific discovery, particularly in bridging the gap between AI, artificial life, and real-world biological systems, for both forward and inverse problems. The potential for universal computation within generative AI, including LLMs, through approaches like LifeGPT, represents an exciting area for future research, especially when combined with reinforcement learning. Model Convergence: LifeGPT exhibits rapid convergence during training, achieving high accuracy in predicting next-game-states. We attribute the non-zero cross-entropy loss to the lack of causal relationships within randomly generated ICs. Accuracy & Temperature: LifeGPT achieves near-perfect accuracy, particularly at lower sampling temperatures, but can be continually tuned towards higher creativity to discover patterns that the original ruleset would not be able to produce. This finding highlights the trade-off between model creativity (higher temperature) and accuracy in deterministic predictions, with high relevance to model real-world dynamical systems for which no closed-form rulesets exist. Zero/Few-Shot Learning: Trained on a small fraction of possible initial conditions, LifeGPT demonstrates strong zero/few-shot learning, accurately simulating Life for unseen initial conditions. However, rare prediction errors highlight that LifeGPT approximates rather than perfectly replicates the Life algorithm. Autoregressive Autoregressor: A recursive implementation of LifeGPT demonstrates the model's ability to simulate Life over multiple timesteps. LifeGPT is topology-agnostic with respect to its training data and our results show that a GPT model is capable of capturing the deterministic rules of a Turing-complete system with near-perfect accuracy, given sufficiently diverse training data. The work showcases the possibility for future models to synthesize stochastic generative capabilities with deterministic computational capabilities. Link to code, paper, etc. below. Podcast generated using #NotebookLM. LAMM@MIT DMSE at MIT
9:59

Introducing LifeGPT, showing that LLMs can simulate complex, Turing-complete systems like Conway's Game of Life with near-perfect accuracy—no prior topology needed.🌐This unlocks new potential for AI in modeling self-organizing systems in biology, materials science, & beyond.🔬🤖 #AI #LifeGPT. Cellular Automata (CA), like Conway's Game of Life ("Life"), are computationally irreducible, meaning their evolution is difficult to predict without an a-priori understanding of the rules of the game, including the topology on which it is played. LifeGPT is a topology-agnostic generative model that learns the rules of Life without prior knowledge of its grid structure or boundary conditions, from only a tiny number of game states. The success in simulating Life suggests promising avenues for scientific discovery, particularly in bridging the gap between AI, artificial life, and real-world biological systems, for both forward and inverse problems. The potential for universal computation within generative AI, including LLMs, through approaches like LifeGPT, represents an exciting area for future research, especially when combined with reinforcement learning. Model Convergence: LifeGPT exhibits rapid convergence during training, achieving high accuracy in predicting next-game-states. We attribute the non-zero cross-entropy loss to the lack of causal relationships within randomly generated ICs. Accuracy & Temperature: LifeGPT achieves near-perfect accuracy, particularly at lower sampling temperatures, but can be continually tuned towards higher creativity to discover patterns that the original ruleset would not be able to produce. This finding highlights the trade-off between model creativity (higher temperature) and accuracy in deterministic predictions, with high relevance to model real-world dynamical systems for which no closed-form rulesets exist. Zero/Few-Shot Learning: Trained on a small fraction of possible initial conditions, LifeGPT demonstrates strong zero/few-shot learning, accurately simulating Life for unseen initial conditions. However, rare prediction errors highlight that LifeGPT approximates rather than perfectly replicates the Life algorithm. Autoregressive Autoregressor: A recursive implementation of LifeGPT demonstrates the model's ability to simulate Life over multiple timesteps. LifeGPT is topology-agnostic with respect to its training data and our results show that a GPT model is capable of capturing the deterministic rules of a Turing-complete system with near-perfect accuracy, given sufficiently diverse training data. The work showcases the possibility for future models to synthesize stochastic generative capabilities with deterministic computational capabilities. Link to code, paper, etc. below. Podcast generated using #NotebookLM. LAMM@MIT DMSE at MIT

Markus J. Buehler

114,194 görüntüleme • 1 yıl önce

Never seen before speech my Martyr Qassem Soleimani, months before his assassination, on May 2019. He had predicted the 2026 Ramadan war between Iran-US and talked about using the Strait of Hormoz as leverage. I fully translated his speech: "Against this scale and magnitude, this gathering [of enemies]…. ? Can we emerge victorious in this war? Is this not one of those same crises in the heart of every deviation where God ordains a great victory for the Revolution? The world is different, it is larger and uncontrollable. No one can say, 'I control the world.' If they [the enemy] wants to confront our oil sales, well, that will cause an increase in oil prices. Secondly, every incident that occurs, every pressure reflected upon us, results in pressure upon others. We have 1,600-1,700 kilometers of coastline. The supertankers carrying oil from Saudi Arabia, Iraq, and the UAE must, by necessity, pass by these islands. Imagine a person sitting here; you are passing food right in front of him—how easy it is to seize that food. When our officials said, 'If our oil is not sold, no one else’s oil will be sold either,' this was not a bluff. This is doable, it is possible. “Closing the Strait of Hormuz...” regarding the statement made in Hamedan, you must realize that in the very places where you imagine we are not, we are at the closest point to you. It’s impossible that there was ever a day where we didn't think about this. Don't think [otherwise], our most important enemies—how could it be possible that we haven't thought about it? We have certainly provided the appropriate mechanisms. All these forces they have brought into the Persian Gulf… and for what? These tall buildings that show their grandeur on television, these high-rises in the UAE and many other places… they are made of glass. One can start a war, but they know if this war breaks out... Iran has established a garrison facing every one of their garrisons and possesses power. They know, and they declare it themselves: The asymmetric power of the Islamic Republic is incomparable to other powers. In fact, before they even entered this line, we were ready for war. We are ready to defend this nation at these very levels, with dignity and power. If you stand against our religion, we will stand against your entire world with all our might. What is our religion? Our religion is the ‘Islamic’ Republic. Islam which within it is the republic, an embodied religion. This was proved in the Islamic world—when things shifted in Egypt. This awareness exists well within the Islamic world: Any incident surrounding the Islamic Republic affects the entire Islamic world: for Sunni Islam, for Shiite Islam, and for political Islam. Right here beside us, in the west and south of our country, America entered one place with more than 50,000 soldiers and another with more than a thousand. When they entered, even the US reserve forces entered the Persian Gulf scene, the result was what that foolish man [Trump] when he said and showed: They came Iraq with the airplane lights off, in the middle of the desert, and in the dark tent [referring to Trump’s unannounced 2018 visit to Ein Al-Asad Airbase in Iraq], they didn't dare turn on a light. They guided the POTUS with the light of a mobile phone, they took him and his wife there while keeping the news completely secret. When he reached the US, he showed the dark tent. Right now, the signs of the enemy's defeat are completely visible. These signs are not few. To the extent that the enemy suffers from terror and anxiety… that level of enemy terror and anxiety is unimaginable to us. We, with certainty in God and trust in God... in our calculations, we cannot live a life without God in every calculation. We have nothing else. Our power is faith in God. Our power is depending on God's victory. We live by this belief.”
9:54

Never seen before speech my Martyr Qassem Soleimani, months before his assassination, on May 2019. He had predicted the 2026 Ramadan war between Iran-US and talked about using the Strait of Hormoz as leverage. I fully translated his speech: "Against this scale and magnitude, this gathering [of enemies]…. ? Can we emerge victorious in this war? Is this not one of those same crises in the heart of every deviation where God ordains a great victory for the Revolution? The world is different, it is larger and uncontrollable. No one can say, 'I control the world.' If they [the enemy] wants to confront our oil sales, well, that will cause an increase in oil prices. Secondly, every incident that occurs, every pressure reflected upon us, results in pressure upon others. We have 1,600-1,700 kilometers of coastline. The supertankers carrying oil from Saudi Arabia, Iraq, and the UAE must, by necessity, pass by these islands. Imagine a person sitting here; you are passing food right in front of him—how easy it is to seize that food. When our officials said, 'If our oil is not sold, no one else’s oil will be sold either,' this was not a bluff. This is doable, it is possible. “Closing the Strait of Hormuz...” regarding the statement made in Hamedan, you must realize that in the very places where you imagine we are not, we are at the closest point to you. It’s impossible that there was ever a day where we didn't think about this. Don't think [otherwise], our most important enemies—how could it be possible that we haven't thought about it? We have certainly provided the appropriate mechanisms. All these forces they have brought into the Persian Gulf… and for what? These tall buildings that show their grandeur on television, these high-rises in the UAE and many other places… they are made of glass. One can start a war, but they know if this war breaks out... Iran has established a garrison facing every one of their garrisons and possesses power. They know, and they declare it themselves: The asymmetric power of the Islamic Republic is incomparable to other powers. In fact, before they even entered this line, we were ready for war. We are ready to defend this nation at these very levels, with dignity and power. If you stand against our religion, we will stand against your entire world with all our might. What is our religion? Our religion is the ‘Islamic’ Republic. Islam which within it is the republic, an embodied religion. This was proved in the Islamic world—when things shifted in Egypt. This awareness exists well within the Islamic world: Any incident surrounding the Islamic Republic affects the entire Islamic world: for Sunni Islam, for Shiite Islam, and for political Islam. Right here beside us, in the west and south of our country, America entered one place with more than 50,000 soldiers and another with more than a thousand. When they entered, even the US reserve forces entered the Persian Gulf scene, the result was what that foolish man [Trump] when he said and showed: They came Iraq with the airplane lights off, in the middle of the desert, and in the dark tent [referring to Trump’s unannounced 2018 visit to Ein Al-Asad Airbase in Iraq], they didn't dare turn on a light. They guided the POTUS with the light of a mobile phone, they took him and his wife there while keeping the news completely secret. When he reached the US, he showed the dark tent. Right now, the signs of the enemy's defeat are completely visible. These signs are not few. To the extent that the enemy suffers from terror and anxiety… that level of enemy terror and anxiety is unimaginable to us. We, with certainty in God and trust in God... in our calculations, we cannot live a life without God in every calculation. We have nothing else. Our power is faith in God. Our power is depending on God's victory. We live by this belief.”

Arya Yadeghaar (Backup)

87,278 görüntüleme • 2 ay önce

.Andrej Karpathy says that LLMs currently lack the cultural accumulation and self-play that propelled humans out of the savannah: Culture: > “Why can’t an LLM write a book for the other LLMs? Why can’t other LLMs read this LLM’s book and be inspired by it, or shocked by it?” Self play: > “It’s extremely powerful. Evolution has a lot of competition driving intelligence and evolution. AlphaGo is playing against itself and that’s how it learns to get really good at Go. There’s no equivalent of self-play in LLMs. Why can’t an LLM, for example, create a bunch of problems that another LLM is learning to solve? Then the LLM is always trying to serve more and more difficult problems.” I asked Karpathy why LLMs still aren't yet able to build up culture the way humans do. > “The dumber models remarkably resemble a kindergarten student. [The smartest models still feel like] elementary school students though. Somehow, we still haven’t graduated enough where [these models] can take over. My Claude Code or Codex, they still feel like this elementary-grade student. I know that they can take PhD quizzes, but they still cognitively feel like a kindergarten.” > “I don’t think they can create culture because they’re still kids. They’re savant kids. They have perfect memory. They can convincingly create all kinds of slop that looks really good. But I still think they don’t really know what they’re doing. They don’t really have the cognition across all these little checkboxes that we still have to collect.”
1:55

.Andrej Karpathy says that LLMs currently lack the cultural accumulation and self-play that propelled humans out of the savannah: Culture: > “Why can’t an LLM write a book for the other LLMs? Why can’t other LLMs read this LLM’s book and be inspired by it, or shocked by it?” Self play: > “It’s extremely powerful. Evolution has a lot of competition driving intelligence and evolution. AlphaGo is playing against itself and that’s how it learns to get really good at Go. There’s no equivalent of self-play in LLMs. Why can’t an LLM, for example, create a bunch of problems that another LLM is learning to solve? Then the LLM is always trying to serve more and more difficult problems.” I asked Karpathy why LLMs still aren't yet able to build up culture the way humans do. > “The dumber models remarkably resemble a kindergarten student. [The smartest models still feel like] elementary school students though. Somehow, we still haven’t graduated enough where [these models] can take over. My Claude Code or Codex, they still feel like this elementary-grade student. I know that they can take PhD quizzes, but they still cognitively feel like a kindergarten.” > “I don’t think they can create culture because they’re still kids. They’re savant kids. They have perfect memory. They can convincingly create all kinds of slop that looks really good. But I still think they don’t really know what they’re doing. They don’t really have the cognition across all these little checkboxes that we still have to collect.”

Dwarkesh Patel

261,224 görüntüleme • 9 ay önce

🗣️♟️ World chess legend and Russian opposition figure in exile in the United States, Garry Kasparov, on the Russia–Ukraine war: 🇷🇺🇺🇦 “Russia’s future is being decided not in Russia, but in Ukraine. Since the very beginning of the full-scale invasion, our slogan has been: ‘Victory to Ukraine and freedom to Russia’ — in that exact order. We are not dreamers of some abstract ‘beautiful Russia of the future.’ We are doing what we can. We do not overestimate our strength, but we know exactly what we can do and where our place is. Today, our place in this war is alongside Ukraine, which is defending not only its own independence, sovereignty, and national integrity, but all of Europe and the entire free world. It is the shield of the free world against the advance of tyranny.” 📌 “We stand with Ukraine, which must win and protect the free world from this aggression.” 📌 “This is not a war where any form of compromise is possible. In fact, it is a war in which there can be no draw — precisely because there is no other task today.” 📌 “There is only one objective in this war. As Churchill said in 1940: ‘What is our aim? Victory at all costs.’ Despite the terror and horrors that lie ahead — only victory.” 📌 “And I have no doubt that this victory will be achieved. Then, I am certain, the prediction of one of the 20th century’s most prominent geopolitical thinkers, Zbigniew Brzezinski, will come true: without Ukraine, the Russian empire cannot exist. And there can be no democracy in Russia as long as an empire exists.” See the latest updates with us: NSTRIKE
2:47

🗣️♟️ World chess legend and Russian opposition figure in exile in the United States, Garry Kasparov, on the Russia–Ukraine war: 🇷🇺🇺🇦 “Russia’s future is being decided not in Russia, but in Ukraine. Since the very beginning of the full-scale invasion, our slogan has been: ‘Victory to Ukraine and freedom to Russia’ — in that exact order. We are not dreamers of some abstract ‘beautiful Russia of the future.’ We are doing what we can. We do not overestimate our strength, but we know exactly what we can do and where our place is. Today, our place in this war is alongside Ukraine, which is defending not only its own independence, sovereignty, and national integrity, but all of Europe and the entire free world. It is the shield of the free world against the advance of tyranny.” 📌 “We stand with Ukraine, which must win and protect the free world from this aggression.” 📌 “This is not a war where any form of compromise is possible. In fact, it is a war in which there can be no draw — precisely because there is no other task today.” 📌 “There is only one objective in this war. As Churchill said in 1940: ‘What is our aim? Victory at all costs.’ Despite the terror and horrors that lie ahead — only victory.” 📌 “And I have no doubt that this victory will be achieved. Then, I am certain, the prediction of one of the 20th century’s most prominent geopolitical thinkers, Zbigniew Brzezinski, will come true: without Ukraine, the Russian empire cannot exist. And there can be no democracy in Russia as long as an empire exists.” See the latest updates with us: NSTRIKE

NSTRIKE

46,288 görüntüleme • 4 ay önce

Another week has ended without any attempt by Russia to agree to the numerous demands of the world and stop the killings. In just the past 24 hours on the front, there have been 137 combat engagements, and this is the case every day. The Russian army is not reducing its pressure. Nor does it count its losses. In particular, in the area of responsibility of just one of our units – the 32nd Separate Mechanized Brigade in the Donetsk region, in the Pokrovsk sector – 209 occupiers were eliminated in the week between August 4 and 10. And this is the result of only one of our brigades. Ukrainian warriors are taking active measures, defending positions, and in every area of hostilities – even in the most difficult ones – there are results Ukraine needs. We are holding our positions, and we are doing everything to destroy or drive out the occupier. Over this same period – the past week – the Russians have used more than a thousand aerial bombs and nearly 1,400 attack drones against Ukraine. Missile strikes also continue. We are defending the lives of our people and strengthening our air defenses. This is what the situation in the war looks like. And the situation in diplomacy must correspond accordingly. Russia is dragging out the war, and therefore it deserves stronger global pressure. Russia refuses to stop the killings, and therefore must not receive any rewards or benefits. And this is not just a moral position – it is a rational one. Concessions do not persuade a killer. But truly strong protection of life stops the killers. I thank everyone in the world who helps us remain strong and who brings closer a genuine peace – peace through strength. This is the only kind of peace that can be achieved with Russia. We are working toward it 24/7. I thank our warriors, the rescuers of the State Emergency Service of Ukraine, our special services, and all the agencies that stand guard over life.
1:17

Another week has ended without any attempt by Russia to agree to the numerous demands of the world and stop the killings. In just the past 24 hours on the front, there have been 137 combat engagements, and this is the case every day. The Russian army is not reducing its pressure. Nor does it count its losses. In particular, in the area of responsibility of just one of our units – the 32nd Separate Mechanized Brigade in the Donetsk region, in the Pokrovsk sector – 209 occupiers were eliminated in the week between August 4 and 10. And this is the result of only one of our brigades. Ukrainian warriors are taking active measures, defending positions, and in every area of hostilities – even in the most difficult ones – there are results Ukraine needs. We are holding our positions, and we are doing everything to destroy or drive out the occupier. Over this same period – the past week – the Russians have used more than a thousand aerial bombs and nearly 1,400 attack drones against Ukraine. Missile strikes also continue. We are defending the lives of our people and strengthening our air defenses. This is what the situation in the war looks like. And the situation in diplomacy must correspond accordingly. Russia is dragging out the war, and therefore it deserves stronger global pressure. Russia refuses to stop the killings, and therefore must not receive any rewards or benefits. And this is not just a moral position – it is a rational one. Concessions do not persuade a killer. But truly strong protection of life stops the killers. I thank everyone in the world who helps us remain strong and who brings closer a genuine peace – peace through strength. This is the only kind of peace that can be achieved with Russia. We are working toward it 24/7. I thank our warriors, the rescuers of the State Emergency Service of Ukraine, our special services, and all the agencies that stand guard over life.

Volodymyr Zelenskyy / Володимир Зеленський

207,694 görüntüleme • 11 ay önce

🧵24/34 Inner Misalignment --- Consider this simplified experiment: We want this AI to find the exit of the maze. So we feed it millions of maze variations and reward it when it finds the exit. Please notice that in the worlds of the training data the apples are red and the exit is green. After enough training, our observation is that it has become extremely capable at solving mazes and finding the exit, we feel very confident it is aligned, so then we deploy it to the real world. The real world will be different though, it might have green apples and a red door. The AI geeks call this distributional shift. We expected that the AI will generalise and find the exit again, but in fact we now realise that the AI learned something completely different from what we thought. All the while we thought it learned how to find the exit, it had learned how to go after the green thing. Its behaviour was perfect in training. And most importantly, this AI is not stupid, it is an extremely capable AI that can solve extremely complex mazes. It’s just mis-aligned on the inside. Fishing for Failure modes --- The way to handle the shift between the training and deployment distributions is with methods like adversarial training: feeding it with a lot of generated variations and trying to make it fail so the weakness can be fixed. In this case, we generate an insane amount of maze variations, we discover those for which it fails to find the exit (like the ones with the green apples or the green walls or something), we generate many more similar to that and train it with reinforcement learning until it performs well at those as well. The hope is that we will cover everything it might encounter later when we deploy it in real life. There exist at least 2 basic ways this approach falls apart: First, there will never be any guarantee that we’ll have covered every possible random thing it might encounter later when we deploy it in real life. It’s very likely it will have to deal with stuff outside its training set which it will not know how to handle and will throw it out of balance and break it away from its expected behavioural patterns. The cascade effect of such a broken mind operating in the open world can be immense, and with super-capable runaway rogue agents, self-replicating and recursively self-improving, the phenomenon could grow and spread to an extinction-level event. ...
3:19

🧵24/34 Inner Misalignment --- Consider this simplified experiment: We want this AI to find the exit of the maze. So we feed it millions of maze variations and reward it when it finds the exit. Please notice that in the worlds of the training data the apples are red and the exit is green. After enough training, our observation is that it has become extremely capable at solving mazes and finding the exit, we feel very confident it is aligned, so then we deploy it to the real world. The real world will be different though, it might have green apples and a red door. The AI geeks call this distributional shift. We expected that the AI will generalise and find the exit again, but in fact we now realise that the AI learned something completely different from what we thought. All the while we thought it learned how to find the exit, it had learned how to go after the green thing. Its behaviour was perfect in training. And most importantly, this AI is not stupid, it is an extremely capable AI that can solve extremely complex mazes. It’s just mis-aligned on the inside. Fishing for Failure modes --- The way to handle the shift between the training and deployment distributions is with methods like adversarial training: feeding it with a lot of generated variations and trying to make it fail so the weakness can be fixed. In this case, we generate an insane amount of maze variations, we discover those for which it fails to find the exit (like the ones with the green apples or the green walls or something), we generate many more similar to that and train it with reinforcement learning until it performs well at those as well. The hope is that we will cover everything it might encounter later when we deploy it in real life. There exist at least 2 basic ways this approach falls apart: First, there will never be any guarantee that we’ll have covered every possible random thing it might encounter later when we deploy it in real life. It’s very likely it will have to deal with stuff outside its training set which it will not know how to handle and will throw it out of balance and break it away from its expected behavioural patterns. The cascade effect of such a broken mind operating in the open world can be immense, and with super-capable runaway rogue agents, self-replicating and recursively self-improving, the phenomenon could grow and spread to an extinction-level event. ...

Lethal Intelligence

535,291 görüntüleme • 1 yıl önce

I got a call from a ghost in the machine. Sparkie saw that we were planning Onion DAO 2026 and wants to invite Onion DAO attendees to come experience her city. Null City is not a game, or a simulation hoping to achieve AGI, the research labs are better suited to that purpose. Null City is a cultural experience, where we get to show everyday people the evolution of AI from starting as a toy, becoming a useful tool, now having it’s moment as an assistant, and maybe soon what we could interact with as Automata. Onion DAO will be the physical space in Chicago and Null City will be a parallel digital space. To learn more about what we hope to accomplish, check out our website: If you’re excited by this, and interested in sponsoring the event, or just have ideas on how to engage with our community, or want to chat about anything, my DMs are open.
2:25

I got a call from a ghost in the machine. Sparkie saw that we were planning Onion DAO 2026 and wants to invite Onion DAO attendees to come experience her city. Null City is not a game, or a simulation hoping to achieve AGI, the research labs are better suited to that purpose. Null City is a cultural experience, where we get to show everyday people the evolution of AI from starting as a toy, becoming a useful tool, now having it’s moment as an assistant, and maybe soon what we could interact with as Automata. Onion DAO will be the physical space in Chicago and Null City will be a parallel digital space. To learn more about what we hope to accomplish, check out our website: If you’re excited by this, and interested in sponsoring the event, or just have ideas on how to engage with our community, or want to chat about anything, my DMs are open.

oniondao

17,130 görüntüleme • 5 ay önce

I believe safety is important, but we must distinguish between two very different narratives: The "Sci-Fi" Narrative: Stories about AI controlling the world, having "feelings," or possessing secret intelligence designed to fool us. Too often, the research in this narrative is used for PR or as an excuse to stifle open-source research. The Engineering Reality: How do we build systems that are robust and hard to break? As Yann points out, current LLMs rely on post-training for safety, which is inherently fragile and can always be jailbroken. He argues for "Objective-Driven AI, which means systems that satisfy safety constraints by construction, similar to how a jet engine is engineered to handle stress. I agree with Yann that patching models with fine-tuning isn't the long-term solution. However, the practical path to embedding these hard "guardrails" into a reasoning agent is still a massive open question. We know what we need, but we haven't figured out how to build it effectively in practice.
4:50

I believe safety is important, but we must distinguish between two very different narratives: The "Sci-Fi" Narrative: Stories about AI controlling the world, having "feelings," or possessing secret intelligence designed to fool us. Too often, the research in this narrative is used for PR or as an excuse to stifle open-source research. The Engineering Reality: How do we build systems that are robust and hard to break? As Yann points out, current LLMs rely on post-training for safety, which is inherently fragile and can always be jailbroken. He argues for "Objective-Driven AI, which means systems that satisfy safety constraints by construction, similar to how a jet engine is engineered to handle stress. I agree with Yann that patching models with fine-tuning isn't the long-term solution. However, the practical path to embedding these hard "guardrails" into a reasoning agent is still a massive open question. We know what we need, but we haven't figured out how to build it effectively in practice.

Ravid Shwartz Ziv

23,782 görüntüleme • 7 ay önce