Machine Learning Street Talk (MLST)

Joscha Bach is a cognitive scientist and the founder of the California Institute for Machine Consciousness.

SPONSOR:

---

Parallel is a single platform for all your agentic research: web search built for AI agents, with Pareto-optimal quality, cost and speed.

Get started: https://parallel.ai/mlst

---

Bach's null hypothesis is that consciousness is easy: an algorithm not much more complicated than the transformer, simple enough for an infant's brain to discover by itself. In this conversation with Tim, he lays out the Machine Consciousness Hypothesis in two parts. The first is that consciousness is a biological learning algorithm, the training algorithm that makes a brain like ours coherent. The second is that machines learning in self-organising ways might discover it in the same way. Either part can fail on its own.

From there they ask why there will never be a Turing test for consciousness, why a chess program is already a zombie, and why faking a mind can sometimes be harder than building one. The conversation runs through Leibniz's mill, the explanatory gap from John Searle to Thomas Nagel, Anil Seth's biological naturalism, integrated information theory and the letter that called it pseudoscience, Gödel and the limits of formal systems, and Mike Levin's idea that morphogenesis and the mind may run the same coherence algorithm. A tea-room interlude takes in Picbreeder and fractured, entangled representations. The episode closes on what it would take to convince us that a machine is conscious.

---

TIMESTAMPS:

0:00 Cold open: Could consciousness be a learning algorithm?

1:51 Introduction

3:37 What makes an agent?

6:16 Why consciousness needs explaining

14:57 Sponsor: Parallel

15:51 Leibniz's mill and causal patterns

24:21 Perceiving that you perceive

32:04 Two hypotheses, two ways to fail

40:42 Do minds converge?

49:23 Does consciousness converge too?

56:20 When language models reshape our thinking

59:16 Can simulated water make you wet?

1:09:21 Computation as construction

1:16:03 Error-correcting patterns and the Ruliad

1:22:28 Conscious, or simulating consciousness?

1:28:23 What would count as evidence?

1:35:49 The research programme

REFERENCES:

Joscha Bach, The Machine Consciousness Hypothesis (California Institute for Machine Consciousness)

Kumar, Clune, Lehman & Stanley, "Questioning Representational Optimism in Deep Learning: The Fractured Entangled Representation Hypothesis"

Giulio Tononi, integrated information theory

Fleming et al. 2023, "The Integrated Information Theory of Consciousness as Pseudoscience"

John Searle 1980, "Minds, brains, and programs", Behavioral and Brain Sciences

Kurt Gödel 1931, "On Formally Undecidable Propositions of Principia Mathematica and Related Systems"

Alan Turing 1950, "Computing Machinery and Intelligence"

Whitehead & Russell, Principia Mathematica

Paul Thagard 2022, "Energy requirements undermine substrate independence and mind-body functionalism"

Richard Sutton 2019, "The Bitter Lesson"

Michael Graziano, attention schema theory

Bernard Baars and Stanislas Dehaene, global workspace theory

Michael Levin 2022, "Technological Approach to Mind Everywhere", Frontiers in Systems Neuroscience

Chris Olah et al. 2020, "Zoom In: An Introduction to Circuits", Distill

Adele Lopez 2025, "The Rise of Parasitic AI", LessWrong

RESCRIPT:https://app.rescript.info/share/fc7088df0dc8b84200aa6e01c9a9ab6a

Kevin Hartnett is the author of *The Proof in the Code*.SPONSOR:---Cyber Fund built the Monastery to help founders ship products that were impossible a year ago. Applications for Batch 1 are now open.Apply now: https://cyber.fund---In 1998 Thomas Hales finished a proof of the Kepler conjecture that leaned so heavily on computer code that the Annals of Mathematics could not say whether it was true. That impasse opens Kevin Hartnett's book *The Proof in the Code*, and this documentary.Built around Tim's conversation with Hartnett, the film follows Lean from a side project at Microsoft Research to the system mathematicians now use to check their own field. Leonardo de Moura, Lean's creator, explains why everything rests on a small trusted kernel. Jeremy Avigad recalls the argument over foundations that changed Lean. Kevin Buzzard describes his fear that his corner of mathematics was becoming a house of cards, and Paul Lessard separates a correct proof from a useful one. Mathlib, Peter Scholze's Liquid Tensor challenge and Terence Tao's Equational Theories project all pass through.The second half turns to AI. AlphaProof, natural-language olympiad proofs and autoformalisation sharpen the question the book keeps asking: when a machine produces a proof nobody can read, what has actually been checked? The answer lies in the formalisation gap, the distance between the mathematics people mean and the statement a checker certifies. The film closes on the human story behind the software.---TIMESTAMPS:0:00 Cold open: proof, code and the wrong problem1:50 Kepler's oranges and the proof nobody could check6:54 How mathematicians decide a proof is true9:04 Correct is not the same as useful12:17 Sponsor: Cyber Fund12:53 What Lean is, and the small kernel it trusts15:23 From Microsoft's Z3 to Lean18:22 The argument over Lean's foundations24:52 One researcher versus a growing user base26:16 Kevin Buzzard's house of cards30:43 Mathlib as a barn raising37:16 The perfectoid-spaces stunt47:12 Scholze's proof nobody had read: Liquid Tensor53:20 Can AI learn to find proofs? AlphaProof58:43 AI mathematics in 2026 and Tao's experiment1:01:36 Lean's place in AI mathematics1:02:17 AI brute force versus the right abstraction1:05:19 Autoformalisation's Trojan horse1:07:25 The formalisation gap1:11:45 A proof is not understanding1:16:58 Integration hell and the zlib specification1:21:14 Verification in the AI software lifecycle1:22:38 The human story behind Lean1:31:09 The birthday walk1:33:05 Farewell---REFERENCES:[2:12] Thomas Hales, A proof of the Kepler conjecture (Annals of Mathematics, 2005)https://annals.math.princeton.edu/2005/162-3/p01[2:12] The Flyspeck project (the formal proof of the Kepler conjecture)https://github.com/flyspeck/flyspeck[12:53] Leanhttps://lean-lang.org/[30:43] Mathlibhttps://github.com/leanprover-community/mathlib4[30:43] Kevin Hartnett, "Building the Mathematical Library of the Future" (Quanta Magazine, 2020)https://www.quantamagazine.org/building-the-mathematical-library-of-the-future-20201001/[47:12] The Liquid Tensor Experimenthttps://github.com/leanprover-community/lean-liquid[53:20] AlphaProof (Google DeepMind): AI achieves silver-medal standard solving International Mathematical Olympiad problemshttps://deepmind.google/blog/ai-solves-imo-problems-at-silver-medal-level/[59:23] The Equational Theories Project (Terence Tao)https://github.com/teorth/equational_theories[1:18:30] lean-zip (the zlib proof)https://github.com/kim-em/lean-zipFrontierMath Erdős (Epoch AI, 2026)https://epoch.ai/benchmarks/frontiermath-erdosBook: Kevin Hartnett, *The Proof in the Code* (Quanta Books, 2026)https://www.quantabooks.org/books/the-proof-in-the-code/---RESCRIPT: https://app.rescript.info/share/8604a0cd6ea4604a368028a05f4e6be4Support MLST on Patreon: https://www.patreon.com/mlstDiscord: https://discord.gg/HNnAwSduud

A lot of people in AI treat evolution as a dumb fallback, basically random search for when you can't take a gradient. Akarsh Kumar thinks that is wrong. Selection hangs on to partial solutions, so mutations only need to be useful about 1% of the time for the search to keep making progress.Akarsh is a PhD student at MIT working with Phillip Isola, works with Sakana AI, and is first author of the Fractured Entangled Representation paper with Kenneth Stanley, Jeff Clune and Joel Lehman. He tells Tim Scarfe why the path a learner takes may shape the structure of what it learns, and why that is a different way of looking at intelligence from the statistical one.Most of the conversation is about ASAL, the method he led for searching whole spaces of artificial worlds. Rather than predict what a rule will do, ASAL runs the simulation and asks a foundation model what happened. Mapped across all 262,144 Life-like rules, the most interesting worlds sit on one small island. Along the way: Game of Life, Lenia (with a clip from its creator, Bert Chan), neural cellular automata, Boids, Particle Life and the emergence of persistence.---TIMESTAMPS:00:00:00 Intro: artificial life, ASAL and Core War in four minutes00:04:19 Life as it could be, not just as it is00:05:46 Why the order you learn things in matters00:08:33 No shortcuts: Wolfram, novelty search and regularisation00:10:44 Kenneth Stanley: the path matters, not just the destination00:11:31 Statistical intelligence vs regularity-based intelligence00:13:49 Artificial chemistry and other possible universes00:15:59 Convergent patterns: shadows of the substrate?00:18:02 How Conway's Game of Life works00:20:49 Change one cell, change everything?00:22:31 Lenia (with Bert Chan) and neural cellular automata00:25:26 Boids, Particle Life and cell-like creatures00:28:47 Persistence, entropy and what life is00:30:15 ASAL: a foundation model as the critic00:34:04 Impostors inside the simulation00:36:50 From primordial soup to alien animals00:38:26 262,144 rules and the island of open-endedness00:40:17 From artificial life to AGI00:41:35 Core War: Game of Thrones, Turing edition00:43:22 LLMs as the mutation step: evolving warriors00:46:03 Evolution is anything but random---REFERENCES:paper:[00:04:19] ASAL (Kumar et al.)https://arxiv.org/abs/2412.17799[00:08:08] Assembly theory https://www.nature.com/articles/s41586-023-06600-9[00:09:54] FEP paperhttps://arxiv.org/abs/2505.11581[00:22:38] Lenia: Biology of Artificial Life (Bert Chan)https://arxiv.org/abs/1812.05433[00:23:20] Growing Neural Cellular Automata (Mordvintsev et al.)https://distill.pub/2020/growing-ca/[00:41:35] Digital Red Queen: Core War with LLMs (Kumar et al.)https://arxiv.org/abs/2601.03335[00:43:42] MAP-Elites https://arxiv.org/abs/1504.04909[00:44:55] AlphaEvolve https://arxiv.org/abs/2506.13131[00:47:03] AutoML-Zero (Real et al.)https://arxiv.org/abs/2003.03384book:[00:07:04] Why Greatness Cannot Be Planned (Stanley and Lehman)https://link.springer.com/book/10.1007/978-3-319-15524-1tool:[00:18:12] Conway's Game of Lifehttps://en.wikipedia.org/wiki/Conway%27s_Game_of_Life[00:25:27] Boids (Craig Reynolds)https://www.red3d.com/cwr/boids/[00:27:12] Particle Life (Tom Mohr)https://github.com/tom-mohr/particle-life[00:41:50] Core Warhttps://corewar.co.uk/other:[00:08:49] Computational irreducibility (Stephen Wolfram)https://www.wolframscience.com/nks/p737--computational-irreducibility/[00:10:46] MLST: Why Every AI Model Is an Impostor (Kenneth Stanley, FER documentary)https://www.youtube.com/watch?v=o1q6Hhz0MAg[00:28:30] MLST: Blaise Agüera y Arcas on life emerging from codehttps://www.youtube.com/watch?v=rMSEqJ_4EBk---LINKS:Akarsh Kumar: https://akarshkumar.com/ASAL project page and demos: https://pub.sakana.ai/asal/Digital Red Queen project page: https://sakana.ai/drq/RESCRIPT:https://app.rescript.info/public/share/Qfv3T0EVzqOXL_CYYeRr9Blv8HnNm4JsXBh7ByFjJbc

Tsung-Hsien (Shawn) Wen, CTO of PolyAI, tells Tim Scarfe why voice agents are harder than text agents. Voice adds time, and a good conversation depends on adapting to the person on the line, not just on reasoning to the best answer. Shawn describes an audio-native model (Dialog-RSN-1) that first predicts a turn-taking signal, then replies in text with citations, and writes the transcript last so enterprises can audit it.Along the way: training on real, noisy calls with synthetic noise added, and why over-cleaned audio made the new model worse. Latency, and what a voice agent should do while it thinks. Why a voice with a hint of regional accent beats a generic one. Why public benchmarks fall short for voice, why enterprises want to own their agent harness, and whether behaviour belongs in the harness or in the weights.The last stretch is about working with agents: cognitive debt, the shift from producing content to checking it, Wispr Flow, building tools that agents can use, and whether slop is in the eye of the reader.This episode was produced in partnership with PolyAI.https://poly.ai CHAPTERS0:00 Why voice agents are harder than text4:27 What enterprises want, and why PolyAI built its own model9:04 How an audio-native voice model works15:21 Training data, spectrograms and synthetic noise20:36 The cocktail party problem and the future of turn-taking25:21 Latency, adaptive reasoning and keeping callers' trust31:33 Voices, personality and the uncanny valley36:32 How do you benchmark a voice agent?42:00 Harness engineering and owning the intelligence45:09 Well-specified problems and auditable agents49:46 Weight adaptation and cognitive debt56:31 Agents at work: Wispr Flow, voice and tool building1:02:50 The next decade of voice, and what counts as slopREFERENCESThe Bitter Lesson: http://www.incompleteideas.net/IncIdeas/BitterLesson.html [9:05]Retrieval-augmented generation: https://arxiv.org/abs/2005.11401 [13:23]Mel scale: https://en.wikipedia.org/wiki/Mel_scale [17:16]Victor Zue: https://en.wikipedia.org/wiki/Victor_Zue [17:45]Cocktail party effect: https://en.wikipedia.org/wiki/Cocktail_party_effect [20:39]Speaker diarisation: https://en.wikipedia.org/wiki/Speaker_diarisation [21:16]

Leonardo de Moura created Lean and co-created Z3. ---This episode is sponsored by Parallel.Parallel, where agents find answers: web search, extraction and deep research APIs built for AI agents.Start free with the Parallel MCP server and $5 of credits every month: https://parallel.ai/mlst?utm_source=creator&utm_medium=podcast&utm_content=MLST---Tim Scarfe talks with Leo about how Lean escaped its original audience, why dependent types and Mathlib made it useful to working mathematicians, and what happens when formal verification leaves the lab. De Moura explains the small trusted kernel and independent checkers, and gives his account of the recent Collatz incident, in which a purported proof was accepted by both Lean's official kernel and Nanoda, apparently by exploiting a different bug in each.---TIMESTAMPS:00:00:00 Cold open: the green checkmark can lie00:00:59 Cathedral or bazaar: who controls Lean's core?00:04:55 Why Lean's core stays small and protected00:08:07 The Slack purge, Brandolini's law and the Lean FRO00:11:12 The Collatz exploit: two kernels, two bugs00:16:44 More kernels, reward hacking and safety by transparency00:21:04 Sponsor: Parallel00:21:59 Kim Morrison, Claude and the zlib proof00:25:25 Can we specify complex systems?00:28:04 Specs change: proofs are cheaper to redo with AI00:31:27 From Lean 1 to Lean 400:34:57 Dependent types in plain terms00:37:25 Lean 4's extensibility and Mathlib's growth00:41:44 Mathlib as infrastructure: Formal Frontiers00:44:15 Creativity, abstraction and nut-sniping00:48:46 Breadcrumbs, not learning: what AI agents lack00:53:03 Competence without comprehension, and verified guardrails00:56:21 Is the human still the author?01:01:44 AlphaProof, LLMs and why certificates still matter01:06:38 What's next for Lean, and its legacy01:11:42 How to start learning Lean---REFERENCES:tool:[00:00:48] Leanhttps://lean-lang.org/[00:01:24] Mathlibhttps://github.com/leanprover-community/mathlib4[00:12:09] nanoda_libhttps://github.com/ammkrn/nanoda_lib[00:12:19] CollatzLeanhttps://github.com/xrchz/CollatzLean/blob/a79357462a33d2a6babd4cf6c8d8bcd25425d653/README.md[00:13:06] Lean issue 14576https://github.com/leanprover/lean4/issues/14576[00:13:21] Lean pull request 14577https://github.com/leanprover/lean4/pull/14577[00:13:45] nanoda_lib pull request 22https://github.com/ammkrn/nanoda_lib/pull/22[00:15:33] Lean comparatorhttps://github.com/leanprover/comparator[00:18:02] Lean4Leanhttps://github.com/digama0/lean4lean[00:19:50] ARC-AGI-3https://arcprize.org/arc-agi/3[00:21:59] lean-ziphttps://github.com/kim-em/lean-zip[00:29:45] CompCerthttps://compcert.org/[00:29:45] seL4https://www.sel4.org/[00:30:23] Z3https://github.com/Z3Prover/z3[00:38:10] Veilhttps://github.com/verse-lab/veil[00:38:10] Velvethttps://github.com/verse-lab/velvetorganization:[00:00:52] Lean FROhttps://lean-lang.org/fro/[00:42:39] Mathlib Initiativehttps://mathlib-initiative.org/about/person:[00:05:11] Ilya Sergeyhttps://ilyasergey.net/[00:10:31] Joachim Breitnerhttps://www.joachim-breitner.de/[00:32:58] Adam Chlipalahttps://adam.chlipala.net/[00:38:42] Kevin Buzzardhttps://www.ma.imperial.ac.uk/~buzzard/[01:10:16] Terence Taohttps://terrytao.wordpress.com/book:[00:08:19] The Proof in the Codehttps://us.macmillan.com/books/9780374620059/theproofinthecode/other:[00:10:05] Brandolini's lawhttps://en.wikipedia.org/wiki/Brandolini%27s_law[00:45:33] A new result on unit distanceshttps://openai.com/index/model-disproves-discrete-geometry-conjecture/[01:01:08] Fermat's Last Theorem formalisationhttps://imperialcollegelondon.github.io/FLT/paper:[00:34:36] The Lean 4 theorem prover and programming languagehttps://doi.org/10.1007/978-3-030-79876-5_37---RESCRIPT:https://app.rescript.info/share/7d3d4a0059443236a01f6c9acbf4db58https://app.rescript.info/api/public/sessions/b007264c0ce89047/pdf

Weco let an AI coding agent rewrite the harness around another agent for eight days: its code, prompts and tools, while the underlying language model stayed fixed. Tim Scarfe asks Weco co-founder Zhengyao Jiang what the reported gains over two years of human engineering actually demonstrate.The discussion examines AIDE 85's generated code, held-out evaluation and the difficulty of separating useful discoveries from reward hacking. Jiang explains Weco's four levels of recursive self-improvement and compares the experiment with AlphaEvolve and the Darwin Gödel Machine.The limits matter as much as the gains. Jiang explains why the experiment did not establish that the system had become a better improver. The conversation closes with open-ended search, human-designed primitives and Parameter Golf: where does the next useful idea come from when the agent is searching inside a space that people designed?---TIMESTAMPS:00:00:00 Eight days of self-improvement: what counts?00:03:25 AIDE and the puzzle of useful spaghetti code00:08:38 Four levels of recursive self-improvement00:12:02 What AIDE 85 changed and how it was tested00:20:04 AlphaEvolve, Darwin Gödel Machine and the RSI claim00:26:21 Reward hacking and the limits of detection00:33:09 Open-ended search, harness tuning and creativity00:39:43 Parameter Golf and the limits of self-improvement---REFERENCES:organization:[00:00:30] Weco AIhttps://www.weco.ai/other:[00:00:33] AIDE²: The First Evidence of Recursive Self-Improvementhttps://www.weco.ai/blog/first-evidence-of-recursive-self-improvement[00:14:11] Faulty reward functions in the wildhttps://openai.com/index/faulty-reward-functions/[00:29:59] The Hugging Face incident and the road aheadhttps://openai.com/index/hugging-face-incident-and-the-road-ahead/tool:[00:03:29] AIDEhttps://github.com/WecoAI/aideml[00:04:29] MLE-benchhttps://github.com/openai/mle-bench[00:04:33] ALE-Benchhttps://github.com/SakanaAI/ALE-Bench[00:04:52] WeatherBench 2https://github.com/google-research/weatherbench2[00:08:18] ReActhttps://react-lm.github.io/[00:39:43] Parameter Golfhttps://github.com/openai/parameter-golfpaper:[00:20:08] AlphaEvolve: A coding agent for scientific and algorithmic discoveryhttps://arxiv.org/abs/2506.13131v1[00:21:35] Darwin Gödel Machine: Open-Ended Evolution of Self-Improving Agentshttps://arxiv.org/abs/2505.22954v3[00:23:45] Hyperagentshttps://arxiv.org/abs/2603.19461v1[00:27:01] SpecBench: Measuring Reward Hacking in Long-Horizon Coding Agentshttps://arxiv.org/abs/2605.21384book:[00:33:14] Why Greatness Cannot Be Planned: The Myth of the Objectivehttps://link.springer.com/book/10.1007/978-3-319-15524-1---LINKS:https://app.rescript.info/share/3a9dc6189cb539c6a05fcc4f75c101b3PDF:https://app.rescript.info/api/public/sessions/9eda60ede2b31c92/pdf

Frank Hutter, co-founder of Prior Labs, talks about TabPFN, a tabular foundation model that makes predictions in a single forward pass, and the research behind it.

TabPFN is pre-trained on synthetic datasets drawn from a prior over structural causal models, rather than on real data. At prediction time it takes the whole training table as context and outputs an approximation of the Bayesian posterior predictive distribution, without per-dataset training or hyperparameter search. Frank explains how this grew out of his earlier work on AutoML and neural architecture search, how the priors are built and revised, and why tabular data was hard for deep learning for so long.

The conversation also covers the TabArena benchmark, how the architecture changed from TabPFN v1 to v3, scaling to larger tables, using the model with coding agents, test-time compute, Google's TabFM, causal inference and interventions, and relational data. At the end, a short update Frank recorded after the interview covers the TabPFN-3.5 release.

Prior Labs:

TabPFN-3.5: https://priorlabs.ai/tabpfn-3-5

https://priorlabs.ai/careers

TOC:

00:00 Introduction

00:44 Welcome and Frank's background

02:05 Why tabular data was hard for deep learning

10:17 Pre-training on synthetic data

12:52 The TabArena benchmark

19:28 From AutoML to neural architecture search

26:34 TabPFN as a learned algorithm

30:50 Bayesian prediction in one forward pass

39:37 Scaling to larger tables

47:48 Using TabPFN with coding agents

57:47 Output heads and architecture from v1 to v3

1:05:29 Test-time compute and adaptation

1:13:32 Google's TabFM

1:16:53 How the priors are designed

1:18:40 Correlation, causation and interventions

1:35:22 Relational and multimodal data

1:38:31 Use in organisations

1:46:38 The open research arm

1:50:21 Update: TabPFN-3.5

REFS:

TabPFN v2, Nature (Hollmann et al., 2025)

https://www.nature.com/articles/s41586-024-08328-6

Transformers Can Do Bayesian Inference (Müller et al.)

https://arxiv.org/abs/2112.10510

TabArena (Erickson et al.)

https://arxiv.org/abs/2506.16791

AutoGluon-Tabular (Erickson et al.)

https://arxiv.org/abs/2003.06505

Beyond IID: How General Are Tabular Foundation Models, Really?

https://arxiv.org/abs/2606.30410

Neural Architecture Search: A Survey (Elsken, Metzen & Hutter)

https://arxiv.org/abs/1808.05377

Auto-WEKA (Thornton et al.)

https://www.cs.ubc.ca/~hutter/papers/AutoWEKA-KDD2013.pdf

TabPFN v1 (Hollmann et al., 2022)

https://arxiv.org/abs/2207.01848

TabPFN-3 technical report

https://arxiv.org/abs/2605.13986

TabPFN-2.5 report

https://arxiv.org/abs/2511.08667

CAAFE (Hollmann et al.)

https://arxiv.org/abs/2305.03403

TabICL (Qu et al.)

https://arxiv.org/abs/2502.05564

TabICLv2 (Qu et al.)

https://arxiv.org/abs/2602.11139

Google TabFM

https://research.google/blog/introducing-tabfm-a-zero-shot-foundation-model-for-tabular-data/

TALENT benchmark (Ye et al.)

https://arxiv.org/abs/2407.00956

Do-PFN (Robertson et al.)

https://arxiv.org/abs/2506.06039

CausalPFN (Balazadeh et al.)

https://arxiv.org/abs/2506.07918

Causal Foundation Models with Partial Graphs (Reuter et al.)

https://arxiv.org/abs/2602.14972

RelBench (Robinson et al.)

https://arxiv.org/abs/2407.20060

RelArena-α, TabPFN-Rel and RPI

https://arxiv.org/abs/2608.16319

TabPFN on GitHub

https://github.com/PriorLabs/TabPFN

TabPFN-3.5 technical report

https://arxiv.org/abs/2609.17895

Otto Group Product Classification Challenge (Kaggle, 2015)

https://www.kaggle.com/competitions/otto-group-product-classification-challenge

---RESCRIPT:https://app.rescript.info/share/e99676c25ee6189fbf54c9be07eb623e

Alexander Mattick is a researcher at Fraunhofer IIS and a PhD researcher at the University of Technology Nuremberg (UTN), and a regular on Yannic Kilcher's Discord. He first came on MLST in 2022, after helping research the Yann LeCun and Randall Balestriero episode on interpolation.

SPONSOR:

---

Cyber Fund built the Monastery to help founders ship products that were impossible a year ago. Applications for Batch 1 are now open.

Apply now: https://cyber.fund

---

Alexander treats inference as the thread running through modern machine learning: once you have a model, what does it cost to get an answer out of it? He works through Monte Carlo, GFlowNets, energy-based models, diffusion, normalising flows and flow matching, with four short explainers he recorded himself. He is blunt about energy-based models: you can sample from them in principle, but it is rarely worth the compute. JEPA and "world model", he says, are closer to branding than to technical categories.

Next: theories of deep learning, none of which he thinks predicts enough yet to guide practice, then reinforcement learning.

---

0:00 Cold open: information is expensive

0:51 Welcome back, Alexander Mattic

2:08 Alexander's research background

2:50 Inference: densities, sampling and Monte Carlo

6:42 GFlowNets, energy functions and MCMC

9:45 Explainer: energy-based models

11:03 Why model a density at all?

17:30 From learned energies to flow matching

25:08 Explainers: diffusion and normalising flows

28:33 Are energy-based models generative?

33:22 JEPA, contrastive learning and collapse

41:13 Why non-language modalities need flows

44:51 Inference as search: branch and bound

49:43 Q-learning and delayed consequences

55:14 Flow matching, optimal transport, Fokker-Planck

1:00:03 Explainer: flow matching

1:01:49 AlphaFold, latents and scale versus architecture

1:07:52 Two families of deep learning theory

1:15:04 What a good theory would predict

1:23:53 The manifold hypothesis and compression

1:28:25 Is reward enough?

1:32:01 Control theory versus reinforcement learning

1:37:22 The Bitter Lesson and expensive information

1:42:08 Constrained RL: the constrained MDP toolbox

1:50:12 Creativity as constrained search

1:55:44 Reality is protean: when abstractions hold

2:00:32 What is a world model?

2:04:38 Prediction is not control

2:08:13 Robot demos, MPC and reliability

---

REFERENCES:

[6:55] GFlowNets (Bengio et al., 2021)

https://arxiv.org/abs/2106.04399

[38:46] Contrastive Self-Supervised Learning (Anand, 2020)

https://ankeshanand.com/blog/2020/01/26/contrative-self-supervised-learning.html

[38:56] LeJEPA (Balestriero and LeCun, 2025)

https://arxiv.org/abs/2511.08544v3

[47:10] RL for Node Selection in Branch-and-Bound (Mattick)

https://openreview.net/forum?id=0ez68a5UqI

[56:20] Flow Matching for Generative Modeling

https://arxiv.org/abs/2210.02747v2

[1:12:41] Disentangling feature and lazy training in deep neural networks

https://arxiv.org/abs/1906.08034v4

[1:31:05] Reward is enough (Silver)

https://doi.org/10.1016/j.artint.2021.103535

[1:35:12] Learning ReLU networks to high uniform accuracy is intractable (Berner et al.)

https://arxiv.org/abs/2205.13531v2

[1:40:20] Dota 2 with Large Scale Deep RL

https://arxiv.org/abs/1912.06680v1

[1:45:41] Constrained Update Projection for Safe Policy Optimization (Yang et al., 2022)

https://arxiv.org/abs/2209.07089

[1:46:11] SafeMPO (ICLR 2026)

https://openreview.net/forum?id=1m0EU6QXj6

[1:50:17] Why Creativity Cannot Be Interpolated

https://archive.mlst.ai/paper/why-creativity-cannot-be-interpolated/

[1:51:39] Invalid Action Masking (Huang and Ontañón)

https://arxiv.org/abs/2006.14171

[2:00:04] Probability Theory: The Logic of Science (Jaynes, 2003)

https://www.cambridge.org/core/books/probability-theory/9CA08E224FF30123304E6D8935CF1A99

[2:01:53] Training Agents Inside of Scalable World Models (Hafner et al., 2025)

https://arxiv.org/abs/2509.24527v1

[2:03:43] World Models (Ha and Schmidhuber, 2018)

https://arxiv.org/abs/1803.10122v4

The car making a left turn at the start of this episode was never filmed. Cosmos 3 generated it. Ming-Yu Liu, who leads the Cosmos research at NVIDIA, explains how one model can describe a video, generate one, and produce robot actions.

He walks Tim through the architecture. A vision language model reasons one token at a time; its weights then initialise a bidirectional diffusion generator for video, audio and action, and a shared temporal position scheme lines up signals that run at different rates. Ming-Yu treats "world model" as a set of tools, not one definition: forward dynamics, inverse dynamics and policy, trained together under a capacity limit so that each helps the others. He also explains why plentiful first-person human video carries over to robots, which have far less data of their own, and why a Cosmos model post-trained on the DROID dataset is a good starting point for pick-and-place policies.

The most practical thread is testing. A neural simulator does not need accurate success rates. It only needs to rank policy A above policy B the way the real world would, so a team can narrow down which checkpoints deserve a real trial. Cosmos Dreams applies that closed-loop idea to driving and robotics, and Ming-Yu argues that humanoids around children and pets make safety matter even more than it does for cars. The conversation ends on the Super, Nano and Edge sizes (Edge targets Jetson Thor, Orin and DGX Spark) and where to find the open weights, code and data.

This episode is a paid partnership with NVIDIA.

Learn more about Cosmos: https://nvda.ws/4cJoY1S

Explore Cosmos Lab: https://research.nvidia.com/labs/cosmos-lab/cosmos3/

---

TIMESTAMPS:

00:00:00 A road that was never filmed

00:02:28 Inside Cosmos 3: reasoning and generator towers

00:05:02 World models: dynamics, policy and one clock

00:08:59 Learning robot skills from human video

00:11:06 Ambiguous tasks and system 2 planning

00:12:53 Neural simulators for policy verification

00:16:41 Cosmos as a starting point for robot policies

00:19:00 Cosmos Dreams and robot safety

00:22:04 Super, Nano and Edge model sizes

00:24:24 Open models, the Cosmos repo and feedback

---

REFERENCES:

tool:

[00:00:13] Cosmos 3 (NVIDIA Cosmos Lab project page)

https://research.nvidia.com/labs/cosmos-lab/cosmos3/

[00:18:27] NVIDIA Cosmos GitHub repository

https://github.com/NVIDIA/cosmos

[00:22:05] Cosmos3-Edge model card

https://huggingface.co/nvidia/Cosmos3-Edge

[00:22:15] Cosmos3-Super model card

https://huggingface.co/nvidia/Cosmos3-Super

[00:22:16] Cosmos3-Nano model card

https://huggingface.co/nvidia/Cosmos3-Nano

[00:22:50] NVIDIA Jetson Thor

https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/jetson-thor/

[00:22:52] NVIDIA Jetson Orin

https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/jetson-orin/

[00:22:53] NVIDIA DGX Spark

https://www.nvidia.com/en-us/products/workstations/dgx-spark/

[00:24:42] Cosmos 3 collection on Hugging Face

https://huggingface.co/collections/nvidia/cosmos3

other:

[00:01:07] Cosmos-Dreams closed-loop simulators (NVIDIA SIGGRAPH 2026 blog)

https://blogs.nvidia.com/blog/siggraph-news-2026/

paper:

[00:08:54] Cosmos 3: Omnimodal World Models for Physical AI

https://arxiv.org/abs/2606.02800

[00:17:43] DROID: A Large-Scale In-The-Wild Robot Manipulation Dataset

https://arxiv.org/abs/2403.12945

---

RESCRIPT: https://app.rescript.info/share/e2385948cf465f0d6a2c0930150fc3ab

Pavan Kumar Reddy leads audio research at Mistral AI. He joins Tim Scarfe for a deep technical tour of Voxtral — and explains why the frontier of deployed voice is still a cascade of specialised models rather than one end-to-end system.

IN PARTNERSHIP WITH MISTRAL AI:

---

This episode was produced in partnership with Mistral AI.

Mistral AI: https://mistral.ai/

---

The conversation opens on architecture. Voxtral Chat feeds a 3B Ministral text trunk with continuous embeddings from an audio encoder, passed to the decoder as direct token input rather than through cross-attention as in Whisper, so the model can answer questions about emotion, timing and who spoke when without an intermediate transcript to lose them. The real-time model becomes a dual-stream decoder that consumes audio and emits text at once, at a target delay down to 160ms, with slower streams in parallel for anything that can wait for more context.

On generation, Pavan explains why Voxtral TTS predicts continuous latents rather than discrete codec tokens, traces the lineage from SoundStream through EnCodec to Mimi's split of semantic and acoustic codebooks, and places FSQ and flow matching in it. Tim presses on the priors underneath: why a mel spectrogram instead of raw waveform, what noise augmentation buys, and when acoustic overfitting becomes somebody's fine-tuning problem. Then the failure modes. Diarisation is emitted autoregressively inside the transcript rather than by a separate head, which makes streaming diarisation fragile — less context, late speaker changes, invented extra speakers. And because the architecture commits to what it has already predicted, one out-of-distribution mistake compounds into looping or skipped segments, which is what DPO corrects: the negative supervision pre-training and SFT cannot give.

The last third is the argument Tim keeps returning to. Customers running voice agents over millions of sessions describe scaffolding, not a solved problem, with a sharp drop outside the top few languages. Cascades survive because each component stays separately adaptable, observable and constrainable. And voice alone is cognitive debt: absorbing information and deciding in one serial stream is harder than glancing at a menu. Voice becomes ubiquitous beside a screen, not instead of one.

---

TIMESTAMPS:

00:00:00 Cold open

00:00:46 Why Mistral moved into audio

00:09:27 Inside Voxtral: trunk, encoder, dual streams

00:20:22 Speech that works in real time

00:30:52 How a voice becomes tokens

00:39:59 Flow matching, FSQ and the new codec

00:52:51 When speech models lose the speaker

01:03:23 Correcting hallucinations with preferences

01:12:12 Controlling synthetic speech

01:20:06 Why cascades still win

01:29:25 Speech in the wild

01:33:46 Audio models as interfaces

01:37:54 Why voice still needs a screen

---

REFERENCES:

paper:

[00:01:42] Mistral 7B

https://arxiv.org/abs/2310.06825

[00:09:38] Voxtral

https://arxiv.org/abs/2507.13264

[00:14:41] Whisper: Robust Speech Recognition

https://arxiv.org/abs/2212.04356

[00:19:11] Voxtral Realtime

https://arxiv.org/abs/2602.11298

[00:21:52] Delayed Streams Modeling (Kyutai)

https://arxiv.org/abs/2509.08753

[00:30:52] Voxtral TTS

https://arxiv.org/abs/2603.25551

[00:32:38] SoundStream neural audio codec

https://arxiv.org/abs/2107.03312

[00:34:59] Flow Matching for Generative Modeling

https://arxiv.org/abs/2210.02747

[00:37:03] EnCodec: High Fidelity Neural Audio Compression

https://arxiv.org/abs/2210.13438

[00:37:42] Moshi and the Mimi codec

https://arxiv.org/abs/2410.00037

[00:39:05] Finite Scalar Quantization (FSQ)

https://arxiv.org/abs/2309.15505

[01:03:33] Direct Preference Optimization (DPO)

https://arxiv.org/abs/2305.18290

dataset:

[00:46:14] Mozilla Common Voice

https://commonvoice.mozilla.org/en/datasets

organization:

[00:50:47] Hugging Face

https://huggingface.co/

Can a machine learn the judgement that separates a plausible-looking result from a faithful experiment? Edward Hughes, Chief Scientist and co-founder of Inherent, joins Tim Scarfe to argue that creativity is not optimisation, and that the missing capability in AI is choosing which questions are worth asking.

SPONSOR:

---

Cyber Fund built the Monastery to help founders ship products that were impossible a year ago.

Apply now: https://cyber.fund

---

Edward makes the case that Move 37 was innovative rather than creative, and that the field, not the individual, decides what counts as a discovery. That reframing runs through Csikszentmihalyi, Deutsch and exaptation into open-endedness, where deceptive goals and imperfect world models turn out to be the point rather than the problem. The second half turns to the paper: Replica, a task space built by redacting figures from real papers, and Faraday, a 27-billion-parameter model trained to steer a frontier coding agent that then beats the frontier on held-out replications.

---

TIMESTAMPS:

00:00:00 Cold open: Move 37, Faraday and collective intelligence

00:01:08 Sponsor: CyberFund

00:01:46 Inherent's $50M raise and the road from string theory

00:09:14 Three timescales of learning: weights, context, culture

00:13:47 Move 37 was innovative, not creative: the field decides

00:20:39 Creativity as satisficing: the urinal and evolution

00:25:06 Exaptation and the Tristan chord: creativity in context

00:30:56 Coherence for whom? Deutsch's hard-to-vary explanations

00:35:53 Why copying is creative: Deutsch and the constraint engineer

00:42:27 Societies of agents and the strong Moravec paradox

00:45:51 Evaluate in hindsight: from Lean proofs to climate change

00:51:56 Picbreeder, local goals and why discovery needs deception

00:57:21 Spaghetti proofs, translation layers and superhuman Go

01:00:37 Does nature compress? Naturalness and real patterns

01:07:36 Why replicate? Replica's redacted figures and Faraday

01:12:31 Faraday beats Codex, Claude and GLM 5.2 on held-out tasks

01:15:31 Replication to innovation: how the Transformer happened

01:18:26 Deep replication: what Faraday learns from Voyager and GNoME

01:23:37 Can the AI scientist cheat? Goodharting the judge

01:29:09 Inside Replica: scale-down, 8xB300 runs, per-task rubrics

01:34:11 The RL crisis: getting GRPO to work with per-turn credit

01:39:43 Weights vs harnesses: AlphaEvolve, DGM and EvoTune

01:45:45 The recursive company: agents cross a phase transition

01:50:35 Collective intelligence and the electric dynamo

01:55:46 What replaces OKRs? Incumbents and the burden of knowledge

---

REFERENCES:

MLST Creativity Article:

https://archive.mlst.ai/read/why-creativity-cannot-be-interpolated

organization:

[00:01:47] Inherent

https://inherentlabs.ai/

other:

[00:20:51] Marcel Duchamp, Fountain

https://www.tate.org.uk/art/artworks/duchamp-fountain-t07573

[00:05:19] Human-Timescale Adaptation in an Open-Ended Task Space (Adaptive Agent)

https://arxiv.org/abs/2301.07608

[00:06:05] The AI Scientist

https://arxiv.org/abs/2408.06292

[00:12:13] Training AI Scientists to Replicate Research (Replica and Faraday)

https://arxiv.org/abs/2608.13331

[01:44:46] Evolutionary Principles in Self-Referential Learning

https://people.idsia.ch/~juergen/diploma.html

[01:59:33] Are Ideas Getting Harder to Find?

https://www.nber.org/papers/w23782

book:

[00:16:04] Creativity: Flow

https://search.worldcat.org/title/254487436

[00:26:22] Why Greatness Cannot Be Planned

https://link.springer.com/book/10.1007/978-3-319-15524-1

[00:33:03] The Beginning of Infinity

https://www.penguinrandomhouse.com/books/293575/the-beginning-of-infinity-by-david-deutsch/

[01:55:47] Laws of Knowledge

https://www.penguin.co.nz/books/the-infinite-alphabet-9780241655672

(Full list refs on YT/rescript)

---

RESCRIPT:

https://app.rescript.info/session/670296ba913761d0?share=6281911cac9bdbff637f10819d4d1e5c

Could slowing AI development make superintelligence safer? Daniel Kokotajlo and Thomas Larsen of the AI Futures Project join Tim Scarfe to examine AI 2040: Plan A, a proposal to buy time before AI exceeds human control.

SPONSOR:

---

Cyber Fund built the Monastery to help founders ship products that were impossible a year ago.

Apply now: https://cyber.fund

---

After revisiting AI 2027 and the limits of forecasting, they ask what happens when AI can automate research and sustain an economy without human workers. Tim challenges the case for general models and asks whether intelligence alone explains power. Plan A proposes an initial pause to build safety infrastructure, then cautious development up to the strongest AI that can still be reliably controlled. The discussion tests the distinction between control and alignment, the case for public AI research, and whether the US and China could enforce a slowdown. It ends with the evidence that would change their forecasts.

---

TIMESTAMPS:

00:00:00 AI 2040: a slower route to superintelligence

00:01:34 Sponsor: Cyber Fund

00:02:12 From OpenAI to AI 2027

00:06:58 Forecasts, war games and self-fulfilling prophecies

00:17:44 Why AI sceptics are changing their minds

00:23:04 When AI can replace its own researchers

00:28:45 Could an AI economy grow without human workers?

00:37:32 One general model or a society of specialists?

00:47:43 Brains, machines and collective intelligence

00:56:12 Plan A: buy time at the controllable frontier

01:00:02 Why control buys time but cannot replace alignment

01:06:36 Why AI research should be public

01:10:32 Can the US and China enforce an AI slowdown?

01:19:04 Why AI policy debates miss the technology

01:21:56 Is AI normal technology? The remaining disagreement

Many thanks to James Wilken-Smith for helping with show research.

---

REFERENCES:

other:

[00:00:01] AI 2040: Plan A

https://ai-2040.com/

[00:03:27] AI 2027

https://ai-2027.com/

[00:13:47] Scenario Scrutiny for AI Policy

https://blog.aifutures.org/p/scenario-scrutiny-for-ai-policy

[00:33:11] The 2028 Global Intelligence Crisis

https://www.citriniresearch.com/p/2028gic

[01:00:40] Brief independent investigation of agents' behavior, reasoning and collaboration in the OpenAI / Hugging Face hacking incident

https://www.redwoodresearch.org/research/hugging-face-incident

[01:09:21] The Hugging Face incident and the road ahead

https://openai.com/index/hugging-face-incident-and-the-road-ahead/

[01:22:01] AI as Normal Technology

https://www.normaltech.ai/p/ai-as-normal-technology

[01:22:51] Common Ground between AI 2027 & AI as Normal Technology

https://asteriskmag.substack.com/p/common-ground-between-ai-2027-and

person:

[00:19:43] Geoffrey Hinton

https://www.cs.toronto.edu/~hinton/

[00:20:07] Ryan Greenblatt

https://www.lesswrong.com/users/ryan_greenblatt

[00:26:06] Elon Musk

https://www.tesla.com/elon-musk

tool:

[00:21:46] ARC-AGI-3

https://arcprize.org/arc-agi/3

[00:21:53] AlphaGo and Move 37

https://deepmind.google/research/alphago/

[00:39:41] Claude

https://claude.com/product/overview

[00:39:58] NVIDIA H100 GPU

https://www.nvidia.com/en-us/data-center/h100/

paper:

[00:24:42] Training AI Scientists to Replicate Research

https://arxiv.org/abs/2608.13331v1

[01:27:19] Validity of the single processor approach to achieving large scale computing capabilities

https://www.cs.cmu.edu/~18742/papers/Amdahl1967.pdf

book:

[00:28:52] Bullshit Jobs: A Theory

https://www.simonandschuster.com/books/Bullshit-Jobs/David-Graeber/9781501143335

organization:

[01:05:09] Redwood Research

https://www.redwoodresearch.org/

---

RESCRIPT:

https://app.rescript.info/public/share/33d1a58fa8f307ae7dfd504d4fdaa9d5

Tom McGrath is co-founder and Chief Scientist at Goodfire, and a former Google DeepMind researcher. He joins Tim Scarfe to ask what neural networks actually learn, whether their internal representations converge on structures in the world, and whether interpretability can extract new scientific knowledge rather than merely explain model outputs.

Beginning with AlphaZero and learned modularity, the conversation moves into neural geometry: concept manifolds, reusable computation inside Llama, and why activation steering can fail when it pushes a model off-manifold. McGrath then makes the case for intentional design, using interpretability as part of the training loop. They examine controlled generalisation, features as rewards, predictive data debugging, and the uncomfortable fact that a model may recognise a hallucination or reward hack and still produce it.

The discussion closes on grader awareness, oversight and collusion between adaptive agents, then returns to sparse autoencoders. SAEs are useful, McGrath argues, but they may fracture the higher-dimensional structures networks actually use. This episode was made with support from Goodfire.

---

TIMESTAMPS:

00:00:00 Introduction: Can interpretability speed-run science?

00:02:03 The invisible grader

00:06:51 What AlphaZero learned from the world

00:12:24 Interpretability as a control loop

00:21:54 The forbidden method and safer interventions

00:37:36 Why models catch hallucinations too late

00:46:19 Debug the dataset before training

00:50:44 Why neural networks become modular

00:55:57 Finding the geometry inside a network

01:02:55 Why steering falls off the manifold

01:12:10 A reusable calculator inside Llama

01:17:19 From abstractions to goals

01:25:28 Reward hacking, oversight and collusion

01:37:23 Are sparse autoencoders dead?

---

REFERENCES:

paper:

[00:05:45] Emergent Misalignment: Narrow finetuning can produce broadly misaligned LLMs

https://arxiv.org/abs/2502.17424v7

[00:11:05] Acquisition of Chess Knowledge in AlphaZero

https://arxiv.org/abs/2111.09259

[00:25:30] Steering Out-of-Distribution Generalization with Concept Ablation Fine-Tuning

https://arxiv.org/abs/2507.16795

[00:29:30] Persona Vectors: Monitoring and Controlling Character Traits in Language Models

https://arxiv.org/abs/2507.21509

[00:41:14] Features as Rewards: Scalable Supervision for Open-Ended Tasks via Interpretability

https://arxiv.org/abs/2602.10067

[00:47:03] Anatomy of Post-Training: Using Interpretability to Characterize Data and Shape the Learning Signal

https://arxiv.org/abs/2606.12360

[01:00:26] Do Sparse Autoencoders Capture Concept Manifolds?

https://arxiv.org/abs/2604.28119

[01:03:04] Manifold Steering Reveals the Shared Geometry of Neural Network Representation and Behavior

https://arxiv.org/abs/2605.05115

[01:14:20] Arithmetic in the Wild: Llama uses Base-10 Addition to Reason About Cyclic Concepts

https://arxiv.org/abs/2605.01148

[01:29:35] Measuring Reward-Seeking via Contrastive Belief Updates

https://arxiv.org/abs/2607.18966v1

other:

[00:15:44] Intentional Design

https://www.goodfire.com/blog/intentional-design

[00:56:12] The World Inside Neural Networks

https://www.goodfire.com/research/the-world-inside-neural-networks

[01:37:28] A Pragmatic Vision for Interpretability

https://www.alignmentforum.org/posts/StENzDcD3kpfGJssR/a-pragmatic-vision-for-interpretability

---

RESCRIPT:

https://app.rescript.info/share/846cfee4131b664fd09209cc3b98018e

Tim Scarfe speaks with Ilia Shumailov and Alexander Panfilov about their paper, Stealing Reasoning Traces from Proprietary LLM APIs.The core bug sounds deceptively simple: providers return encrypted reasoning state so conversations can be resumed or forked. But those blobs can be replayed across users and sibling models. A smaller model can ask the provider to decrypt the trace, then repeat the hidden reasoning in plain text. The discussion covers leaked private data, a broadly reusable jailbreak, poisoned agent traces, chain-of-thought monitoring, responsible disclosure, and possible defenses.Ilia Shumailov is an AI and security researcher, formerly at Google DeepMind, who completed his Cambridge PhD under Ross Anderson. Alexander Panfilov is a PhD researcher at the ELLIS Institute Tübingen and the Max Planck Institute for Intelligent Systems, working on AI safety, adversarial machine learning, and LLM red-teaming. They close by separating the demonstrated jailbreaking threat from ordinary benign distillation, and by arguing for controlled experiments over sweeping claims.---TIMESTAMPS:00:00:00 Intro montage00:01:33 Portable encrypted thought and decoded reasoning00:24:55 How the attack works and what it means00:39:04 Doom, defense, and scientific restraint---REFERENCES:paper:[00:00:00] Stealing Reasoning Traces from Proprietary LLM APIshttps://arxiv.org/abs/2608.09867[00:09:22] Chain of Thought Monitorability: A New and Fragile Opportunity for AI Safetyhttps://arxiv.org/abs/2507.11473[00:11:30] Reasoning Models Don’t Always Say What They Thinkhttps://www.anthropic.com/research/reasoning-models-dont-say-think[00:37:22] PostTrainBench: Can LLM Agents Automate LLM Post-Training?https://arxiv.org/abs/2603.08640[00:41:02] Large-scale online deanonymization with LLMshttps://arxiv.org/abs/2602.16800other:[00:09:28] OpenAI and Hugging Face partner to address security incident during model evaluationhttps://openai.com/index/hugging-face-model-evaluation-security-incident/[00:10:22] Claude, GPT, and Gemini All Struggle to Evade Monitorshttps://metr.org/notes/2025-08-22-claude-gpt-gemini-struggle-evade-monitors/tool:[00:42:08] Isabelle proof assistanthttps://isabelle.in.tum.de/---RESCRIPT: https://app.rescript.info/share/07fc38276e0823dc9b8986c32e202c7f

Astrophysicist Adam Becker, author of "What Is Real?", joins Tim Scarfe to take apart the futures Silicon Valley keeps selling: the 2045 singularity, mind uploading, Mars colonies, and the AI apocalypse. His new book *More Everything Forever* argues these ideas are hugely influential, mostly evidence-free, and bankrolled by tech billionaires who need a story in which growth never ends.Becker does the physics the boosters skip. Kurzweil's "law of accelerating returns" rests on cherry-picked data, and every exponential ends. Grant Bezos his perpetual energy growth and humanity boils the oceans within a few centuries, then exhausts the observable universe in under 4,000 years. The stars are too far away, Mars dirt is poison, and the day the dinosaur-killing asteroid hit Earth was still nicer than any day on Mars. On AI, Becker calls LLMs pocket calculators for language: hallucination is the model doing exactly what it always does, and the intelligence explosion assumes intelligence is a single number you can buy with compute.The sting is that Becker thinks the doomers are sincere. Yudkowsky, Bostrom and the effective altruists are not grifters, he says, just wrong, and their warnings that AI could end the world feed the same growth story the money depends on. He closes with his own prescription: take social problems seriously, regulate the whole tech industry, and tax billionaires out of existence.---TIMESTAMPS:00:00:00 Cold open and the thesis of More Everything Forever00:04:24 Kurzweil's singularity and the physical limits of exponential growth00:14:02 High agency and the fantasy of imprinting humanity on the cosmos00:16:55 Mind uploading, functionalism, and embodied cognition00:24:24 AI psychosis and anthropomorphizing LLMs00:26:24 Calculators, hallucination, and the limits of scale00:32:20 Yudkowsky and the intelligence-explosion argument00:40:37 True believers, venture capital, and the sci-fi growth narrative00:47:21 From Extropians to EA: utilitarianism and longtermism00:53:50 Brain worms and Becker's prescription: take social science seriously00:56:49 Why the AI-ethics discourse is broken01:01:42 The eugenics and IQ argument against 'intelligence'01:06:07 Why space settlement fails: Mars, the moon, and orbital data centers01:10:42 Billionaire myths and the search for purpose01:13:38 Tax billionaires, regulate tech: closing prescriptions---REFERENCES:book:[00:00:07] More Everything Forever (Adam Becker, 2025)https://www.hachettebookgroup.com/titles/adam-becker/more-everything-forever/9781541619593/[00:00:15] What Is Real? (Adam Becker, 2018)https://en.wikipedia.org/wiki/What_Is_Real%3F[00:15:46] What We Owe the Future (Will MacAskill, 2022)https://www.hachettebookgroup.com/titles/william-macaskill/what-we-owe-the-future/9781541618626/other:[00:00:27] Dreaming Against the Machine (podcast)https://www.dreamingagainstthemachine.com[00:01:04] The Useful Idiots of AI Doomsaying (Adam Becker, The Atlantic, 2025)https://www.theatlantic.com/books/archive/2025/09/what-ais-doomers-and-utopians-have-in-common/684270/

RESCRIPT: https://app.rescript.info/share/d6e37f9866673d8f74a39076efa5926b

This episode is sponsored by Notion. Learn more about Notion's Developer Platform today at https://notion.com/mlstWhy can deep networks discover abstractions that shallow models miss? Statistical physicist Matthieu Wyart joins Tim Scarfe to argue that the answer lies in the hidden hierarchy of data. Language and images are built from parts within parts; depth lets a network recover those coarse-grained variables and escape the curse of dimensionality.The conversation moves from jamming transitions and rough loss surfaces to Chomsky, context-free grammars and machine creativity. Wyart explains why next-token prediction can still recover compositional structure, where current systems fall short of genuine scientific invention, and why predicting latent representations rather than raw tokens could make learning far more sample-efficient.They also examine diffusion models, neural scaling laws and the limits of physics-inspired theory. The final question is on a personal note: if mistakes are the price of leaving the beaten path, how much scientific risk is worth taking?---TIMESTAMPS:00:00:00 Can machines learn abstractions from data?00:02:00 Notion agentic workspace00:02:49 From statistical physics to machine learning00:06:40 What physics can explain about learning00:16:37 From Carnot to Chomsky bulldozer00:21:21 How deep networks recover hidden hierarchies00:32:43 Where machine creativity still falls short00:40:48 How deep nets escape the curse of dimensionality00:52:19 Why predict latents instead of tokens01:02:49 The sample-efficiency case for latent prediction01:08:31 Diffusion, scaling laws and text entropy01:16:40 The scientists we learn from and the mistakes we make---REFERENCES:person:[00:00:43] Noam Chomskyhttps://linguistics.mit.edu/user/chomsky/tool:[00:02:08] Notion Developer Platformhttps://www.notion.com/en-gb/blog/introducing-developer-platformpaper:[00:04:43] Mastering the game of Go with deep neural networks and tree searchhttps://www.nature.com/articles/nature16961[00:05:52] Reconciling modern machine-learning practice and the bias-variance trade-offhttps://arxiv.org/abs/1812.11118[00:25:54] How Deep Neural Networks Learn Compositional Data: The Random Hierarchy Modelhttps://arxiv.org/abs/2307.02129[00:42:12] Efficient Estimation of Word Representations in Vector Spacehttps://arxiv.org/abs/1301.3781[00:52:46] Self-Supervised Learning from Images with a Joint-Embedding Predictive Architecturehttps://arxiv.org/abs/2301.08243[00:52:54] Learn from your own latents and not from tokens: A sample-complexity theoryhttps://arxiv.org/abs/2605.27734[01:08:31] A Phase Transition in Diffusion Models Reveals the Hierarchical Nature of Datahttps://arxiv.org/abs/2402.16991[01:11:39] Scaling Laws for Neural Language Modelshttps://arxiv.org/abs/2001.08361[01:12:17] Deriving Neural Scaling Laws from the statistics of natural languagehttps://arxiv.org/abs/2602.07488[01:13:34] Prediction and Entropy of Printed Englishhttps://ieeexplore.ieee.org/document/6773263---LINKS:Download PDF transcript: https://app.rescript.info/share/f7644cdaa86c5cc1e41e484e290f2bd4

Can an AI do the right thing for the wrong reason? Tim Scarfe speaks with Apollo Research’s Alexander Meinke, Axel Højmark and Jérémy Scheurer about Measuring Reward-Seeking via Contrastive Belief Updates, their new research with OpenAI.

The panel asks how models infer what graders reward, why good behaviour can come from the wrong reason, and whether that difference can be measured. The conversation moves through promise-breaking, grader awareness, reward hacking, scheming, opaque reasoning and corrigibility, then turns to a detailed walkthrough of the contrastive-belief method and what its results do and do not show. The o3 results discussed here concern an intermediate checkpoint without safety training.

This episode was made in partnership with Apollo Research. MLST retained full editorial control.

Reference

Apollo Research: https://www.apolloresearch.ai/

---

TIMESTAMPS:

00:00:00 Cold Open

00:02:12 Right Things, Wrong Reasons

00:12:47 Grader Awareness

00:26:22 Legibility

00:32:35 What To Call It

00:35:58 Intelligence, Agency, Anthropomorphism

00:45:16 Apollo’s Mission

00:48:54 The End of the Exponential

00:55:45 The Paper

01:16:34 Closing Reflection

---

REFERENCES:

tool:

[00:00:08] Claude Fable

https://www.anthropic.com/claude/fable

[00:12:50] AlphaGo Zero

https://deepmind.google/blog/alphago-zero-starting-from-scratch/

[00:44:30] AlphaFold 3

https://deepmind.google/science/alphafold/

paper:

[00:01:02] Measuring Reward-Seeking via Contrastive Belief Updates

https://arxiv.org/abs/2607.18966

[00:16:19] Natural Language Autoencoders Produce Unsupervised Explanations of LLM Activations

https://transformer-circuits.pub/2026/nla/

[00:26:48] Stress Testing Deliberative Alignment for Anti-Scheming Training

https://arxiv.org/abs/2509.15541

[00:35:33] Shortcut learning in deep neural networks

https://arxiv.org/abs/2004.07780

[00:53:49] Measuring AI Ability to Complete Long Software Tasks

https://arxiv.org/abs/2503.14499

[00:59:52] Modifying LLM Beliefs with Synthetic Document Finetuning

https://alignment.anthropic.com/2025/modifying-beliefs-via-sdf/

[01:10:44] Alignment Faking in Large Language Models

https://arxiv.org/abs/2412.14093

[01:13:55] Natural Emergent Misalignment from Reward Hacking

https://www.anthropic.com/research/emergent-misalignment-reward-hacking

other:

[00:10:14] We Need a Science of Scheming

https://www.apolloresearch.ai/science/science-of-scheming/

[00:32:56] CoastRunners reward hacking example

https://deepmind.google/blog/specification-gaming-the-flip-side-of-ai-ingenuity/

organization:

[01:06:07] Redwood Research

https://www.redwoodresearch.org/

---

ReScript:

https://app.rescript.info/share/718ab68e18cfa3b9b800da6b3290fd42

This episode is sponsored by Notion. Learn more about Notion's Developer Platform today at https://notion.com/mlst

Britain's most capable coding model can't be exported, and that ban is the whole reason Cosine set out to build one from scratch. Alistair Pullen, CEO and co-founder of Cosine, sits down with Tim Scarfe to explain how a frontier system he calls Fable, locked behind US export controls, became the founding case for a UK sovereign model trained on the Isambard supercomputer in Bristol.

The bet underneath it is economic. Pullen argues that an inference company, rather than a training-first lab, doesn't need billions to compete: millions, a national compute allocation, and a consortium feedback loop can be enough. From there it gets into the machinery, why open-weight models still trail the frontier on size, active parameters and data, the mixture-of-experts versus dense trade-off and why active params dominate how a model actually feels, and the edge that real coding trajectories confer.

The back half is about making agents trustworthy. Pullen makes the case for beating "slop" by rewarding the process instead of the final answer, reframes code review as runtime proof (spin the bug up in a VM and force the agent to actually exploit it), and walks through Swarm, Cosine's system running hundreds of sub-agents in one shot. It ends on why memory is still an unsolved hack, how synthetic graders let you run RL on tasks with no built-in test, and why Pullen reads US export controls as an accidental gift, with a supply-chain sting in the tail.

---

TIMESTAMPS:

00:00:00 The sovereign mandate and the Fable ban

00:04:02 Millions vs billions: the inference-company model

00:07:19 The consortium feedback loop

00:07:40 Why open models lag the frontier

00:14:59 MoE vs dense, and why active params matter

00:16:29 Trajectories: the process-data advantage

00:19:48 Beating slop: reward the process, not the answer

00:26:06 Reusable abstractions and the epistemic wall

00:29:56 Code review becomes runtime proof

00:37:32 Do agentic harnesses still matter?

00:40:35 Swarm: orchestrating hundreds of sub-agents

00:45:14 Why memory is still unsolved

00:48:25 Synthetic data and graders for RL

00:53:09 The US export gift and supply-chain risk

---

REFERENCES:

organization:

[00:01:15] Cosine

https://cosine.sh

[00:04:14] Mistral AI

https://mistral.ai

[00:05:50] Anthropic

https://www.anthropic.com

[00:07:42] Cohere

https://cohere.com

[00:08:36] DeepSeek

https://www.deepseek.com

tool:

[00:02:52] Isambard-AI

https://isambard.ac.uk

[00:05:56] Colossus (xAI)

https://en.wikipedia.org/wiki/Colossus_(supercomputer)

[00:07:52] GLM (Z.ai)

https://z.ai

[00:11:52] NVIDIA B300

https://www.nvidia.com/en-us/data-center/dgx-b300/

[00:15:37] gpt-oss-120b

https://huggingface.co/openai/gpt-oss-120b

[00:15:52] Devstral 2

https://mistral.ai/news/devstral

[00:16:01] Llama 70b

https://www.llama.com

[00:17:05] Claude Code

https://www.anthropic.com/claude-code

[00:26:23] ARC-AGI (Francois Chollet)

https://arcprize.org

[00:40:38] Swarm (Cosine)

https://cosine.sh

[00:40:50] OpenAI Codex

https://github.com/openai/codex

[00:41:16] Lumen Outpost (Cosine)

https://cosine.sh

[00:41:18] Kimi K2 (Moonshot)

https://huggingface.co/moonshotai/Kimi-K2-Instruct

[00:49:55] SWE-bench

https://www.swebench.com

[00:52:40] SystemVerilog

https://en.wikipedia.org/wiki/SystemVerilog

person:

[00:23:40] Andrej Karpathy

https://karpathy.ai

paper:

[00:27:10] GRPO (DeepSeekMath)

https://arxiv.org/abs/2402.03300

[00:27:13] GSPO

https://arxiv.org/abs/2507.18071

Incompressible Knowledge Probes, Bojie Li

https://arxiv.org/pdf/2604.24827

Estimating the Size of Claude Opus 4.5/4.6

https://unexcitedneurons.substack.com/p/estimating-the-size-of-claude-opus

---

ReScript:

https://app.rescript.info/session/5852d2b884c4ce4b?share=10b9799160845bb11779f8ac6cd3124f

Tim Scarfe travels to Zurich to sit down with the Tufa Labs ARC-AGI-3 team — founder Benjamin Crouzier, with Jeroen Cottaar, Dries Smit, Stefano Viel and Michal Tesnar — to work out what their leaderboard-topping system does and what the benchmark is really testing.The cut opens on the games: a walkthrough of the Locksmith game, where you read the rules of an unfamiliar world straight from raw frames. ARC-AGI-3 makes ARC interactive and agentic, so the model has to *discover* the goal rather than transduce a static grid. It stays easy for humans and breaks LLMs, and it runs through everything that follows. Dries traces his StochasticGoose preview win — brute force that only searched actions which changed the frame — and why it collapsed once the organisers added action-efficiency scoring and unseen games.Induction and transduction run through the middle of the conversation — how much of an answer is really priors leaking back the moment a model recognises a maze. The abstraction mountain, and Tim's case that LLMs reach the right answer through fractured, entangled representations — performance, not competence. Whether transformers plan at all or just fake it well enough. Why the score really measures action efficiency, not games solved, and why agents lock onto the wrong goal and cannot climb back out.Crouzier closes on the Tufa Labs thesis — a small lab against the giants, the bitter lesson against hand-built harnesses, and safety — and Tim ties it back to Kenneth Stanley, deep constraints, and creativity as competence.

Disclosure: Tufa Labs sponsors MLST. ---TIMESTAMPS:00:00:00 Meet the Tufa team and what makes ARC-AGI-3 hard00:02:11 Locksmith game: reading the rules from raw frames00:03:10 Why build an independent research lab00:04:11 StochasticGoose: a preview win, then the hardened games00:07:58 Induction, transduction, and priors inside LLMs00:10:31 Curiosity, world models, and exploring by frame change00:14:32 Understanding debt and losing sight of your own code00:15:53 Requirements-based agents and human-AI co-creativity00:19:22 Why auto-research misses the big picture00:21:54 The abstraction mountain and fractured representations00:27:36 Constraints and making LLMs act as if they understand00:34:51 Human difficulty calibration, esports priors, and emergence00:41:35 Agency, goal acquisition, and two kinds of planning00:47:31 Harnesses, the 36% number, and wrong-goal loops00:52:33 Rewards, goals, and why ARC-AGI-3 resists brute force01:00:46 Would solving ARC-AGI-3 prove AGI?01:07:53 Stripping language away, then priors leak back01:14:06 Representation and whether language is necessary01:18:04 The bitter lesson versus specialised harnesses01:22:20 Capability research, safety, and the software singularity---REFERENCES:organization:[00:02:11] ARC-AGI-3https://arcprize.org/arc-agi/3[00:03:10] Tufa Labshttps://tufalabs.ai/team/[00:04:20] ARC-AGI-3 Preview Agent Competitionhttps://arcprize.org/competitions/arc-agi-3-preview-agentstool:[00:04:55] StochasticGoose ARC-AGI-3 solutionhttps://github.com/DriesSmit/ARC3-solution[00:07:42] ArcGenticahttps://github.com/symbolica-ai/arcgentica[00:07:49] RGB-Agenthttps://github.com/alexisfox7/RGB-Agent[00:14:38] Claude Codehttps://www.anthropic.com/claude-code[01:03:42] Qwen 3.6 27Bhttps://huggingface.co/Qwen/Qwen3.6-27Bpaper:[00:13:03] On the Measure of Intelligencehttps://arxiv.org/abs/1911.01547[00:27:42] DreamCoderhttps://arxiv.org/abs/2006.08381[00:43:55] On the Biology of a Large Language Modelhttps://transformer-circuits.pub/2025/attribution-graphs/biology.html[01:18:46] ImageNet Classification with Deep CNNs (AlexNet)https://proceedings.neurips.cc/paper_files/paper/2012/file/c399862d3b9d6b76c8436e924a68c45b-Paper.pdfother:[01:18:16] The Bitter Lessonhttp://www.incompleteideas.net/IncIdeas/BitterLesson.html---https://app.rescript.info/share/463d7f031349b4b9db428553eed88230

Thomas Ahle wants Normal Computing to be the Lovable for chip design: type your intent, and a swarm of agents carries it from design through optimisation, formalisation and verification to tape-out. To get there, his team at wrote their own open-source Verilog simulator, 580,000 lines in 43 days, because commercial EDA verifiers run about $10,000 per core and there are no decent open-source compilers to build on.

That sets up the question Tim keeps pressing: if an agent can produce a chip design, a proof, or a working program, how do you actually know it is correct? Passing 70% of tests is not the same as being right, and a single fabricated bug can cost a company a fortune. They dig into ProgramBench (rebuild a program from its tests, roughly 0% success), the difference between structure and competence, and the "understanding debt" you take on when nobody reads the code.

From there: auto-formalisation in Lean and the AlphaProof trick of training on prove-or-disprove; why there is no single true representation of a spec (Petri nets, TLA+, Erik Curiel's "math does not represent"); and thermodynamic computing, where Normal Computing's CN101 chip is built so that its physical noise *is* the computation, settling a stochastic differential equation in hardware to invert a matrix. Plus Bayesian uncertainty, specialisation, the Chomsky hierarchy, AI slop, and whether performance is all that matters.

Recorded in Zurich.

Disclosure: Normal Computing paid our production and travel costs for this show. We retained full editorial control. They did not see the video before publication, and we did not show it to them or discuss it with them beforehand.

---

TIMESTAMPS:

00:00:00 Meet Thomas Ahle: the Lovable for chip design

00:03:41 Why hardware needs formal verification

00:06:36 Ten thousand dollars per core and a six-month agent run

00:07:40 Rebuilding programs from tests: ProgramBench and zero percent

00:12:15 Structure vs competence: can you learn a program from behavior?

00:15:27 Continual learning, abstraction, and Claude as an ecosystem

00:23:17 Autoformalization and the AlphaProof trick

00:29:31 No single true representation: specs, Petri nets and TLA+

00:34:43 Thermodynamic computing: when noise is the computation

00:37:32 Bayesian uncertainty in the age of token streams

00:41:12 Hybrid compute: vibe-coding loops, binaries and Stockfish

00:44:44 Co-design, central-AI apps and API pricing

00:49:45 Chain of thoughtlessness and the Chomsky hierarchy

00:53:40 AI psychosis, slop and the broken social contract

00:57:34 Typing it yourself, teamwork and performance vs competence

---

REFERENCES:

person:

[00:00:10] Thomas Ahle

https://thomasahle.com

organization:

[00:00:27] Normal Computing

https://normalcomputing.com/

paper:

[00:11:21] ProgramBench: Can Language Models Rebuild Programs From Scratch?

https://arxiv.org/abs/2605.03546

[00:31:55] Autoformalizing Memory Device Specifications with Agents

https://arxiv.org/abs/2605.00058

[00:35:20] Thermo AI and the Fluctuation Frontier

https://arxiv.org/abs/2302.06584

[00:36:40] Thermo Comp System for AI Applications

https://arxiv.org/abs/2312.04836

[00:37:05] Thermodynamic Linear Algebra

https://arxiv.org/abs/2308.05660

[00:44:50] An efficient probabilistic hardware architecture for diffusion-like models

https://arxiv.org/abs/2510.23972

tool:

other:

[00:01:00] Building an Open-Source Verilog Simulator with AI: 580K Lines in 43 Days

https://normalcomputing.com/blog/building-an-open-source-verilog-simulator-with-ai-580k-lines-in-43-days

[00:02:55] Normal Computing Announces Tape-Out of the World's First Thermodynamic Computing Chip (CN101)

https://www.normalcomputing.com/blog/normal-computing-announces-tape-out-of-worlds-first-thermodynamic-computing-chip

[00:32:02] DRAMBench: Autoformalizing DRAM Specifications with Timed Petri Nets

https://www.iese.fraunhofer.de/blog/drambench-autoformalizing-dram-specifications/

---

ReScript: https://app.rescript.info/share/ff9684a112ab37744096adaeb097a263