Episodit
-
Neel Nanda, a senior research scientist at Google DeepMind, leads their mechanistic interpretability team. In this extensive interview, he discusses his work trying to understand how neural networks function internally. At just 25 years old, Nanda has quickly become a prominent voice in AI research after completing his pure mathematics degree at Cambridge in 2020.
Nanda reckons that machine learning is unique because we create neural networks that can perform impressive tasks (like complex reasoning and software engineering) without understanding how they work internally. He compares this to having computer programs that can do things no human programmer knows how to write. His work focuses on "mechanistic interpretability" - attempting to uncover and understand the internal structures and algorithms that emerge within these networks.
SPONSOR MESSAGES:
***
CentML offers competitive pricing for GenAI model deployment, with flexible options to suit a wide range of models, from small to large-scale deployments.
https://centml.ai/pricing/
Tufa AI Labs is a brand new research lab in Zurich started by Benjamin Crouzier focussed on ARC and AGI, they just acquired MindsAI - the current winners of the ARC challenge. Are you interested in working on ARC, or getting involved in their events? Goto https://tufalabs.ai/
***
SHOWNOTES, TRANSCRIPT, ALL REFERENCES (DONT MISS!):
https://www.dropbox.com/scl/fi/36dvtfl3v3p56hbi30im7/NeelShow.pdf?rlkey=pq8t7lyv2z60knlifyy17jdtx&st=kiutudhc&dl=0
We riff on:
* How neural networks develop meaningful internal representations beyond simple pattern matching
* The effectiveness of chain-of-thought prompting and why it improves model performance
* The importance of hands-on coding over extensive paper reading for new researchers
* His journey from Cambridge to working with Chris Olah at Anthropic and eventually Google DeepMind
* The role of mechanistic interpretability in AI safety
NEEL NANDA:
https://www.neelnanda.io/
https://scholar.google.com/citations?user=GLnX3MkAAAAJ&hl=en
https://x.com/NeelNanda5
Interviewer - Tim Scarfe
TOC:
1. Part 1: Introduction
[00:00:00] 1.1 Introduction and Core Concepts Overview
2. Part 2: Outside Interview
[00:06:45] 2.1 Mechanistic Interpretability Foundations
3. Part 3: Main Interview
[00:32:52] 3.1 Mechanistic Interpretability
4. Neural Architecture and Circuits
[01:00:31] 4.1 Biological Evolution Parallels
[01:04:03] 4.2 Universal Circuit Patterns and Induction Heads
[01:11:07] 4.3 Entity Detection and Knowledge Boundaries
[01:14:26] 4.4 Mechanistic Interpretability and Activation Patching
5. Model Behavior Analysis
[01:30:00] 5.1 Golden Gate Claude Experiment and Feature Amplification
[01:33:27] 5.2 Model Personas and RLHF Behavior Modification
[01:36:28] 5.3 Steering Vectors and Linear Representations
[01:40:00] 5.4 Hallucinations and Model Uncertainty
6. Sparse Autoencoder Architecture
[01:44:54] 6.1 Architecture and Mathematical Foundations
[02:22:03] 6.2 Core Challenges and Solutions
[02:32:04] 6.3 Advanced Activation Functions and Top-k Implementations
[02:34:41] 6.4 Research Applications in Transformer Circuit Analysis
7. Feature Learning and Scaling
[02:48:02] 7.1 Autoencoder Feature Learning and Width Parameters
[03:02:46] 7.2 Scaling Laws and Training Stability
[03:11:00] 7.3 Feature Identification and Bias Correction
[03:19:52] 7.4 Training Dynamics Analysis Methods
8. Engineering Implementation
[03:23:48] 8.1 Scale and Infrastructure Requirements
[03:25:20] 8.2 Computational Requirements and Storage
[03:35:22] 8.3 Chain-of-Thought Reasoning Implementation
[03:37:15] 8.4 Latent Structure Inference in Language Models
-
Jonas Hübotter, PhD student at ETH Zurich's Institute for Machine Learning, discusses his groundbreaking research on test-time computation and local learning. He demonstrates how smaller models can outperform larger ones by 30x through strategic test-time computation and introduces a novel paradigm combining inductive and transductive learning approaches.
Using Bayesian linear regression as a surrogate model for uncertainty estimation, Jonas explains how models can efficiently adapt to specific tasks without massive pre-training. He draws an analogy to Google Earth's variable resolution system to illustrate dynamic resource allocation based on task complexity.
The conversation explores the future of AI architecture, envisioning systems that continuously learn and adapt beyond current monolithic models. Jonas concludes by proposing hybrid deployment strategies combining local and cloud computation, suggesting a future where compute resources are allocated based on task complexity rather than fixed model size.
This research represents a significant shift in machine learning, prioritizing intelligent resource allocation and adaptive learning over traditional scaling approaches.
SPONSOR MESSAGES:
CentML offers competitive pricing for GenAI model deployment, with flexible options to suit a wide range of models, from small to large-scale deployments.
https://centml.ai/pricing/
Tufa AI Labs is a brand new research lab in Zurich started by Benjamin Crouzier focussed on ARC and AGI, they just acquired MindsAI - the current winners of the ARC challenge. Are you interested in working on ARC, or getting involved in their events? Goto https://tufalabs.ai/
Transcription, references and show notes PDF download:
https://www.dropbox.com/scl/fi/cxg80p388snwt6qbp4m52/JonasFinal.pdf?rlkey=glk9mhpzjvesanlc14rtpvk4r&st=6qwi8n3x&dl=0
Jonas Hübotter
https://jonhue.github.io/
https://scholar.google.com/citations?user=pxi_RkwAAAAJ
Transductive Active Learning: Theory and Applications (NeurIPS 2024)
https://arxiv.org/pdf/2402.15898
EFFICIENTLY LEARNING AT TEST-TIME: ACTIVE FINE-TUNING OF LLMS (SIFT)
https://arxiv.org/pdf/2410.08020
TOC:
1. Test-Time Computation Fundamentals
[00:00:00] Intro
[00:03:10] 1.1 Test-Time Computation and Model Performance Comparison
[00:05:52] 1.2 Retrieval Augmentation and Machine Teaching Strategies
[00:09:40] 1.3 In-Context Learning vs Fine-Tuning Trade-offs
2. System Architecture and Intelligence
[00:15:58] 2.1 System Architecture and Intelligence Emergence
[00:23:22] 2.2 Active Inference and Constrained Agency in AI
[00:29:52] 2.3 Evolution of Local Learning Methods
[00:32:05] 2.4 Vapnik's Contributions to Transductive Learning
3. Resource Optimization and Local Learning
[00:34:35] 3.1 Computational Resource Allocation in ML Models
[00:35:30] 3.2 Historical Context and Traditional ML Optimization
[00:37:55] 3.3 Variable Resolution Processing and Active Inference in ML
[00:43:01] 3.4 Local Learning and Base Model Capacity Trade-offs
[00:48:04] 3.5 Active Learning vs Local Learning Approaches
4. Information Retrieval and Model Interpretability
[00:51:08] 4.1 Information Retrieval and Nearest Neighbor Limitations
[01:03:07] 4.2 Model Interpretability and Surrogate Models
[01:15:03] 4.3 Bayesian Uncertainty Estimation and Surrogate Models
5. Distributed Systems and Deployment
[01:23:56] 5.1 Memory Architecture and Controller Systems
[01:28:14] 5.2 Evolution from Static to Distributed Learning Systems
[01:38:03] 5.3 Transductive Learning and Model Specialization
[01:41:58] 5.4 Hybrid Local-Cloud Deployment Strategies
-
Puuttuva jakso?
-
Professor Swarat Chaudhuri from the University of Texas at Austin and visiting researcher at Google DeepMind discusses breakthroughs in AI reasoning, theorem proving, and mathematical discovery. Chaudhuri explains his groundbreaking work on COPRA (a GPT-based prover agent), shares insights on neurosymbolic approaches to AI.
Professor Swarat Chaudhuri:
https://www.cs.utexas.edu/~swarat/
SPONSOR MESSAGES:
CentML offers competitive pricing for GenAI model deployment, with flexible options to suit a wide range of models, from small to large-scale deployments.
https://centml.ai/pricing/
Tufa AI Labs is a brand new research lab in Zurich started by Benjamin Crouzier focussed on ARC and AGI, they just acquired MindsAI - the current winners of the ARC challenge. Are you interested in working on ARC, or getting involved in their events? Goto https://tufalabs.ai/
TOC:
[00:00:00] 0. Introduction / CentML ad, Tufa ad
1. AI Reasoning: From Language Models to Neurosymbolic Approaches
[00:02:27] 1.1 Defining Reasoning in AI
[00:09:51] 1.2 Limitations of Current Language Models
[00:17:22] 1.3 Neuro-symbolic Approaches and Program Synthesis
[00:24:59] 1.4 COPRA and In-Context Learning for Theorem Proving
[00:34:39] 1.5 Symbolic Regression and LLM-Guided Abstraction
2. AI in Mathematics: Theorem Proving and Concept Discovery
[00:43:37] 2.1 AI-Assisted Theorem Proving and Proof Verification
[01:01:37] 2.2 Symbolic Regression and Concept Discovery in Mathematics
[01:11:57] 2.3 Scaling and Modularizing Mathematical Proofs
[01:21:53] 2.4 COPRA: In-Context Learning for Formal Theorem-Proving
[01:28:22] 2.5 AI-driven theorem proving and mathematical discovery
3. Formal Methods and Challenges in AI Mathematics
[01:30:42] 3.1 Formal proofs, empirical predicates, and uncertainty in AI mathematics
[01:34:01] 3.2 Characteristics of good theoretical computer science research
[01:39:16] 3.3 LLMs in theorem generation and proving
[01:42:21] 3.4 Addressing contamination and concept learning in AI systems
REFS:
00:04:58 The Chinese Room Argument, https://plato.stanford.edu/entries/chinese-room/
00:11:42 Software 2.0, https://medium.com/@karpathy/software-2-0-a64152b37c35
00:11:57 Solving Olympiad Geometry Without Human Demonstrations, https://www.nature.com/articles/s41586-023-06747-5
00:13:26 Lean, https://lean-lang.org/
00:15:43 A General Reinforcement Learning Algorithm That Masters Chess, Shogi, and Go Through Self-Play, https://www.science.org/doi/10.1126/science.aar6404
00:19:24 DreamCoder (Ellis et al., PLDI 2021), https://arxiv.org/abs/2006.08381
00:24:37 The Lambda Calculus, https://plato.stanford.edu/entries/lambda-calculus/
00:26:43 Neural Sketch Learning for Conditional Program Generation, https://arxiv.org/pdf/1703.05698
00:28:08 Learning Differentiable Programs With Admissible Neural Heuristics, https://arxiv.org/abs/2007.12101
00:31:03 Symbolic Regression With a Learned Concept Library (Grayeli et al., NeurIPS 2024), https://arxiv.org/abs/2409.09359
00:41:30 Formal Verification of Parallel Programs, https://dl.acm.org/doi/10.1145/360248.360251
01:00:37 Training Compute-Optimal Large Language Models, https://arxiv.org/abs/2203.15556
01:18:19 Chain-of-Thought Prompting Elicits Reasoning in Large Language Models, https://arxiv.org/abs/2201.11903
01:18:42 Draft, Sketch, and Prove: Guiding Formal Theorem Provers With Informal Proofs, https://arxiv.org/abs/2210.12283
01:19:49 Learning Formal Mathematics From Intrinsic Motivation, https://arxiv.org/pdf/2407.00695
01:20:19 An In-Context Learning Agent for Formal Theorem-Proving (Thakur et al., CoLM 2024), https://arxiv.org/pdf/2310.04353
01:23:58 Learning to Prove Theorems via Interacting With Proof Assistants, https://arxiv.org/abs/1905.09381
01:39:58 An In-Context Learning Agent for Formal Theorem-Proving (Thakur et al., CoLM 2024), https://arxiv.org/pdf/2310.04353
01:42:24 Programmatically Interpretable Reinforcement Learning (Verma et al., ICML 2018), https://arxiv.org/abs/1804.02477
-
Nora Belrose, Head of Interpretability Research at EleutherAI, discusses critical challenges in AI safety and development. The conversation begins with her technical work on concept erasure in neural networks through LEACE (LEAst-squares Concept Erasure), while highlighting how neural networks' progression from simple to complex learning patterns could have important implications for AI safety.
Many fear that advanced AI will pose an existential threat -- pursuing its own dangerous goals once it's powerful enough. But Belrose challenges this popular doomsday scenario with a fascinating breakdown of why it doesn't add up.
Belrose also provides a detailed critique of current AI alignment approaches, particularly examining "counting arguments" and their limitations when applied to AI safety. She argues that the Principle of Indifference may be insufficient for addressing existential risks from advanced AI systems. The discussion explores how emergent properties in complex AI systems could lead to unpredictable and potentially dangerous behaviors that simple reductionist approaches fail to capture.
The conversation concludes by exploring broader philosophical territory, where Belrose discusses her growing interest in Buddhism's potential relevance to a post-automation future. She connects concepts of moral anti-realism with Buddhist ideas about emptiness and non-attachment, suggesting these frameworks might help humans find meaning in a world where AI handles most practical tasks. Rather than viewing this automated future with alarm, she proposes that Zen Buddhism's emphasis on spontaneity and presence might complement a society freed from traditional labor.
SPONSOR MESSAGES:
CentML offers competitive pricing for GenAI model deployment, with flexible options to suit a wide range of models, from small to large-scale deployments.
https://centml.ai/pricing/
Tufa AI Labs is a brand new research lab in Zurich started by Benjamin Crouzier focussed on ARC and AGI, they just acquired MindsAI - the current winners of the ARC challenge. Are you interested in working on ARC, or getting involved in their events? Goto https://tufalabs.ai/
Nora Belrose:
https://norabelrose.com/
https://scholar.google.com/citations?user=p_oBc64AAAAJ&hl=en
https://x.com/norabelrose
SHOWNOTES:
https://www.dropbox.com/scl/fi/38fhsv2zh8gnubtjaoq4a/NORA_FINAL.pdf?rlkey=0e5r8rd261821g1em4dgv0k70&st=t5c9ckfb&dl=0
TOC:
1. Neural Network Foundations
[00:00:00] 1.1 Philosophical Foundations and Neural Network Simplicity Bias
[00:02:20] 1.2 LEACE and Concept Erasure Fundamentals
[00:13:16] 1.3 LISA Technical Implementation and Applications
[00:18:50] 1.4 Practical Implementation Challenges and Data Requirements
[00:22:13] 1.5 Performance Impact and Limitations of Concept Erasure
2. Machine Learning Theory
[00:32:23] 2.1 Neural Network Learning Progression and Simplicity Bias
[00:37:10] 2.2 Optimal Transport Theory and Image Statistics Manipulation
[00:43:05] 2.3 Grokking Phenomena and Training Dynamics
[00:44:50] 2.4 Texture vs Shape Bias in Computer Vision Models
[00:45:15] 2.5 CNN Architecture and Shape Recognition Limitations
3. AI Systems and Value Learning
[00:47:10] 3.1 Meaning, Value, and Consciousness in AI Systems
[00:53:06] 3.2 Global Connectivity vs Local Culture Preservation
[00:58:18] 3.3 AI Capabilities and Future Development Trajectory
4. Consciousness Theory
[01:03:03] 4.1 4E Cognition and Extended Mind Theory
[01:09:40] 4.2 Thompson's Views on Consciousness and Simulation
[01:12:46] 4.3 Phenomenology and Consciousness Theory
[01:15:43] 4.4 Critique of Illusionism and Embodied Experience
[01:23:16] 4.5 AI Alignment and Counting Arguments Debate
(TRUNCATED, TOC embedded in MP3 file with more information)
-
Prof. Gennady Pekhimenko (CEO of CentML, UofT) joins us in this *sponsored episode* to dive deep into AI system optimization and enterprise implementation. From NVIDIA's technical leadership model to the rise of open-source AI, Pekhimenko shares insights on bridging the gap between academic research and industrial applications. Learn about "dark silicon," GPU utilization challenges in ML workloads, and how modern enterprises can optimize their AI infrastructure. The conversation explores why some companies achieve only 10% GPU efficiency and practical solutions for improving AI system performance. A must-watch for anyone interested in the technical foundations of enterprise AI and hardware optimization.
CentML offers competitive pricing for GenAI model deployment, with flexible options to suit a wide range of models, from small to large-scale deployments. Cheaper, faster, no commitments, pay as you go, scale massively, simple to setup. Check it out!
https://centml.ai/pricing/
SPONSOR MESSAGES:
MLST is also sponsored by Tufa AI Labs - https://tufalabs.ai/
They are hiring cracked ML engineers/researchers to work on ARC and build AGI!
SHOWNOTES (diarised transcript, TOC, references, summary, best quotes etc)
https://www.dropbox.com/scl/fi/w9kbpso7fawtm286kkp6j/Gennady.pdf?rlkey=aqjqmncx3kjnatk2il1gbgknk&st=2a9mccj8&dl=0
TOC:
1. AI Strategy and Leadership
[00:00:00] 1.1 Technical Leadership and Corporate Structure
[00:09:55] 1.2 Open Source vs Proprietary AI Models
[00:16:04] 1.3 Hardware and System Architecture Challenges
[00:23:37] 1.4 Enterprise AI Implementation and Optimization
[00:35:30] 1.5 AI Reasoning Capabilities and Limitations
2. AI System Development
[00:38:45] 2.1 Computational and Cognitive Limitations of AI Systems
[00:42:40] 2.2 Human-LLM Communication Adaptation and Patterns
[00:46:18] 2.3 AI-Assisted Software Development Challenges
[00:47:55] 2.4 Future of Software Engineering Careers in AI Era
[00:49:49] 2.5 Enterprise AI Adoption Challenges and Implementation
3. ML Infrastructure Optimization
[00:54:41] 3.1 MLOps Evolution and Platform Centralization
[00:55:43] 3.2 Hardware Optimization and Performance Constraints
[01:05:24] 3.3 ML Compiler Optimization and Python Performance
[01:15:57] 3.4 Enterprise ML Deployment and Cloud Provider Partnerships
4. Distributed AI Architecture
[01:27:05] 4.1 Multi-Cloud ML Infrastructure and Optimization
[01:29:45] 4.2 AI Agent Systems and Production Readiness
[01:32:00] 4.3 RAG Implementation and Fine-Tuning Considerations
[01:33:45] 4.4 Distributed AI Systems Architecture and Ray Framework
5. AI Industry Standards and Research
[01:37:55] 5.1 Origins and Evolution of MLPerf Benchmarking
[01:43:15] 5.2 MLPerf Methodology and Industry Impact
[01:50:17] 5.3 Academic Research vs Industry Implementation in AI
[01:58:59] 5.4 AI Research History and Safety Concerns
-
Eliezer Yudkowsky and Stephen Wolfram discuss artificial intelligence and its potential existen‑
tial risks. They traversed fundamental questions about AI safety, consciousness, computational irreducibility, and the nature of intelligence.
The discourse centered on Yudkowsky’s argument that advanced AI systems pose an existential threat to humanity, primarily due to the challenge of alignment and the potential for emergent goals that diverge from human values. Wolfram, while acknowledging potential risks, approached the topic from a his signature measured perspective, emphasizing the importance of understanding computational systems’ fundamental nature and questioning whether AI systems would necessarily develop the kind of goal‑directed behavior Yudkowsky fears.
***
MLST IS SPONSORED BY TUFA AI LABS!
The current winners of the ARC challenge, MindsAI are part of Tufa AI Labs. They are hiring ML engineers. Are you interested?! Please goto https://tufalabs.ai/
***
TOC:
1. Foundational AI Concepts and Risks
[00:00:01] 1.1 AI Optimization and System Capabilities Debate
[00:06:46] 1.2 Computational Irreducibility and Intelligence Limitations
[00:20:09] 1.3 Existential Risk and Species Succession
[00:23:28] 1.4 Consciousness and Value Preservation in AI Systems
2. Ethics and Philosophy in AI
[00:33:24] 2.1 Moral Value of Human Consciousness vs. Computation
[00:36:30] 2.2 Ethics and Moral Philosophy Debate
[00:39:58] 2.3 Existential Risks and Digital Immortality
[00:43:30] 2.4 Consciousness and Personal Identity in Brain Emulation
3. Truth and Logic in AI Systems
[00:54:39] 3.1 AI Persuasion Ethics and Truth
[01:01:48] 3.2 Mathematical Truth and Logic in AI Systems
[01:11:29] 3.3 Universal Truth vs Personal Interpretation in Ethics and Mathematics
[01:14:43] 3.4 Quantum Mechanics and Fundamental Reality Debate
4. AI Capabilities and Constraints
[01:21:21] 4.1 AI Perception and Physical Laws
[01:28:33] 4.2 AI Capabilities and Computational Constraints
[01:34:59] 4.3 AI Motivation and Anthropomorphization Debate
[01:38:09] 4.4 Prediction vs Agency in AI Systems
5. AI System Architecture and Behavior
[01:44:47] 5.1 Computational Irreducibility and Probabilistic Prediction
[01:48:10] 5.2 Teleological vs Mechanistic Explanations of AI Behavior
[02:09:41] 5.3 Machine Learning as Assembly of Computational Components
[02:29:52] 5.4 AI Safety and Predictability in Complex Systems
6. Goal Optimization and Alignment
[02:50:30] 6.1 Goal Specification and Optimization Challenges in AI Systems
[02:58:31] 6.2 Intelligence, Computation, and Goal-Directed Behavior
[03:02:18] 6.3 Optimization Goals and Human Existential Risk
[03:08:49] 6.4 Emergent Goals and AI Alignment Challenges
7. AI Evolution and Risk Assessment
[03:19:44] 7.1 Inner Optimization and Mesa-Optimization Theory
[03:34:00] 7.2 Dynamic AI Goals and Extinction Risk Debate
[03:56:05] 7.3 AI Risk and Biological System Analogies
[04:09:37] 7.4 Expert Risk Assessments and Optimism vs Reality
8. Future Implications and Economics
[04:13:01] 8.1 Economic and Proliferation Considerations
SHOWNOTES (transcription, references, summary, best quotes etc):
https://www.dropbox.com/scl/fi/3st8dts2ba7yob161dchd/EliezerWolfram.pdf?rlkey=b6va5j8upgqwl9s2muc924vtt&st=vemwqx7a&dl=0
-
Francois Chollet, a prominent AI expert and creator of ARC-AGI, discusses intelligence, consciousness, and artificial intelligence.
Chollet explains that real intelligence isn't about memorizing information or having lots of knowledge - it's about being able to handle new situations effectively. This is why he believes current large language models (LLMs) have "near-zero intelligence" despite their impressive abilities. They're more like sophisticated memory and pattern-matching systems than truly intelligent beings.
***
MLST IS SPONSORED BY TUFA AI LABS!
The current winners of the ARC challenge, MindsAI are part of Tufa AI Labs. They are hiring ML engineers. Are you interested?! Please goto https://tufalabs.ai/
***
He introduced his "Kaleidoscope Hypothesis," which suggests that while the world seems infinitely complex, it's actually made up of simpler patterns that repeat and combine in different ways. True intelligence, he argues, involves identifying these basic patterns and using them to understand new situations.
Chollet also talked about consciousness, suggesting it develops gradually in children rather than appearing all at once. He believes consciousness exists in degrees - animals have it to some extent, and even human consciousness varies with age and circumstances (like being more conscious when learning something new versus doing routine tasks).
On AI safety, Chollet takes a notably different stance from many in Silicon Valley. He views AGI development as a scientific challenge rather than a religious quest, and doesn't share the apocalyptic concerns of some AI researchers. He argues that intelligence itself isn't dangerous - it's just a tool for turning information into useful models. What matters is how we choose to use it.
ARC-AGI Prize:
https://arcprize.org/
Francois Chollet:
https://x.com/fchollet
Shownotes:
https://www.dropbox.com/scl/fi/j2068j3hlj8br96pfa7bi/CHOLLET_FINAL.pdf?rlkey=xkbr7tbnrjdl66m246w26uc8k&st=0a4ec4na&dl=0
TOC:
1. Intelligence and Model Building
[00:00:00] 1.1 Intelligence Definition and ARC Benchmark
[00:05:40] 1.2 LLMs as Program Memorization Systems
[00:09:36] 1.3 Kaleidoscope Hypothesis and Abstract Building Blocks
[00:13:39] 1.4 Deep Learning Limitations and System 2 Reasoning
[00:29:38] 1.5 Intelligence vs. Skill in LLMs and Model Building
2. ARC Benchmark and Program Synthesis
[00:37:36] 2.1 Intelligence Definition and LLM Limitations
[00:41:33] 2.2 Meta-Learning System Architecture
[00:56:21] 2.3 Program Search and Occam's Razor
[00:59:42] 2.4 Developer-Aware Generalization
[01:06:49] 2.5 Task Generation and Benchmark Design
3. Cognitive Systems and Program Generation
[01:14:38] 3.1 System 1/2 Thinking Fundamentals
[01:22:17] 3.2 Program Synthesis and Combinatorial Challenges
[01:31:18] 3.3 Test-Time Fine-Tuning Strategies
[01:36:10] 3.4 Evaluation and Leakage Problems
[01:43:22] 3.5 ARC Implementation Approaches
4. Intelligence and Language Systems
[01:50:06] 4.1 Intelligence as Tool vs Agent
[01:53:53] 4.2 Cultural Knowledge Integration
[01:58:42] 4.3 Language and Abstraction Generation
[02:02:41] 4.4 Embodiment in Cognitive Systems
[02:09:02] 4.5 Language as Cognitive Operating System
5. Consciousness and AI Safety
[02:14:05] 5.1 Consciousness and Intelligence Relationship
[02:20:25] 5.2 Development of Machine Consciousness
[02:28:40] 5.3 Consciousness Prerequisites and Indicators
[02:36:36] 5.4 AGI Safety Considerations
[02:40:29] 5.5 AI Regulation Framework
-
Anil Ananthaswamy is an award-winning science writer and former staff writer and deputy news editor for the London-based New Scientist magazine.
Machine learning systems are making life-altering decisions for us: approving mortgage loans, determining whether a tumor is cancerous, or deciding if someone gets bail. They now influence developments and discoveries in chemistry, biology, and physics—the study of genomes, extrasolar planets, even the intricacies of quantum systems. And all this before large language models such as ChatGPT came on the scene.
We are living through a revolution in machine learning-powered AI that shows no signs of slowing down. This technology is based on relatively simple mathematical ideas, some of which go back centuries, including linear algebra and calculus, the stuff of seventeenth- and eighteenth-century mathematics. It took the birth and advancement of computer science and the kindling of 1990s computer chips designed for video games to ignite the explosion of AI that we see today. In this enlightening book, Anil Ananthaswamy explains the fundamental math behind machine learning, while suggesting intriguing links between artificial and natural intelligence. Might the same math underpin them both?
As Ananthaswamy resonantly concludes, to make safe and effective use of artificial intelligence, we need to understand its profound capabilities and limitations, the clues to which lie in the math that makes machine learning possible.
Why Machines Learn: The Elegant Math Behind Modern AI:
https://amzn.to/3UAWX3D
https://anilananthaswamy.com/
Sponsor message:
DO YOU WANT WORK ON ARC with the MindsAI team (current ARC winners)?
Interested? Apply for an ML research position: [email protected]
Shownotes:
https://www.dropbox.com/scl/fi/wpv22m5jxyiqr6pqfkzwz/anil.pdf?rlkey=9c233jo5armr548ctwo419n6p&st=xzhahtje&dl=0
Chapters:
1. ML Fundamentals and Prerequisites
[00:00:00] 1.1 Differences Between Human and Machine Learning
[00:00:35] 1.2 Mathematical Prerequisites and Societal Impact of ML
[00:02:20] 1.3 Author's Journey and Book Background
[00:11:30] 1.4 Mathematical Foundations and Core ML Concepts
[00:21:45] 1.5 Bias-Variance Tradeoff and Modern Deep Learning
2. Deep Learning Architecture
[00:29:05] 2.1 Double Descent and Overparameterization in Deep Learning
[00:32:40] 2.2 Mathematical Foundations and Self-Supervised Learning
[00:40:05] 2.3 High-Dimensional Spaces and Model Architecture
[00:52:55] 2.4 Historical Development of Backpropagation
3. AI Understanding and Limitations
[00:59:13] 3.1 Pattern Matching vs Human Reasoning in ML Models
[01:00:20] 3.2 Mathematical Foundations and Pattern Recognition in AI
[01:04:08] 3.3 LLM Reliability and Machine Understanding Debate
[01:12:50] 3.4 Historical Development of Deep Learning Technologies
[01:15:21] 3.5 Alternative AI Approaches and Bio-inspired Methods
4. Ethical and Neurological Perspectives
[01:24:32] 4.1 Neural Network Scaling and Mathematical Limitations
[01:31:12] 4.2 AI Ethics and Societal Impact
[01:38:30] 4.3 Consciousness and Neurological Conditions
[01:46:17] 4.4 Body Ownership and Agency in Neuroscience
-
Professor Michael Levin explores the revolutionary concept of diverse intelligence, demonstrating how cognitive capabilities extend far beyond traditional brain-based intelligence. Drawing from his groundbreaking research, he explains how even simple biological systems like gene regulatory networks exhibit learning, memory, and problem-solving abilities. Levin introduces key concepts like "cognitive light cones" - the scope of goals a system can pursue - and shows how these ideas are transforming our approach to cancer treatment and biological engineering. His insights challenge conventional views of intelligence and agency, with profound implications for both medicine and artificial intelligence development. This deep discussion reveals how understanding intelligence as a spectrum, from molecular networks to human minds, could be crucial for humanity's future technological development. Contains technical discussion of biological systems, cybernetics, and theoretical frameworks for understanding emergent cognition.
Prof. Michael Levin
https://as.tufts.edu/biology/people/faculty/michael-levin
https://x.com/drmichaellevin
Sponsor message:
DO YOU WANT WORK ON ARC with the MindsAI team (current ARC winners)?
Interested? Apply for an ML research position: [email protected]
TOC
1. Intelligence Fundamentals and Evolution
[00:00:00] 1.1 Future Evolution of Human Intelligence and Consciousness
[00:03:00] 1.2 Science Fiction's Role in Exploring Intelligence Possibilities
[00:08:15] 1.3 Essential Characteristics of Human-Level Intelligence and Relationships
[00:14:20] 1.4 Biological Systems Architecture and Intelligence
2. Biological Computing and Cognition
[00:24:00] 2.1 Agency and Intelligence in Biological Systems
[00:30:30] 2.2 Learning Capabilities in Gene Regulatory Networks
[00:35:37] 2.3 Biological Control Systems and Competency Architecture
[00:39:58] 2.4 Scientific Metaphors and Polycomputing Paradigm
3. Systems and Collective Intelligence
[00:43:26] 3.1 Embodiment and Problem-Solving Spaces
[00:44:50] 3.2 Perception-Action Loops and Biological Intelligence
[00:46:55] 3.3 Intelligence, Wisdom and Collective Systems
[00:53:07] 3.4 Cancer and Cognitive Light Cones
[00:57:09] 3.5 Emergent Intelligence and AI Agency
Shownotes:
https://www.dropbox.com/scl/fi/i2vl1vs009thg54lxx5wc/LEVIN.pdf?rlkey=dtk8okhbsejryiu2vrht19qp6&st=uzi0vo45&dl=0
REFS:
[0:05:30] A Fire Upon the Deep - Vernor Vinge sci-fi novel on AI and consciousness
[0:05:35] Maria Chudnovsky - MacArthur Fellow, Princeton mathematician, graph theory expert
[0:14:20] Bow-tie architecture in biological systems - Network structure research by Csete & Doyle
[0:15:40] Richard Watson - Southampton Professor, evolution and learning systems expert
[0:17:00] Levin paper on human issues in AI and evolution
[0:19:00] Bow-tie architecture in Darwin's agential materialism - Levin
[0:22:55] Philip Goff - Work on panpsychism and consciousness in Galileo's Error
[0:23:30] Strange Loop - Hofstadter's work on self-reference and consciousness
[0:25:00] The Hard Problem of Consciousness - Van Gulick
[0:26:15] Daniel Dennett - Theories on consciousness and intentional systems
[0:29:35] Principle of Least Action - Light path selection in physics
[0:29:50] Free Energy Principle - Friston's unified behavioral framework
[0:30:35] Gene regulatory networks - Learning capabilities in biological systems
[0:36:55] Minimal networks with learning capacity - Levin
[0:38:50] Multi-scale competency in biological systems - Levin
[0:41:40] Polycomputing paradigm - Biological computation by Bongard & Levin
[0:45:40] Collective intelligence in biology - Levin et al.
[0:46:55] Niche construction and stigmergy - Torday
[0:53:50] Tasmanian Devil Facial Tumor Disease - Transmissible cancer research
[0:55:05] Cognitive light cone - Computational boundaries of self - Levin
[0:58:05] Cognitive properties in sorting algorithms - Zhang, Goldstein & Levin
-
Will Williams is CTO of Speechmatics in Cambridge. In this sponsored episode - he shares deep technical insights into modern speech recognition technology and system architecture. The episode covers several key technical areas:
* Speechmatics' hybrid approach to ASR, which focusses on unsupervised learning methods, achieving comparable results with 100x less data than fully supervised approaches. Williams explains why this is more efficient and generalizable than end-to-end models like Whisper.
* Their production architecture implementing multiple operating points for different latency-accuracy trade-offs, with careful latency padding (up to 1.8 seconds) to ensure consistent user experience. The system uses lattice-based decoding with language model integration for improved accuracy.
* The challenges and solutions in real-time ASR, including their approach to diarization (speaker identification), handling cross-talk, and implicit source separation. Williams explains why these problems remain difficult even with modern deep learning approaches.
* Their testing and deployment infrastructure, including the use of mirrored environments for catching edge cases in production, and their strategy of maintaining global models rather than allowing customer-specific fine-tuning.
* Technical evolution in ASR, from early days of custom CUDA kernels and manual memory management to modern frameworks, with Williams offering interesting critiques of current PyTorch memory management approaches and arguing for more efficient direct memory allocation in production systems.
Get coding with their API! This is their URL:
https://www.speechmatics.com/
DO YOU WANT WORK ON ARC with the MindsAI team (current ARC winners)?
MLST is sponsored by Tufa Labs:
Focus: ARC, LLMs, test-time-compute, active inference, system2 reasoning, and more.
Interested? Apply for an ML research position: [email protected]
TOC
1. ASR Core Technology & Real-time Architecture
[00:00:00] 1.1 ASR and Diarization Fundamentals
[00:05:25] 1.2 Real-time Conversational AI Architecture
[00:09:21] 1.3 Neural Network Streaming Implementation
[00:12:49] 1.4 Multi-modal System Integration
2. Production System Optimization
[00:29:38] 2.1 Production Deployment and Testing Infrastructure
[00:35:40] 2.2 Model Architecture and Deployment Strategy
[00:37:12] 2.3 Latency-Accuracy Trade-offs
[00:39:15] 2.4 Language Model Integration
[00:40:32] 2.5 Lattice-based Decoding Architecture
3. Performance Evaluation & Ethical Considerations
[00:44:00] 3.1 ASR Performance Metrics and Capabilities
[00:46:35] 3.2 AI Regulation and Evaluation Methods
[00:51:09] 3.3 Benchmark and Testing Challenges
[00:54:30] 3.4 Real-world Implementation Metrics
[01:00:51] 3.5 Ethics and Privacy Considerations
4. ASR Technical Evolution
[01:09:00] 4.1 WER Calculation and Evaluation Methodologies
[01:10:21] 4.2 Supervised vs Self-Supervised Learning Approaches
[01:21:02] 4.3 Temporal Learning and Feature Processing
[01:24:45] 4.4 Feature Engineering to Automated ML
5. Enterprise Implementation & Scale
[01:27:55] 5.1 Future AI Systems and Adaptation
[01:31:52] 5.2 Technical Foundations and History
[01:34:53] 5.3 Infrastructure and Team Scaling
[01:38:05] 5.4 Research and Talent Strategy
[01:41:11] 5.5 Engineering Practice Evolution
Shownotes:
https://www.dropbox.com/scl/fi/d94b1jcgph9o8au8shdym/Speechmatics.pdf?rlkey=bi55wvktzomzx0y5sic6jz99y&st=6qwofv8t&dl=0
-
Dr. Sanjeev Namjoshi, a machine learning engineer who recently submitted a book on Active Inference to MIT Press, discusses the theoretical foundations and practical applications of Active Inference, the Free Energy Principle (FEP), and Bayesian mechanics. He explains how these frameworks describe how biological and artificial systems maintain stability by minimizing uncertainty about their environment.
DO YOU WANT WORK ON ARC with the MindsAI team (current ARC winners)?
MLST is sponsored by Tufa Labs:
Focus: ARC, LLMs, test-time-compute, active inference, system2 reasoning, and more.
Future plans: Expanding to complex environments like Warcraft 2 and Starcraft 2.
Interested? Apply for an ML research position: [email protected]
Namjoshi traces the evolution of these fields from early 2000s neuroscience research to current developments, highlighting how Active Inference provides a unified framework for perception and action through variational free energy minimization. He contrasts this with traditional machine learning approaches, emphasizing Active Inference's natural capacity for exploration and curiosity through epistemic value.
He sees Active Inference as being at a similar stage to deep learning in the early 2000s - poised for significant breakthroughs but requiring better tools and wider adoption. While acknowledging current computational challenges, he emphasizes Active Inference's potential advantages over reinforcement learning, particularly its principled approach to exploration and planning.
Dr. Sanjeev Namjoshi
https://snamjoshi.github.io/
TOC:
1. Theoretical Foundations: AI Agency and Sentience
[00:00:00] 1.1 Intro
[00:02:45] 1.2 Free Energy Principle and Active Inference Theory
[00:11:16] 1.3 Emergence and Self-Organization in Complex Systems
[00:19:11] 1.4 Agency and Representation in AI Systems
[00:29:59] 1.5 Bayesian Mechanics and Systems Modeling
2. Technical Framework: Active Inference and Free Energy
[00:38:37] 2.1 Generative Processes and Agent-Environment Modeling
[00:42:27] 2.2 Markov Blankets and System Boundaries
[00:44:30] 2.3 Bayesian Inference and Prior Distributions
[00:52:41] 2.4 Variational Free Energy Minimization Framework
[00:55:07] 2.5 VFE Optimization Techniques: Generalized Filtering vs DEM
3. Implementation and Optimization Methods
[00:58:25] 3.1 Information Theory and Free Energy Concepts
[01:05:25] 3.2 Surprise Minimization and Action in Active Inference
[01:15:58] 3.3 Evolution of Active Inference Models: Continuous to Discrete Approaches
[01:26:00] 3.4 Uncertainty Reduction and Control Systems in Active Inference
4. Safety and Regulatory Frameworks
[01:32:40] 4.1 Historical Evolution of Risk Management and Predictive Systems
[01:36:12] 4.2 Agency and Reality: Philosophical Perspectives on Models
[01:39:20] 4.3 Limitations of Symbolic AI and Current System Design
[01:46:40] 4.4 AI Safety Regulation and Corporate Governance
5. Socioeconomic Integration and Modeling
[01:52:55] 5.1 Economic Policy and Public Sentiment Modeling
[01:55:21] 5.2 Free Energy Principle: Libertarian vs Collectivist Perspectives
[01:58:53] 5.3 Regulation of Complex Socio-Technical Systems
[02:03:04] 5.4 Evolution and Current State of Active Inference Research
6. Future Directions and Applications
[02:14:26] 6.1 Active Inference Applications and Future Development
[02:22:58] 6.2 Cultural Learning and Active Inference
[02:29:19] 6.3 Hierarchical Relationship Between FEP, Active Inference, and Bayesian Mechanics
[02:33:22] 6.4 Historical Evolution of Free Energy Principle
[02:38:52] 6.5 Active Inference vs Traditional Machine Learning Approaches
Transcript and shownotes with refs and URLs:
https://www.dropbox.com/scl/fi/qj22a660cob1795ej0gbw/SanjeevShow.pdf?rlkey=w323r3e8zfsnve22caayzb17k&st=el1fdgfr&dl=0
-
Dr. Joscha Bach discusses advanced AI, consciousness, and cognitive modeling. He presents consciousness as a virtual property emerging from self-organizing software patterns, challenging panpsychism and materialism. Bach introduces "Cyberanima," reinterpreting animism through information processing, viewing spirits as self-organizing software agents.
He addresses limitations of current large language models and advocates for smaller, more efficient AI models capable of reasoning from first principles. Bach describes his work with Liquid AI on novel neural network architectures for improved expressiveness and efficiency.
The interview covers AI's societal implications, including regulation challenges and impact on innovation. Bach argues for balancing oversight with technological progress, warning against overly restrictive regulations.
Throughout, Bach frames consciousness, intelligence, and agency as emergent properties of complex information processing systems, proposing a computational framework for cognitive phenomena and reality.
SPONSOR MESSAGE:
DO YOU WANT WORK ON ARC with the MindsAI team (current ARC winners)?MLST is sponsored by Tufa Labs:Focus: ARC, LLMs, test-time-compute, active inference, system2 reasoning, and more.Future plans: Expanding to complex environments like Warcraft 2 and Starcraft 2.Interested? Apply for an ML research position: [email protected]
TOC
[00:00:00] 1.1 Consciousness and Intelligence in AI Development
[00:07:44] 1.2 Agency, Intelligence, and Their Relationship to Physical Reality
[00:13:36] 1.3 Virtual Patterns and Causal Structures in Consciousness
[00:25:49] 1.4 Reinterpreting Concepts of God and Animism in Information Processing Terms
[00:32:50] 1.5 Animism and Evolution as Competition Between Software Agents
2. Self-Organizing Systems and Cognitive Models in AI
[00:37:59] 2.1 Consciousness as self-organizing software
[00:45:49] 2.2 Critique of panpsychism and alternative views on consciousness
[00:50:48] 2.3 Emergence of consciousness in complex systems
[00:52:50] 2.4 Neuronal motivation and the origins of consciousness
[00:56:47] 2.5 Coherence and Self-Organization in AI Systems
3. Advanced AI Architectures and Cognitive Processes
[00:57:50] 3.1 Second-Order Software and Complex Mental Processes
[01:01:05] 3.2 Collective Agency and Shared Values in AI
[01:05:40] 3.3 Limitations of Current AI Agents and LLMs
[01:06:40] 3.4 Liquid AI and Novel Neural Network Architectures
[01:10:06] 3.5 AI Model Efficiency and Future Directions
[01:19:00] 3.6 LLM Limitations and Internal State Representation
4. AI Regulation and Societal Impact
[01:31:23] 4.1 AI Regulation and Societal Impact
[01:49:50] 4.2 Open-Source AI and Industry Challenges
Refs in shownotes and MP3 metadata
Shownotes:
https://www.dropbox.com/scl/fi/g28dosz19bzcfs5imrvbu/JoschaInterview.pdf?rlkey=s3y18jy192ktz6ogd7qtvry3d&st=10z7q7w9&dl=0
-
Alessandro Palmarini is a post-baccalaureate researcher at the Santa Fe Institute working under the supervision of Melanie Mitchell. He completed his undergraduate degree in Artificial Intelligence and Computer Science at the University of Edinburgh. Palmarini's current research focuses on developing AI systems that can efficiently acquire new skills from limited data, inspired by François Chollet's work on measuring intelligence. His work builds upon the DreamCoder program synthesis system, introducing a novel approach called "dream decompiling" to improve library learning in inductive program synthesis. Palmarini is particularly interested in addressing the Abstraction and Reasoning Corpus (ARC) challenge, aiming to create AI systems that can perform abstract reasoning tasks more efficiently than current approaches. His research explores the balance between computational efficiency and data efficiency in AI learning processes.
DO YOU WANT WORK ON ARC with the MindsAI team (current ARC winners)?MLST is sponsored by Tufa Labs:Focus: ARC, LLMs, test-time-compute, active inference, system2 reasoning, and more.Future plans: Expanding to complex environments like Warcraft 2 and Starcraft 2.Interested? Apply for an ML research position: [email protected]
TOC:
1. Intelligence Measurement in AI Systems
[00:00:00] 1.1 Defining Intelligence in AI Systems
[00:02:00] 1.2 Research at Santa Fe Institute
[00:04:35] 1.3 Impact of Gaming on AI Development
[00:05:10] 1.4 Comparing AI and Human Learning Efficiency
2. Efficient Skill Acquisition in AI
[00:06:40] 2.1 Intelligence as Skill Acquisition Efficiency
[00:08:25] 2.2 Limitations of Current AI Systems in Generalization
[00:09:45] 2.3 Human vs. AI Cognitive Processes
[00:10:40] 2.4 Measuring AI Intelligence: Chollet's ARC Challenge
3. Program Synthesis and ARC Challenge
[00:12:55] 3.1 Philosophical Foundations of Program Synthesis
[00:17:14] 3.2 Introduction to Program Induction and ARC Tasks
[00:18:49] 3.3 DreamCoder: Principles and Techniques
[00:27:55] 3.4 Trade-offs in Program Synthesis Search Strategies
[00:31:52] 3.5 Neural Networks and Bayesian Program Learning
4. Advanced Program Synthesis Techniques
[00:32:30] 4.1 DreamCoder and Dream Decompiling Approach
[00:39:00] 4.2 Beta Distribution and Caching in Program Synthesis
[00:45:10] 4.3 Performance and Limitations of Dream Decompiling
[00:47:45] 4.4 Alessandro's Approach to ARC Challenge
[00:51:12] 4.5 Conclusion and Future Discussions
Refs:
Full reflist on YT VD, Show Notes and MP3 metadata
Show Notes: https://www.dropbox.com/scl/fi/x50201tgqucj5ba2q4typ/Ale.pdf?rlkey=0ubvk7p5gtyx1gpownpdadim8&st=5pniu3nq&dl=0
-
François Chollet discusses the limitations of Large Language Models (LLMs) and proposes a new approach to advancing artificial intelligence. He argues that current AI systems excel at pattern recognition but struggle with logical reasoning and true generalization.
This was Chollet's keynote talk at AGI-24, filmed in high-quality. We will be releasing a full interview with him shortly. A teaser clip from that is played in the intro!
Chollet introduces the Abstraction and Reasoning Corpus (ARC) as a benchmark for measuring AI progress towards human-like intelligence. He explains the concept of abstraction in AI systems and proposes combining deep learning with program synthesis to overcome current limitations. Chollet suggests that breakthroughs in AI might come from outside major tech labs and encourages researchers to explore new ideas in the pursuit of artificial general intelligence.
TOC
1. LLM Limitations and Intelligence Concepts
[00:00:00] 1.1 LLM Limitations and Composition
[00:12:05] 1.2 Intelligence as Process vs. Skill
[00:17:15] 1.3 Generalization as Key to AI Progress
2. ARC-AGI Benchmark and LLM Performance
[00:19:59] 2.1 Introduction to ARC-AGI Benchmark
[00:20:05] 2.2 Introduction to ARC-AGI and the ARC Prize
[00:23:35] 2.3 Performance of LLMs and Humans on ARC-AGI
3. Abstraction in AI Systems
[00:26:10] 3.1 The Kaleidoscope Hypothesis and Abstraction Spectrum
[00:30:05] 3.2 LLM Capabilities and Limitations in Abstraction
[00:32:10] 3.3 Value-Centric vs Program-Centric Abstraction
[00:33:25] 3.4 Types of Abstraction in AI Systems
4. Advancing AI: Combining Deep Learning and Program Synthesis
[00:34:05] 4.1 Limitations of Transformers and Need for Program Synthesis
[00:36:45] 4.2 Combining Deep Learning and Program Synthesis
[00:39:59] 4.3 Applying Combined Approaches to ARC Tasks
[00:44:20] 4.4 State-of-the-Art Solutions for ARC
Shownotes (new!): https://www.dropbox.com/scl/fi/i7nsyoahuei6np95lbjxw/CholletKeynote.pdf?rlkey=t3502kbov5exsdxhderq70b9i&st=1ca91ewz&dl=0
[0:01:15] Abstraction and Reasoning Corpus (ARC): AI benchmark (François Chollet)
https://arxiv.org/abs/1911.01547
[0:05:30] Monty Hall problem: Probability puzzle (Steve Selvin)
https://www.tandfonline.com/doi/abs/10.1080/00031305.1975.10479121
[0:06:20] LLM training dynamics analysis (Tirumala et al.)
https://arxiv.org/abs/2205.10770
[0:10:20] Transformer limitations on compositionality (Dziri et al.)
https://arxiv.org/abs/2305.18654
[0:10:25] Reversal Curse in LLMs (Berglund et al.)
https://arxiv.org/abs/2309.12288
[0:19:25] Measure of intelligence using algorithmic information theory (François Chollet)
https://arxiv.org/abs/1911.01547
[0:20:10] ARC-AGI: GitHub repository (François Chollet)
https://github.com/fchollet/ARC-AGI
[0:22:15] ARC Prize: $1,000,000+ competition (François Chollet)
https://arcprize.org/
[0:33:30] System 1 and System 2 thinking (Daniel Kahneman)
https://www.amazon.com/Thinking-Fast-Slow-Daniel-Kahneman/dp/0374533555
[0:34:00] Core knowledge in infants (Elizabeth Spelke)
https://www.harvardlds.org/wp-content/uploads/2017/01/SpelkeKinzler07-1.pdf
[0:34:30] Embedding interpretive spaces in ML (Tennenholtz et al.)
https://arxiv.org/abs/2310.04475
[0:44:20] Hypothesis Search with LLMs for ARC (Wang et al.)
https://arxiv.org/abs/2309.05660
[0:44:50] Ryan Greenblatt's high score on ARC public leaderboard
https://arcprize.org/
-
Ivan Zhang, co-founder of Cohere, discusses the company's enterprise-focused AI solutions. He explains Cohere's early emphasis on embedding technology and training models for secure environments.
Zhang highlights their implementation of Retrieval-Augmented Generation in healthcare, significantly reducing doctor preparation time. He explores the shift from monolithic AI models to heterogeneous systems and the importance of improving various AI system components. Zhang shares insights on using synthetic data to teach models reasoning, the democratization of software development through AI, and how his gaming skills transfer to running an AI company.
He advises young developers to fully embrace AI technologies and offers perspectives on AI reliability, potential risks, and future model architectures.
https://cohere.com/
https://ivanzhang.ca/
https://x.com/1vnzh
TOC:
00:00:00 Intro
00:03:20 AI & Language Model Evolution
00:06:09 Future AI Apps & Development
00:09:29 Impact on Software Dev Practices
00:13:03 Philosophical & Societal Implications
00:16:30 Compute Efficiency & RAG
00:20:39 Adoption Challenges & Solutions
00:22:30 GPU Optimization & Kubernetes Limits
00:24:16 Cohere's Implementation Approach
00:28:13 Gaming's Professional Influence
00:34:45 Transformer Optimizations
00:36:45 Future Models & System-Level Focus
00:39:20 Inference-Time Computation & Reasoning
00:42:05 Capturing Human Thought in AI
00:43:15 Research, Hiring & Developer Advice
REFS:
00:02:31 Cohere, https://cohere.com/
00:02:40 The Transformer architecture, https://arxiv.org/abs/1706.03762
00:03:22 The Innovator's Dilemma, https://www.amazon.com/Innovators-Dilemma-Technologies-Management-Innovation/dp/1633691780
00:09:15 The actor model, https://en.wikipedia.org/wiki/Actor_model
00:14:35 John Searle's Chinese Room Argument, https://plato.stanford.edu/entries/chinese-room/
00:18:00 Retrieval-Augmented Generation, https://arxiv.org/abs/2005.11401
00:18:40 Retrieval-Augmented Generation, https://docs.cohere.com/v2/docs/retrieval-augmented-generation-rag
00:35:39 Let’s Verify Step by Step, https://arxiv.org/pdf/2305.20050
00:39:20 Adaptive Inference-Time Compute, https://arxiv.org/abs/2410.02725
00:43:20 Ryan Greenblatt ARC entry, https://redwoodresearch.substack.com/p/getting-50-sota-on-arc-agi-with-gpt
Disclaimer: This show is part of our Cohere partnership series
-
Prof. Tim Rocktäschel, AI researcher at UCL and Google DeepMind, talks about open-ended AI systems. These systems aim to keep learning and improving on their own, like evolution does in nature.
Ad: Are you a hardcore ML engineer who wants to work for Daniel Cahn at SlingshotAI building AI for mental health? Give him an email! - [email protected]
TOC:
00:00:00 Introduction to Open-Ended AI and Key Concepts
00:01:37 Tim Rocktäschel's Background and Research Focus
00:06:25 Defining Open-Endedness in AI Systems
00:10:39 Subjective Nature of Interestingness and Learnability
00:16:22 Open-Endedness in Practice: Examples and Limitations
00:17:50 Assessing Novelty in Open-ended AI Systems
00:20:05 Adversarial Attacks and AI Robustness
00:24:05 Rainbow Teaming and LLM Safety
00:25:48 Open-ended Research Approaches in AI
00:29:05 Balancing Long-term Vision and Exploration in AI Research
00:37:25 LLMs in Program Synthesis and Open-Ended Learning
00:37:55 Transition from Human-Based to Novel AI Strategies
00:39:00 Expanding Context Windows and Prompt Evolution
00:40:17 AI Intelligibility and Human-AI Interfaces
00:46:04 Self-Improvement and Evolution in AI Systems
Show notes (New!) https://www.dropbox.com/scl/fi/5avpsyz8jbn4j1az7kevs/TimR.pdf?rlkey=pqjlcqbtm3undp4udtgfmie8n&st=x50u1d1m&dl=0
REFS:
00:01:47 - UCL DARK Lab (Rocktäschel) - AI research lab focusing on RL and open-ended learning - https://ucldark.com/
00:02:31 - GENIE (Bruce) - Generative interactive environment from unlabelled videos - https://arxiv.org/abs/2402.15391
00:02:42 - Promptbreeder (Fernando) - Self-referential LLM prompt evolution - https://arxiv.org/abs/2309.16797
00:03:05 - Picbreeder (Secretan) - Collaborative online image evolution - https://dl.acm.org/doi/10.1145/1357054.1357328
00:03:14 - Why Greatness Cannot Be Planned (Stanley) - Book on open-ended exploration - https://www.amazon.com/Why-Greatness-Cannot-Planned-Objective/dp/3319155237
00:04:36 - NetHack Learning Environment (Küttler) - RL research in procedurally generated game - https://arxiv.org/abs/2006.13760
00:07:35 - Open-ended learning (Clune) - AI systems for continual learning and adaptation - https://arxiv.org/abs/1905.10985
00:07:35 - OMNI (Zhang) - LLMs modeling human interestingness for exploration - https://arxiv.org/abs/2306.01711
00:10:42 - Observer theory (Wolfram) - Computationally bounded observers in complex systems - https://writings.stephenwolfram.com/2023/12/observer-theory/
00:15:25 - Human-Timescale Adaptation (Rocktäschel) - RL agent adapting to novel 3D tasks - https://arxiv.org/abs/2301.07608
00:16:15 - Open-Endedness for AGI (Hughes) - Importance of open-ended learning for AGI - https://arxiv.org/abs/2406.04268
00:16:35 - POET algorithm (Wang) - Open-ended approach to generate and solve challenges - https://arxiv.org/abs/1901.01753
00:17:20 - AlphaGo (Silver) - AI mastering the game of Go - https://deepmind.google/technologies/alphago/
00:20:35 - Adversarial Go attacks (Dennis) - Exploiting weaknesses in Go AI systems - https://www.ifaamas.org/Proceedings/aamas2024/pdfs/p1630.pdf
00:22:00 - Levels of AGI (Morris) - Framework for categorizing AGI progress - https://arxiv.org/abs/2311.02462
00:24:30 - Rainbow Teaming (Samvelyan) - LLM-based adversarial prompt generation - https://arxiv.org/abs/2402.16822
00:25:50 - Why Greatness Cannot Be Planned (Stanley) - 'False compass' and 'stepping stone collection' concepts - https://www.amazon.com/Why-Greatness-Cannot-Planned-Objective/dp/3319155237
00:27:45 - AI Debate (Khan) - Improving LLM truthfulness through debate - https://proceedings.mlr.press/v235/khan24a.html
00:29:40 - Gemini (Google DeepMind) - Advanced multimodal AI model - https://deepmind.google/technologies/gemini/
00:30:15 - How to Take Smart Notes (Ahrens) - Effective note-taking methodology - https://www.amazon.com/How-Take-Smart-Notes-Nonfiction/dp/1542866502
(truncated, see shownotes)
-
Ben Goertzel discusses AGI development, transhumanism, and the potential societal impacts of superintelligent AI. He predicts human-level AGI by 2029 and argues that the transition to superintelligence could happen within a few years after. Goertzel explores the challenges of AI regulation, the limitations of current language models, and the need for neuro-symbolic approaches in AGI research. He also addresses concerns about resource allocation and cultural perspectives on transhumanism.
TOC:
[00:00:00] AGI Timeline Predictions and Development Speed
[00:00:45] Limitations of Language Models in AGI Development
[00:02:18] Current State and Trends in AI Research and Development
[00:09:02] Emergent Reasoning Capabilities and Limitations of LLMs
[00:18:15] Neuro-Symbolic Approaches and the Future of AI Systems
[00:20:00] Evolutionary Algorithms and LLMs in Creative Tasks
[00:21:25] Symbolic vs. Sub-Symbolic Approaches in AI
[00:28:05] Language as Internal Thought and External Communication
[00:30:20] AGI Development and Goal-Directed Behavior
[00:35:51] Consciousness and AI: Expanding States of Experience
[00:48:50] AI Regulation: Challenges and Approaches
[00:55:35] Challenges in AI Regulation
[00:59:20] AI Alignment and Ethical Considerations
[01:09:15] AGI Development Timeline Predictions
[01:12:40] OpenCog Hyperon and AGI Progress
[01:17:48] Transhumanism and Resource Allocation Debate
[01:20:12] Cultural Perspectives on Transhumanism
[01:23:54] AGI and Post-Scarcity Society
[01:31:35] Challenges and Implications of AGI Development
New! PDF Show notes: https://www.dropbox.com/scl/fi/fyetzwgoaf70gpovyfc4x/BenGoertzel.pdf?rlkey=pze5dt9vgf01tf2wip32p5hk5&st=svbcofm3&dl=0
Refs:
00:00:15 Ray Kurzweil's AGI timeline prediction, Ray Kurzweil, https://en.wikipedia.org/wiki/Technological_singularity
00:01:45 Ben Goertzel: SingularityNET founder, Ben Goertzel, https://singularitynet.io/
00:02:35 AGI Conference series, AGI Conference Organizers, https://agi-conf.org/2024/
00:03:55 Ben Goertzel's contributions to AGI, Wikipedia contributors, https://en.wikipedia.org/wiki/Ben_Goertzel
00:11:05 Chain-of-Thought prompting, Subbarao Kambhampati, https://arxiv.org/abs/2405.04776
00:11:35 Algorithmic information content, Pieter Adriaans, https://plato.stanford.edu/entries/information-entropy/
00:12:10 Turing completeness in neural networks, Various contributors, https://plato.stanford.edu/entries/turing-machine/
00:16:15 AlphaGeometry: AI for geometry problems, Trieu, Li, et al., https://www.nature.com/articles/s41586-023-06747-5
00:18:25 Shane Legg and Ben Goertzel's collaboration, Shane Legg, https://en.wikipedia.org/wiki/Shane_Legg
00:20:00 Evolutionary algorithms in music generation, Yanxu Chen, https://arxiv.org/html/2409.03715v1
00:22:00 Peirce's theory of semiotics, Charles Sanders Peirce, https://plato.stanford.edu/entries/peirce-semiotics/
00:28:10 Chomsky's view on language, Noam Chomsky, https://chomsky.info/1983____/
00:34:05 Greg Egan's 'Diaspora', Greg Egan, https://www.amazon.co.uk/Diaspora-post-apocalyptic-thriller-perfect-MIRROR/dp/0575082097
00:40:35 'The Consciousness Explosion', Ben Goertzel & Gabriel Axel Montes, https://www.amazon.com/Consciousness-Explosion-Technological-Experiential-Singularity/dp/B0D8C7QYZD
00:41:55 Ray Kurzweil's books on singularity, Ray Kurzweil, https://www.amazon.com/Singularity-Near-Humans-Transcend-Biology/dp/0143037889
00:50:50 California AI regulation bills, California State Senate, https://sd18.senate.ca.gov/news/senate-unanimously-approves-senator-padillas-artificial-intelligence-package
00:56:40 Limitations of Compute Thresholds, Sara Hooker, https://arxiv.org/abs/2407.05694
00:56:55 'Taming Silicon Valley', Gary F. Marcus, https://www.penguinrandomhouse.com/books/768076/taming-silicon-valley-by-gary-f-marcus/
01:09:15 Kurzweil's AGI prediction update, Ray Kurzweil, https://www.theguardian.com/technology/article/2024/jun/29/ray-kurzweil-google-ai-the-singularity-is-nearer
-
AI expert Prof. Gary Marcus doesn't mince words about today's artificial intelligence. He argues that despite the buzz, chatbots like ChatGPT aren't as smart as they seem and could cause real problems if we're not careful.
Marcus is worried about tech companies putting profits before people. He thinks AI could make fake news and privacy issues even worse. He's also concerned that a few big tech companies have too much power. Looking ahead, Marcus believes the AI hype will die down as reality sets in. He wants to see AI developed in smarter, more responsible ways. His message to the public? We need to speak up and demand better AI before it's too late.
Buy Taming Silicon Valley:
https://amzn.to/3XTlC5s
Gary Marcus:
https://garymarcus.substack.com/
https://x.com/GaryMarcus
Interviewer:
Dr. Tim Scarfe
(Refs in top comment)
TOC
[00:00:00] AI Flaws, Improvements & Industry Critique
[00:16:29] AI Safety Theater & Image Generation Issues
[00:23:49] AI's Lack of World Models & Human-like Understanding
[00:31:09] LLMs: Superficial Intelligence vs. True Reasoning
[00:34:45] AI in Specialized Domains: Chess, Coding & Limitations
[00:42:10] AI-Generated Code: Capabilities & Human-AI Interaction
[00:48:10] AI Regulation: Industry Resistance & Oversight Challenges
[00:54:55] Copyright Issues in AI & Tech Business Models
[00:57:26] AI's Societal Impact: Risks, Misinformation & Ethics
[01:23:14] AI X-risk, Alignment & Moral Principles Implementation
[01:37:10] Persistent AI Flaws: System Limitations & Architecture Challenges
[01:44:33] AI Future: Surveillance Concerns, Economic Challenges & Neuro-Symbolic AI
YT version with refs: https://youtu.be/o9MfuUoGlSw
-
Prof. Mark Solms, a neuroscientist and psychoanalyst, discusses his groundbreaking work on consciousness, challenging conventional cortex-centric views and emphasizing the role of brainstem structures in generating consciousness and affect.
MLST is sponsored by Brave:
The Brave Search API covers over 20 billion webpages, built from scratch without Big Tech biases or the recent extortionate price hikes on search API access. Perfect for AI model training and retrieval augmentated generation. Try it now - get 2,000 free queries monthly at http://brave.com/api.
Key points discussed:
The limitations of vision-centric approaches to consciousness studies.
Evidence from decorticated animals and hydranencephalic children supporting the brainstem's role in consciousness.
The relationship between homeostasis, the free energy principle, and consciousness.
Critiques of behaviorism and modern theories of consciousness.
The importance of subjective experience in understanding brain function.
The discussion also explored broader topics:
The potential impact of affect-based theories on AI development.
The role of the SEEKING system in exploration and learning.
Connections between neuroscience, psychoanalysis, and philosophy of mind.
Challenges in studying consciousness and the limitations of current theories.
Mark Solms:
https://neuroscience.uct.ac.za/contacts/mark-solms
Show notes and transcript: https://www.dropbox.com/scl/fo/roipwmnlfmwk2e7kivzms/ACjZF-VIGC2-Suo30KcwVV0?rlkey=53y8v2cajfcgrf17p1h7v3suz&st=z8vu81hn&dl=0
TOC (*) are best bits
00:00:00 1. Intro: Challenging vision-centric approaches to consciousness *
00:02:20 2. Evidence from decorticated animals and hydranencephalic children *
00:07:40 3. Emotional responses in hydranencephalic children
00:10:40 4. Brainstem stimulation and affective states
00:15:00 5. Brainstem's role in generating affective consciousness *
00:21:50 6. Dual-aspect monism and the mind-brain relationship
00:29:37 7. Information, affect, and the hard problem of consciousness *
00:37:25 8. Wheeler's participatory universe and Chalmers' theories
00:48:51 9. Homeostasis, free energy principle, and consciousness *
00:59:25 10. Affect, voluntary behavior, and decision-making
01:05:45 11. Psychoactive substances, REM sleep, and consciousness research
01:12:14 12. Critiquing behaviorism and modern consciousness theories *
01:24:25 13. The SEEKING system and exploration in neuroscience
Refs:
1. Mark Solms' book "The Hidden Spring" [00:20:34] (MUST READ!)
https://amzn.to/3XyETb3
2. Karl Friston's free energy principle [00:03:50]
https://www.nature.com/articles/nrn2787
3. Hydranencephaly condition [00:07:10]
https://en.wikipedia.org/wiki/Hydranencephaly
4. Periaqueductal gray (PAG) [00:08:57]
https://en.wikipedia.org/wiki/Periaqueductal_gray
5. Positron Emission Tomography (PET) [00:13:52]
https://en.wikipedia.org/wiki/Positron_emission_tomography
6. Paul MacLean's triune brain theory [00:03:30]
https://en.wikipedia.org/wiki/Triune_brain
7. Baruch Spinoza's philosophy of mind [00:23:48]
https://plato.stanford.edu/entries/spinoza-epistemology-mind
8. Claude Shannon's "A Mathematical Theory of Communication" [00:32:15]
https://people.math.harvard.edu/~ctm/home/text/others/shannon/entropy/entropy.pdf
9. Francis Crick's "The Astonishing Hypothesis" [00:39:57]
https://en.wikipedia.org/wiki/The_Astonishing_Hypothesis
10. Frank Jackson's Knowledge Argument [00:40:54]
https://plato.stanford.edu/entries/qualia-knowledge/
11. Mesolimbic dopamine system [01:11:51]
https://en.wikipedia.org/wiki/Mesolimbic_pathway
12. Jaak Panksepp's SEEKING system [01:25:23]
https://en.wikipedia.org/wiki/Jaak_Panksepp#Affective_neuroscience
-
Dr. Patrick Lewis, who coined the term RAG (Retrieval Augmented Generation) and now works at Cohere, discusses the evolution of language models, RAG systems, and challenges in AI evaluation.
MLST is sponsored by Brave:
The Brave Search API covers over 20 billion webpages, built from scratch without Big Tech biases or the recent extortionate price hikes on search API access. Perfect for AI model training and retrieval augmented generation. Try it now - get 2,000 free queries monthly at http://brave.com/api.
Key topics covered:
- Origins and evolution of Retrieval Augmented Generation (RAG)
- Challenges in evaluating RAG systems and language models
- Human-AI collaboration in research and knowledge work
- Word embeddings and the progression to modern language models
- Dense vs sparse retrieval methods in information retrieval
The discussion also explored broader implications and applications:
- Balancing faithfulness and fluency in RAG systems
- User interface design for AI-augmented research tools
- The journey from chemistry to AI research
- Challenges in enterprise search compared to web search
- The importance of data quality in training AI models
Patrick Lewis: https://www.patricklewis.io/
Cohere Command Models, check them out - they are amazing for RAG!
https://cohere.com/command
TOC
00:00:00 1. Intro to RAG
00:05:30 2. RAG Evaluation: Poll framework & model performance
00:12:55 3. Data Quality: Cleanliness vs scale in AI training
00:15:13 4. Human-AI Collaboration: Research agents & UI design
00:22:57 5. RAG Origins: Open-domain QA to generative models
00:30:18 6. RAG Challenges: Info retrieval, tool use, faithfulness
00:42:01 7. Dense vs Sparse Retrieval: Techniques & trade-offs
00:47:02 8. RAG Applications: Grounding, attribution, hallucination prevention
00:54:04 9. UI for RAG: Human-computer interaction & model optimization
00:59:01 10. Word Embeddings: Word2Vec, GloVe, and semantic spaces
01:06:43 11. Language Model Evolution: BERT, GPT, and beyond
01:11:38 12. AI & Human Cognition: Sequential processing & chain-of-thought
Refs:
1. Retrieval Augmented Generation (RAG) paper / Patrick Lewis et al. [00:27:45]
https://arxiv.org/abs/2005.11401
2. LAMA (LAnguage Model Analysis) probe / Petroni et al. [00:26:35]
https://arxiv.org/abs/1909.01066
3. KILT (Knowledge Intensive Language Tasks) benchmark / Petroni et al. [00:27:05]
https://arxiv.org/abs/2009.02252
4. Word2Vec algorithm / Tomas Mikolov et al. [01:00:25]
https://arxiv.org/abs/1301.3781
5. GloVe (Global Vectors for Word Representation) / Pennington et al. [01:04:35]
https://nlp.stanford.edu/projects/glove/
6. BERT (Bidirectional Encoder Representations from Transformers) / Devlin et al. [01:08:00]
https://arxiv.org/abs/1810.04805
7. 'The Language Game' book / Nick Chater and Morten H. Christiansen [01:11:40]
https://amzn.to/4grEUpG
Disclaimer: This is the sixth video from our Cohere partnership. We were not told what to say in the interview. Filmed in Seattle in June 2024.
- Näytä enemmän
