Voices

Deep Medicine; clinical AI evidence standards

AlphaFold lead

AlphaFold/AlphaTensor science portfolio

ML for drug discovery and oncology

State-space models; Mamba co-author

GPT-1/2/3 and CLIP first author; the quiet architect of the LLM era

Vision-to-LLM arc; the field's best explainer of how models work

Google Brain co-founder; the field's largest teaching footprint

Visual geometry and recognition (VGG)

Attention Is All You Need lead author

Meta-learning (MAML); robot learning

ML systems and AI engineering canon

T5; open-model training science

Visual recognition and video understanding

AlphaGo/AlphaZero lead; RL at scale

VAE and Adam author

Keras; On the Measure of Intelligence; ARC benchmark

3Blue1Brown; the visual mathematics of neural networks

GANs; adversarial examples; Deep Learning textbook

Chain-of-thought and emergent-abilities papers

The Illustrated Transformer; visual ML pedagogy

fast.ai; made deep learning learnable; answer.ai

Computer vision foundations

Reproducibility standards; open research leadership

ResNet; representation learning

Statistical learning canon in Chinese; industrial NLP research

Safety systems research; canonical technical blog

VAE co-author; graph networks; ML-for-physics

Dive into Deep Learning; parameter server; ML systems teaching

HELM evaluation; foundation-models framing (CRFM)

Robot learning; apprenticeship learning

Chinese NLP curriculum; MOSS early Chinese LLM

AutoML, seq2seq, LaMDA lineage

Hardware lottery; multilingual frontier research

LLM-from-scratch pedagogy; open technical writing

Deep robotic learning; offline RL

Primary technical analysis of LLM tooling; prompt-injection naming

Vision and cognitive architectures; returned to lead BIGAI

word2vec

FlashAttention; Mamba co-author

Transfer and federated learning

Commonsense reasoning; delphi and value learning

Watermelon Book; ensemble learning; China's ML curriculum

Economics of transformative AI; AGI preparedness

Prediction-machines economics of AI

Frey-Osborne automation-exposure estimates; Technology Trap

Power and Progress; task-displacement economics; AI growth skepticism

Task framework of labour and automation

Second Machine Age; productivity-J-curve; Turing trap

GPTs-are-GPTs labour exposure research

Dataset audits; relational ethics of AI

Data practices critique; The AI Con co-author

AI Snake Oil; evaluation skepticism with technical depth

Privacy Is Power

Weapons of Math Destruction; algorithmic auditing practice

Differential privacy; formal fairness

Interpretable-models-first position

Stochastic Parrots; linguistic critique of LLM claims

Black Box Society; new laws of robotics

Deep-learning limits; hybrid AI advocacy; policy testimony

Algorithmic auditing frameworks

Algorithmic Justice League; face-recognition audits

Atlas of AI; material and political costs of AI

Philosophy of information; digital ethics in EU policy

Model Cards; ML documentation practice

AI ethics and political philosophy of AI

Artificial Unintelligence; technochauvinism

AI Now co-founder; surveillance-business-model critique from inside industry

Equalized odds; fairness-and-ML textbook

Race After Technology; the New Jim Code

Algorithms of Oppression

Counterfactual explanations; GDPR-AI bridge

AI Snake Oil co-author; reproducibility in ML-based science

Technology and the virtues; The AI Mirror

Gender Shades, Datasheets, Stochastic Parrots; DAIR

Responsible AI as engineering discipline; UN advisory roles

Reinforcement learning co-founder

Complexity theory; built China's elite AI pipeline (Yao Class)

Kernel methods; causal ML

Statistical and neural NLP

Probabilistic graphical models; ML for biomedicine

ImageNet; human-centered AI; spatial intelligence

Backpropagation, deep learning revival; left Google to warn on risk

Hopfield networks; physics of neural computation

Bayesian networks and causality

LSTM lineage; meta-learning; priority debates

Neocognitron, ancestor of CNNs

PAC learning; computational learning theory

Complexity, analogy, and the limits of current AI

Probabilistic ML foundations; critic of AI framing

Multi-agent systems; AI history for the public

AIMA textbook; data-over-cleverness empiricism

Reinforcement learning; The Bitter Lesson

LSTM; modern xLSTM revival

AIMA textbook; provably beneficial AI agenda

Boltzmann machines; computational neuroscience

Statistical learning theory; SVMs

Convolutional networks; open-research advocate and AGI-risk skeptic

Deep learning foundations; now leads international AI safety reporting

Transformer co-author; enterprise LLMs

Data engine for frontier training; Meta superintelligence push

Europe's frontier-lab bet

Open-model ecosystem's central platform

Anthropic co-founder; operations and policy posture

Robotics research leadership; liquid networks

Anthropic co-founder; scaling-plus-safety thesis; Machines of Loving Grace essay

DeepMind founder; AlphaGo-to-AlphaFold arc

OpenAI co-founder turned competitor; Grok/xAI; policy lightning rod

OpenAI co-founder; engineering of frontier training

LLaMA lineage; Mistral research

AlexNet, seq2seq, GPT lineage; superintelligence-focused lab

Frontier model training leadership at OpenAI

Scaling laws; Anthropic science direction

Google's ML infrastructure era; TensorFlow, TPU programme

Built the compute substrate of the AI era

PPO and RLHF in practice

AI Superpowers; China-US ecosystem bridge; 01.AI

DeepMind research-to-product engine

DeepSeek; open frontier models at disruptive cost

The credible second source for AI compute

Legged robotics from lab to industry

Led ChatGPT/GPT-4 product era; new frontier lab

DeepMind/Inflection co-founder; The Coming Wave

Transformer co-author; MoE; Character.AI and return to Google

seq2seq, AlphaStar, Gemini technical leadership

Scaled OpenAI from lab to consumer platform

Coined AGI usage at DeepMind; long-horizon safety

GLM model family; Tsinghua-to-industry pipeline

ERNIE; China's first large-scale LLM deployments

OpenAI co-founder; robotics-to-LLM arc

Long-context LLMs; Kimi

Founded AI governance as research field (GovAI)

AI Bill of Rights blueprint

The Brussels Effect; digital-empires framing

ALLAI; EU high-level expert group on AI

NIST AI Risk Management Framework

Architect of the EU AI Act text

Normative infrastructure for AI; regulatory markets

Frontier-lab governance; CSET analysis; OpenAI board episode

State of AI report; UK frontier taskforce/AISI chair

Import AI; policy from inside a frontier lab

ChinAI; diffusion-centric view of AI power

Tech Coup; EU digital policy voice in the US

Frontier-regulation architecture

China's AI regulatory system analysis

Runtime governance for AI agents; AI compliance as policies on execution paths

Researcher at Eindhoven University of Technology and Kyvvu B.V. (Nijmegen). Corresponding author of the 2026 framework 'Runtime Governance for AI Agents: Policies on Paths,' which models AI-agent compliance as deterministic policies evaluated on execution paths.

Malicious-use report; AGI readiness critique after OpenAI exit

Enterprise AI governance curation and frameworks

Autonomous weapons; military AI doctrine

Algorithmic audit practice; red-teaming at scale

Genius Makers; deep-learning era chronicle

Primary-source long interviews with the frontier; Scaling Era

SemiAnalysis; compute supply-chain analysis

Co-Intelligence; empirical workplace-AI experiments

Empire of AI; defining investigative record of OpenAI era

Clearview exposé; privacy reporting

Code Dependent; FT AI editorship

State of AI Report

Interconnects; open post-training analysis

Supremacy; FT Business Book of the Year 2024

Transformer newsletter; AI policy reporting

The most thorough weekly synthesis of AI developments

Biological-anchors timelines; AI takeover analysis

Human-robot interaction; value alignment in practice

Dangerous-capability evaluations as policy instrument

Founded mechanistic interpretability

Open-source GPT replication turned doom advocacy

ML safety benchmarks; CAIS extinction-risk statement

Founded the alignment problem discourse; doom case

Mesa-optimization; Sleeper Agents

Most-important-century framing; safety funding architecture

ML robustness and forecasting; measurement-driven safety

Led superalignment; public resignations over safety priority

Life 3.0; FLI pause letter and policy campaigns

MIRI research leadership; co-author of the 2025 risk bestseller

Mech-interp field-building and training pipeline

Superintelligence; existential-risk framing

RLHF origins; alignment research agenda; now in government-adjacent safety

Alignment Newsletter; DeepMind AGI safety

Uncontrollability thesis

LLM evaluation; safety cases at frontier scale

Value extrapolation; safe AI startup research

The Precipice; existential risk quantification

Specification gaming; FLI co-founder