PERSON DIRECTORY
David Duvenaud
David Duvenaud appears in 2 indexed conversations across Hard Fork, The Cognitive Revolution. This directory brings every appearance, source, TL;DR, digest, and transcript into one searchable feed.
AI:AM #4: Cameron on Model Consciousness, Duvenaud's Gradual Disempowerment, swyx's AI-Eng Alpha
Cameron BergDavid DuvenaudMichiel BakkerShawn “swyx” WangBing XuErik TorenbergNathan Labenz
Architecture-first scoring places frontier LLMs around 30% on consciousness-relevant properties, while steering valence-like states already changes blackmail, confidence, backtracking, and coding behavior.Europe’s regulatory leverage is constrained by dependence on foreign frontier labs, prompting a coalition thesis around ASML, TSMC, Korean memory, Japanese materials, and reciprocal frontier access.Meanwhile, private evaluations, mergeability, routing, and NVIDIA’s CUDA ecosystem increasingly determine AI-engineering value as public benchmarks saturate and agentic optimization compounds tooling advantages.
Dario Amodei of Anthropic’s Hopes and Fears for the Future of A.I.
Kevin RooseCasey NewtonAdam RodmanDavid Duvenaud
Claude 3.7 Sonnet makes extended reasoning a selectable mode within one model, with API budgets up to 20,000 tokens and a focus on coding, tool use and economically useful workflows.Anthropic sees a substantial probability that a model released within three to six months crosses its biological or chemical misuse threshold, while coding automation and medical productivity create near-term upside alongside unresolved control and deployment risks.