Some discovery features could not load. You can still browse the latest articles below.
Anthropic CEO Dario Amodei weighs an exponential revenue curve, model-by-model economics, and adoption lags against the risk that models err strangely, regulation misfires, or growth plateaus.
Anthropic CEO Dario Amodei tells Dwarkesh Patel scaling works and AI geniuses may arrive in 1-3 years, but uncertain economic diffusion, not technology, limits compute purchases.
DeepMind's Demis Hassabis, accepting the RSA Albert Medal, argues AI will arrive fast and cannot be stopped, yet insists its outcome depends on society's choices about value, economics and purpose.
OpenAI's Sam Altman says capabilities cannot outrun alignment and monitoring, yet concedes no lab has solved alignment, and that a sandbox escape forced the company to change course.
Vercel CEO Guillermo Rauch argues that agents reward API-first infrastructure and verification loops, while insisting humans must keep understanding and judgment as autonomy grows.
Elon Musk argues that power, not chips, will limit AI scaling soon, while interviewers press him on space economics, alignment, humanoid robots and how he manages bottlenecks.
Steve Eisman, who bet against subprime, argues the AI capex behind half of US growth depends on two money-losing companies, while private credit risks look painful but not systemic.
Cognitive psychologist Steven Pinker argues AI demands liability and engineering-style safety standards rather than mysticism, questioning whether superintelligence talk reflects technical reality or incentives to inflate investor expectations.
Khan Academy founder Sal Khan, usually an optimist, argues AI-driven dislocation will arrive within years while society lacks reskilling answers, and is building a skills-based degree as one response.
Theorist Scott Aaronson argues AI is clearing career-defining problems, leaving humans to interpret results, choose questions and supervise unaligned systems, though he doubts any such role is permanent.
Cameron Berg argues that scaled neural networks may already have functional emotions, and that AI safety and AI welfare share one unresolved root: nobody knows what these systems are.
Signal AI analysts Andrew Smith and Scott Wilkinson argue that Marvell's reported Lumentum bid is about owning the whole optical stack, even though it would mean competing with Marvell's own DSP customers.
Google DeepMind's Nenad Tomasev argues that AI capability risks outrunning the laboratories, human expertise, alignment methods and multi-agent governance needed to turn it into reliable scientific and social benefit.
David Tisch of Box Group defends his 150-company, low-ownership seed model against a host who once doubted it, arguing larger outcomes and thinner dilution have rewritten venture math.