Guests make confident forecasts on podcasts, but memory is short. We log high-stakes claims, timestamp the exact evidence, and track whether they prove true.
On Demis Hassabis - Scaling to AGI, AlphaFold 3, & The Physics of Mind
"AI will compress the discovery timeline of novel small-molecule cancer drugs from years into weeks before 2027."
Context: Explaining the real-world biological impact of AlphaFold 3's protein-ligand interaction modeling.
On Nikhil Kamath - Why 99% of Startups Shouldn't Raise VC Money
"Over 50% of funded Indian consumer fintech startups will consolidate or shut down due to unsustainable customer acquisition costs."
Context: Debating why burning VC capital on user discounts fails when financial products are commoditized.
On Yann LeCun - Why LLMs Will Never Achieve Human-Level Intelligence
"Within 3 years, the mainstream AI research community will accept that autoregressive LLMs cannot achieve AGI and will pivot to hierarchical world models."
Context: Discussing the exponential accumulation of errors in multi-step autoregressive generation.
