Hidden Layers: AI and the People Behind It

Dense with genuine scientific explanation (protein folding, AlphaFold, diffusion models, CAR-T mechanisms, the verification dilemma) that a curious listener would learn from, though much is educational background rather than novel operator-relevant insight, and there's notable repetition and meandering.

“it's sort of like running cloud code, except you can never run the code”

“actually like half of the designs didn't even like results in really like functional T cells at all”

Some fresh framing (the code-can't-run analogy, generated proteins looking unlike natural sequences, the honest pessimistic note about low hit rates and lack of a 'CASP for design'), but much of the protein-folding narrative is standard popular-science recounting.

“these generated proteins look very different from like regular proteins in terms of the sequences”

“there's no casp there's no like critical assessment of these methods you know head to head”

Guests are PhD candidates who actually organized and ran the BitsToBinders challenge—genuine hands-on practitioners on the specific topic—but they are early-career students rather than senior operators who have done this at scale.

“all of whom are PhD student candidates at the University of Texas at Austin”

“very young PhD students maybe in their first year who are interested in this space”

Rich in concrete detail: exact counts, success rates, sequence constraints, named companies and named winning teams, specific proteins and timelines—well above average for evidentiary grounding.

“we tested 12,000 sequences in pooled screens”

“the best hit rate was nuclear UK London and they had a 38% hit rate”

Host asks several pointed, well-structured questions and follows up (the digital-to-real gap, hardest step, student-led vs. corporate constraints), though the tone is largely celebratory with little genuine pushback, and one guest had to volunteer the skeptical counterpoint himself.

“What was actually the hardest part? Was it what you thought it was going to be?”

“How big is that gap? Like what did you learn... is the gap bigger than you expected?”