PodCompanions
Discussions: 2026-W43 — page 54
Discussions added each week. These are publication dates here; the original episodes may be older.
- 0:00–5:00
Hsu recalls an encounter with Richard Feynman and the personal advice he says Feynman offered.
- 25:01–29:58
Hsu discusses evolutionary arguments for additive variation and argues that changing environments complicate ideas of biological optimisation.
- 58:24–1:03:20
Hsu acknowledges environmental effects on height and reduced prediction quality across ancestry groups, discussing possible explanations and future data needs.
- 2:09:45–2:12:48
Hsu offers family connections, peer exposure and networking as possible explanations for career advantages associated with elite universities.
- 1:18:53–1:23:25
Hsu distinguishes educational attainment from cognitive test performance and describes differences in predictive power within sibling pairs.
- 16:17–21:04
Hsu describes his move from physics into genomics, compressed sensing and his group’s claims about the sample sizes needed for prediction.
- 43:20–46:25
Hsu discusses clinic relationships, data access, patents and the uncertain durability of his company’s scientific advantage.
- 1:23:25–1:28:17
The speakers discuss scientific communication and possible consumer uptake; Hsu distinguishes the IVF market from broader speculation about human evolution.
- 1:11:18–1:15:22
Hsu argues that better cognitive measurements could improve predictors and offers his interpretation of obstacles to collecting those data.
- 2:17:48–2:20:49
The speakers conclude their discussion and point listeners to Hsu’s writing and Manifold podcast.
- 1:03:20–1:06:45
The speakers consider distinguishing causal variants from statistical tags and discuss uncertainties and controversies around interpreting population differences.
- 1:58:02–2:03:02
The speakers discuss transitions out of physics and use stories about Jeff Bezos to consider different combinations of abilities.
- 1:28:17–1:33:02
Hsu discusses adoption within IVF and praises independent scholarship, while considering the opportunities and constraints of cross-disciplinary work.
- 21:04–25:01
Hsu explains additive predictors, correlated variants and sparse optimisation; Patel asks why selection has not maximised the traits discussed.
- 1:06:45–1:11:18
The speakers speculate about regulation and access, then discuss historical and political attitudes toward reproductive technologies.
- 2:03:02–2:07:18
Hsu argues that cognitive abilities are multidimensional and discusses technical problem-solving and founders’ communication skills.
- 30:52–35:13
Kimmel explains why detecting and amplifying rare pluripotent cells made Yamanaka’s problem different from therapeutic rejuvenation.
- 45:23–50:07
Kimmel discusses viral interactions with gene regulation and explains physical challenges in targeting transcription factors with different drug types.
- 1:14:59–1:19:21
Kimmel distinguishes identifying useful targets from making molecules, then describes predictive cell models as a proposed route to better discovery.
- 1:06:22–1:09:35
Kimmel considers engineered transcription factors and suggests that repairing extracellular structures may require additional cell states.
- 1:19:21–1:24:15
Kimmel separates predicting expression changes from choosing desirable cell states and clarifies that his team is not currently using reinforcement learning.
- 1:24:15–1:28:02
Kimmel explains Perturb-seq, describes technical improvements and says combinatorial experiments require reliable identification of multiple interventions.
- 1:31:32–1:36:01
Kimmel offers hypotheses about reimbursement and direct purchasing for durable treatments while setting aside broader intellectual-property enforcement questions.
- 1:03:04–1:06:22
Kimmel discusses payload size and hopes for durable effects while explicitly limiting the available reprogramming evidence to weeks.