2026-06-09

giant

Abhishaike Mahajan on the mechanisms behind cancer vaccines and why they have finally started working as a result of cheap sequencing and advances in bioinformatics.

Eryney Marrogi arguing against mandatory screening of synthesized DNA orders inspired by fears of AI-enabled bioterrorism. Personally, it seems to me that claims that Anthropic and other AI companies are offloading regulation downstream is somewhat unfair, given that they are also doing things like adding safeguards on their own end. It would clearly be better if biotech could undergo regulatory reforms, such as proposed by Adam Kroetsch, instead of only adding additional regulations, but it seems perhaps somewhat of a conflation to treat them as if they are mutually exclusive (Edit: Cremieux with similar claims to Marrogi, which again I personally read as originating out of an expectation that existing regulations will never be reformed, in which case any additional regulations must be avoided at all costs in order to attempt to keep American biotech competitive).

Sasha Gusev with some interesting thoughts on AI in academic research; not fully endorsed, but it seems to be a good development that people in academia are starting to take things more seriously.

Anthropic system card for Mythos 5, and Ethan Mollick commentary on the capabilities of Fable 5.

Andy Hall overview of AI discourse in popular culture.

Harry Law in Cosmos Institute with an interesting partial repudiation of futurist views of politics in which policies are determined through straightforward aggregation of individual preferences across society, noting that it ignores the question of where individual preferences originate in the first place. One way of answering this question if as a result of genetics, or alternatively through a shared cultural upbringing which is itself determined by aggregated societal preference. But in practice, it’s exactly because people are diverse that the formation of shared cultural values cannot be generated by a fixed process generated through aggregation, but only by one which is responsive and adapted to local context, particularly if those values include things like agency and civic participation. Personally, given that the strength of AI is exactly that it can deliver individuation at scale, it’s unclear to me whether his assertion that there is “no AI system, however carefully designed, that can deliver [a formative account of civic life]”, but it does seem that attempts to do so may be somewhat prone to producing dystopian situations.

Stetson interview with Darby Saxbe, probably the most interesting interview of her current circuit promoting Dad Brain (edit: Elena Bridgers interview also quite good, which discusses something similar to my theory of how contradictory desires are complicating modern courtship).

Tailcalled describes their generation of their (non-lexical) factor model of personality.

S. Y. Lee with some interesting commentary on the Seoul reelection of Oh Se-hoon. Tangentially related, Oliver Traldi with a nuanced position on the Great Man theory of history, on the nonequivalence of greatness and uniqueness; unclear to me to what extent this applies to people like Alexander.

Daniel Muñoz Ian Shapiro on America and the west in the post-Cold war era, including an interesting theory on the unpopularity of neoliberalism as a result of voter backlash to feelings of economic insecurity in the abscence of lifelong employment, presumably a result of union-busting, deliberate government non-interference in the market, and the experience of the GFC (and now, AI-induced breakdown in formal hiring). Tangentially related, Naomi Kanakia review of Liberties Magazine (edit: and Grace Byron on the cultural niche of the little magazine).

SemiAnalysis (partial paywall) profile of Unitree Robotics.

Chinese Doomscroll with a special edition of translated social media commentary on the Gaokao. Somewhat related, Zilan Qian on technological innovation in Gaokao cheating and countermeasures.

Jing Jing P. review of Love Letters to Grandma.

Celine Nguyen literature linkthread.