2026-04-29

Lucius Caviola with initial musings on questions of AI ethics of global geostrategic importance.

Anton Leicht arguing against raising the political salience of artificial intelligence as a legislative strategy, noting that when precision is required, the volatility of public polarization should be avoided. Though notably those who are most eager to raise the salience of AI in the public are exactly those who do not believe that precise solutions are necessary, or even possible.

Ezra Klein retrospective on the effects of Abundance, and to what extent people are actually embracing the fundamental goal of increasing supply, rather than getting sidetracked into various adjacent arguments1.

Alexander Wales with a short story on persisting in futile individual action, along with assorted commentary on starfish strandings.

Dialectic interview with Nicholas Thompson as a representation of the Atlantic worldview, a personally appealing combination of humility and discipline which the world could do with more of. Goodwill now given, it is disappointing to me that the journalistic profession still has not realized the actual reason for the dissatisfaction that tech is showing them. Yes, the Atlantic has probably the best editors in the world, and many of the best writers, but unfortunately intelligence and writing skill alone is insufficient for fair and accurate reporting of complicated topics. The real problem with the view from nowhere isn’t that neutrality is impossible, because neither is perfection in writing, accuracy, or interestingness, and yet all these are things worth striving for. The actual problem is that neutrality is hard, because to actually make everything balance requires a finely-tuned understanding of how much weight every point actually deserves, without making the mistake of assuming that attention is equivalent to truth, selection effects which reduce a publication into a mere amplifier of existing distortions. Complaints about the illusion of neutrality are actually complaints about the illusion of expertise2 among not only journalists, but also managers, lawyers, and politicians: that your arguments are bad; your regulations are wrongheaded; you have no idea what you are actually doing; you are making a mess that we will later have to clean up after you. To some extent, this state of affairs is what is being acknowledged by claims of constraining and speaking truth to power, but given how hard it is to do things nowadays, then insofar as the role of journalism is acknowledged to be restraining the actions of others, then perhaps it’s a good thing that it’s importance is in decline.

Paola on some interesting techniques of covert social manipulation3.

Ariel Zeleznikow-Johnston on the need for high quality brains to develop techniques of mapping the human connectome.

Free Radicals podcast with James Peyer on longevity medicine. It seems somewhat ironic to me that so much money is entering this field only once the boomers have already gotten rather old; clearly a suboptimal decision of timing on their part.

Dan Schwartz in Asterisk Mag asking to what extent prediction markets actually produce significant positive second-order effects in providing useful information to external observers. Somewhat related, MAbromov has a controversial LessWrong post recently which argued that forecasting in general is overrated. Personally, it seems to me that even if AI continues to consistently outperform human forecasters across most topics, it’s useful to retain human involvement if only as an additional verification and a means to resist disempowerment.

Benjamin Breen commentaries inspired by Talkie-1930.

Henry Stanley linkthread of clearly very highly impactful scientific papers.

Alexander Kustov immigration linkthread.