2024-07-25
free
Emily Chamlee-Wright on liberalism, via Arnold Kling. It’s a good analogy, that in the same way that the market is best positioned to solve market failures, that liberalism is the best way to solve the problems of liberalism. Direct intervention puts the system in a state of stasis that is unresponsive to the next wave of failures.
Elle Griffin on ESOP (Employee Stock Ownership Plan) companies, which have more aligned incentives. It always puzzles me when people push for unionization at tech companies, exactly because tech companies are typically run this way (excluding some cases like video game developers, which perhaps need better worker protections).
Noah Smith (and Noah Millman) comments on a JD Vance interview where he mentions “childless cat ladies”, which is interesting, because it wouldn’t surprise me if the quote came directly from the Collins’. The response seems largely negative, with the left taking offense because one’s worth isn’t related to their children or lack thereof and it is sexist to imply otherwise, and the right disavowing it because they all have relatives who are childless not by choice but because of infertility or the inability to find a spouse and so it seems needlessly mean. Of course, he probably only means to refer to the purposely childfree (possibly even as a way to alleviate the concerns mentioned in these Based Camp episodes about whether he’s really a “tech-elite” Republican versus a populist social conservative), but the mental framework around that distinction isn’t really widespread. Probably he needs to say childfree instead of childless, and something else rather than lady or cat lady, which comes with baggage. I wonder whether this will become part of the mainstream discourse, and if it does, whether it actually promotes a further backlash in which having children becomes (even more?) right-coded.
Chris Jesu Lee on the male personal essay discourse. Basically everyone knows why this is, which is that the personal is political, which means it is unlikely that your essay will be both true and palatable. But I don’t necessarily see this as a bad thing. But the personal essay seems to be a local optima that people get stuck into simply because its so much easier to put raw emotions in when writing them. Bereft of this easy path, maybe we’ll get something different out, instead of just another copy of Knausgaard or Roth.
Adam of Experimental History in Seeds of Science saying to abolish peer review.
Llama 3.1 paper, Dean Ball analysis. He references a quote saying that it will bring all open-source models 3-9 months ahead. What’s interesting is whether the closed AI labs will be able to take anything from it, or if Meta was merely running catch up. Notably, Llama is not MoE, which anecdotally is what OpenAI is using. Will be an interesting data point to the question of how sparse the space of performant models actually is.
Bayesian Conspiracy on martial arts. Reminds me of this article on the moat of low status, probably the most influential thing I read recently which has totally changed how I approached 2024.
