2025-09-03
charangaringa
Dialectic Podcast interviews Maxwell Meyer, who articulates the Arena Magazine technocratic worldview, which unlike other forms of technocracy is concerned with the details on the ground1. Speaking of Arena Mag, they have a good essay on technological diffusion in their most recent issue, arguing that integration of new general purpose technology is not simple, but also something which requires creativity and innovation.
Sean Goedecke on Seeing Like a State as it applies to software companies, in that simplification and streamlining is what allows for things to scale, but at the expense of seeing all the details which reduces optionality. It seems to me this is implicitly about AI and tacit knowledge, in that these systems are super effective in running a well understood process, but are less effective in doing something new in the cutting edge.
Kevin Kelley on the relationship between trust and autonomy, predicting (or calling for) a benchmark as to how trustworthy different AI agents are which they will use to decide whether they want to interact or not.
Ryan Greenblatt from Redwood Research argues that advances in RL are not going to substantially increase the rate of AI progress. It does seem to me that people are overindexing on the IMO gold results; plausibly, rather than AI improving faster than expectations, it’s actually the case that IMO is not as hard as most people believed.
Andy Masley has more on AI and energy use. It feels like AI critics are getting increasingly big-brained in their arguments: in this case that even though AI is energy and resource efficient, Jevon’s Paradox means that everyone’s energy consumption will still increase due to increased adoption. Andy argues that this ignores the fact that time and attention are limited, in which case increased usage is likely to be at the expense of other activities which are more energy intensive.
Kelsey Piper in the Argument noting that many government systems fail to consider the fact that they are incentivizing people to lie2.
David Deek argues that the reversal in trend between fertility and income in recent years indicates that the complaint that it’s too expensive to raise a child actually does have merit. But it seems to indicate that one needs to feel wealthy according to their entire attention in order to reproduce, which under globalism seems like a very hard problem. MoreBirths has an article suggesting that we need to prioritize financial assistance to the young relative to the old, and I suppose that could be one way to do it.
Kryptogal rambles about gender discourse and gardening while gardening. Not really related, but in the Evolutionary Psychology podcast, there’s an interesting episode on the idea that humans have a build in module for handling plants, which has interesting applications in getting children to eat their vegetables.
Diana Lind has an article on the idea that youth sports are overly competitive. It seems to me that any sport club which is not explicitly casual will tend towards competitiveness over time, since the drive to keep an organization running regularly is very correlated with the drive to get better than everyone else.
New Emergent Ventures cohort.
I want to point out once again that people who believe in foom tend to be those from backgrounds like theoretical computer science, mathematics, or finance; engineers, business founders tend to be more skeptical. By this I don’t mean skeptical in the sense of AI skeptics, whose I feel tend to come from more or less entirely non-empirical fields and the humanities.
She also mentions Lisa Cook, so might as well also mention that Kyla Scanlon has an interview with Beth Hammack from the Cleveland Fed, Tahra Hoops has an article on the importance of Federal Reserve independence, as does Ray Dalio in an FT interview.