2025-08-05
we are nowhere and it's now
OpenAI releases their open source model.
Harvey Lederman has an interesting post on the end of human greatness in a world dominated by AI. It’s something interesting to think about if you read about this recent Quanta Magazine article on Hannah Cairo1 and then consider the current rate of progress, including the 2025 IMO gold medals. Harvey correctly points out that most people already find themselves in a place where their work is not personally meaningful and feel their lives are largely out of their own control, dependent on more powerful agents who are not aligned with themselves. But assuming that AI is compute limited, I think that while it will be more or less impossible to receive fulfillment through becoming the best at anything, it should still be possible to achieve something like fulfillment through the pursuit of results which are not economically viable, akin to how some people might enjoy carving miniature figurines for their kids. For example, one justification for R&D in areas like mathematics is that eventually everything becomes applied, but if you assume sufficiently powerful AI, then for reasons of economic efficiency most R&D will probably be done as needed just-in-time; in which case, even if AI could trivially solve some mathematical problem, if you do it first you will still be the one cited for the result. Meaning will increasingly migrate out of economic usefulness into these more symbolic victories.
Scott Alexander notes that while people find great enjoyment and meaning from community, most people do not prioritize it for economic reasons. This will be another area where, if AI destroys our economic value, we may be able to derive meaning from non-economically viable activities.
Edit: Adam Mastroianni with more on this topic: he notes that LLMs do not compete with humans for status. The way I’m viewing things is that money represents our usefulness to each other, and LLMs rendering humans economically unviable would represent a future where humans can become perfectly self-sufficient, able to meet all their needs without relying on anyone else. The result would be hyperindividualization as never seen before, which would mean a substantial increase in hermits living out in the middle of nowhere. But also, since many (most?) people actually pursue self-actualization through community, the average connectedness would probably actually increase under that regime.
Alex Kesin on Novartis’ R&D lab NIBR, an example of R&D which is beneficial but shut down due to being economically unviable.
Tahra Hoops on the new bipartisan housing bill (ROAD to Housing Act of 2025).
Scott Sumner with misconceptions around tariff narratives.
The presentation slides she created for her results are nicely personalized.