2025-04-06
unsubstantiated
Owl Posting on the lack of success for AI pathology companies. It’s another reminder that AI will probably take a while to diffuse throughout the economy, given physical constraints on the generation of training data and potential costs of integrating the AI within an existing process. The other interesting thing is the question I had of why biotech startups always try to own the entire stack, when the service model has been proven by SaaS software to be a generally more efficient use of resources, which maintains your competitive advantage while also aiding the buildup of an ecosystem around you. This post seems to be a partial answer to that question, that it does seem to be due to habits from the broader industry, and something due for a change.
Asimov Press on the under-celebrated history of vaccine adjuvants.
Alex Honnold interviews Janja Garnbret. There are a few different ways that people who don’t like talking about general intelligence covertly ask about it, and one of them is whether people who are at the top of their field would have also succeeded in others. It seems to me that, more often than not, these are people who were just good at everything when they were younger (Magnus Carlsen is another example). That they are often one-dimensional is because they made the choice, faced with the prospect of becoming merely good, to dedicate their life to a single pursuit. I’m starting to believe that what separates these two decisions is how much intrinsic motivation they can bring to a task: if your drive is high enough, then a relentless and repetitive training schedule can feel fun and enjoyable. People like me can’t do that, so that’s why we’re constantly chasing novelty.
Max Nussenbaum on Ross Douthat and faith. Recently, I’ve found it useful to define one’s strength of faith as the difference between one’s justified and actual level of belief. Under this lens, efforts such as Douthat’s feel misguided because they actually decrease the average level of faith, because increasing the evidence floor only converts those who previously felt the level of unjustified belief required was too high. This is a self-perpetuating system, because it is the rationally-minded who have this trait, and as they continue to look for evidence that probably does not exist, the key pillar holding everything up becomes belief liquidity, which is subject to Ponzi dynamics and can collapse on a dime.
