there was a post shared out there recentlyish (blogpost? substack? can't find it. enjoy vagueposting) that was about how ai companies have no clue what they're doing, and compared them to alchemy in that they also had no idea what they were doing, and over time moved on to more realistic goals, but while they had funding for these unrealistic goals they invented distillation and crystallization and black powder and shit. and the same for ai would be buildout of infra that can be presumably used for something later (citation needed).

so comparison of this entire ea/lw/openai milieu to alchemists is unjust for alchemists. alchemy has that benefit on its side that it was developed before scientific method was a proper thing, modern chatbot peddlers can't really claim that what they're doing is protoscience. what is similar is that alchemy and failed ea scifi writers claim that magic tech will get you similar things. cure for all disease (nanobots), immortality (cryonics or mind uploading or nanobots), infinite wisdom (chatbots), transformation of any matter at will (nanobots again), mind control derived from supreme rationality (ok this one comes from magic), synthetic life (implied by ai bioweapons, but also agi itself). when chinese alchemists figured out that mercury pills kill people and don't make them immortal, there was a shift to "inner alchemy" that is spiritual practices (mental tech). maybe eliezer &co are last alchemists (so far) and not first ai-safety-researchers