| ▲ | qlte 10 hours ago | |
It's kinda surreal thinking back to 2023ish when OpenAI was discussed like it could be the final monopoly to end all monopolies. Altman was performatively discouraging investments because money might be worthless in a post AGI world (and still taken somewhat seriously). OpenAI promised to shut down for the benefit of humanity if they lost the lead (lol). LLaMA derived open models seemed like they could be stuck years behind the smallest SOTA model from OpenAI. Anthropic was seen as intentionally stymied by their honest commitment to safety and wouldn't recklessly try to leapfrog OpenAI in public releases to fuel a dangerous arms race. Seems like vibes started to change after Altman's return, then he started to spend less time on robot gods and instead lots of waffling about the definition of AGI being like, "us making a lot of money" or else maybe "does the job of a mediocre office worker". Anthropic catching up and, despite all their rhetoric/Dario's grievances about Sam, now appeared more or less indistinguishable from OpenAI. Deepseek was the nail in the coffin for any AI lab having a real moat or defensible IP. Both companies/leaders increasingly seemed more concerned with about IPO timing, ads, B2B sales and financial engineering while their obligatory mentions of singularity/extinction just felt forced and hard to take seriously. Things have definitely changed since the days when Sam Altman/OpenAI was the symbol for everything AI, and Anthropic existed only for alignment/safety. | ||