a lot of eugenicists in the early days, and then during the cold war when uni's got huge influxes of cash it was all from the DoD and related MIL, so the computing researchers who thrived were the ones who could stomach taking blood money. it arguably shaped that+later generations of researchers

after 4 years the "critical" tech community will be comprised of objectively uncritical computer scientists who "defiantly" use words like "ethical", "participatory", or "human-centered" in proposals to train hegemonic algorithmic systems, because that's the form & extent of their hashtag-resistance

the tech stack musk is using is incidental. computer scientists who happen to be interested in LLMs will make careers out of the next 4 years chasing the idiosyncrasies of the implementations of these systems, condescendingly dismissing the study of the political movement animating all of this shit

Can't fix institutional design problems with "literacy" solutions. See this often in AI policy: public tech education, awareness, training programs proposed as solutions to low trust in AI, tech companies, & regulators - without addressing why the public mistrusts those institutions to begin with

... the only solution right now would be to spend another full day on chat/phone support. Otherwise, just wait. They rolled this out without permission, and they HAVE AN INTERFACE FOR USERS TO DISABLE IT, they just aren't letting most users access it without jumping through endless hoops.

"You also have confirmed that unless users go through a similar set of support, that took me six hours today, there will never be any way to turn this off."

Google: "Yes."

He later followed up/said instead that there eventually will be a way (no timeline).

Congratulations to you and team