GPU utilization and active infrastructure remain high for now.
This is not an AI pause. It's the government thrashing around after realizing that the people who told them AI was NBD were wrong, and being severely understaffed.
AI development is not even close to stopped RN + this is not what you'd do if you wanted it to stop
Many users criticized AI policies favoring safe deployment as nefarious pay-to-play tactics or a boogeyman that stifle open progress and favor insiders over real talent.
No Digg Deeper questions have been answered for this story yet.
The understaffed thing is totally solvable, btw, but it would take someone at the level of Trump/Vance/Wiles to make a decision to ease up on loyalty/ideology/affiliation based vetting (see e.g. Collin Burns lasting a few days simply because he used to work at Anthropic)
It is bad for AI *deployment,* which is different from development.
On a longer timescale (months, not weeks), yes, that will hurt funding + slow compute purchases etc. + slow development. TBD if it gets sorted before then.
For now, rest assured the GPUs are humming
It is bad for AI *deployment,* which is different from development.
On a longer timescale (months, not weeks), yes, that will hurt funding + slow compute purchases etc. + slow development. TBD if it gets sorted before then.
For now, rest assured the GPUs are humming
This is not an AI pause. It's the government thrashing around after realizing that the people who told them AI was NBD were wrong, and being severely understaffed.
AI development is not even close to stopped RN + this is not what you'd do if you wanted it to stop
@AndrewCurran_ But people are so used to race v pause that they'll read everything on that valence when this is clearly a fight on a different axis within Race
@AndrewCurran_ My view is the Good outcome is in ten years civilians are still only a couple gens behind frontier. Civies cant be frontier in 2036 jesus christ think about what that would mean
Yes we obviously need to slow down open source deployment as "frontier" becomes more capable! Eventually we're going to try to actually kill it but import controls will probably come first. And similarly export controls are inevitable because they'll permit more frontier adjacent US deployment because while "citizens + select others" arent a safe pool to deploy in they clearly are a different risk / reward tradeoff than "everyone".
@AndrewCurran_ I hope to see massive annual improvements in the capabilities i control and also never be able to hack my bank.
@AndrewCurran_ Which again makes sense. You really think you and your bank, or you and the drone network, should have the same coding capabilities guys? Really? You absolutely sure about that?
@AndrewCurran_ I really feel like everyone else is just kind of being insane here
If we freeze the civilian frontier at 2026 i agree *thats* bad, but the super hacker is like 4 months old here i think its super premature to say thats what's what we're doing and doesnt really make sense given open source anyway which we arent (yet) trying in earnest to crush.
We're clearly just trying to create some defender / attacker separation in a sort of clumsy way because most fast time pressured approaches will not be perfectly surgical.
@Miles_Brundage If anything they've incentivized the labs to keep their most explosive progress to themselves
@AndrewCurran_ Open source was a useful boogeyman to milk for incentives to lock in a "Race" mechanic when Luddism had a shot but its just about get real o'clock
@Miles_Brundage This is way more nefarious than that, man. This is pay to play gating using government to edge out people who aren't paying the people at the helm.
@Miles_Brundage the smartest people on AI chose to build it, not regulate it. the government didn't lose a staffing war, it lost a talent war.