Critics fear arbitrary White House authority over frontier model access.
I used to give the current admin a nod for saying their AI policies are fairly reasonable (chips stuff far messier).
In the last ~3 months they've destroyed all that trust and dug a giant hole of uncertainty - where the US's AI leadership will now be actively degrading.
Many users blame the US Administration's erratic reversals of AI model policies for creating unpredictable chaos that deters investment and erodes trust, often accusing incompetence or cronyism.
No Digg Deeper questions have been answered for this story yet.
Actually it's potentially worse.
We're in a major step change & acceleration in progress and the admin just did a complete 180 in policy on models.
It would've been better to have no policy before. Now, the policy has changed to a vibe check & we have no idea what comes next.
I used to give the current admin a nod for saying their AI policies are fairly reasonable (chips stuff far messier).
In the last ~3 months they've destroyed all that trust and dug a giant hole of uncertainty - where the US's AI leadership will now be actively degrading.
@natolambert blame anthropic
@natolambert USG was extremely pro open source and competition. I loved this about them. I don’t know what happened.
@the_dismal_tide there are different blames for different people
@natolambert @AtaeiMe USG should have resisted the China military distillation narrative / pause urgency (obviously what’s under this per reporting on SK Telecom) but the party relentlessly misleading our military and the American people is way, way, way more culpable
@natolambert Read Situational Awareness and AI2027. The government is not going to let the public experiment with what it believes is an important military technology before it does anymore.
@natolambert @AtaeiMe Such as that DeepSeek doing 150k calls of judge inference constitutes a “distillation attack” (obviously does not qualify even under the most expansive notion) or most recently that Alibaba’s alleged use of transfer station traces was linked to PLA (almost certainly not)
@natolambert @AtaeiMe I agree that USG both should have done more work upfront to understand AI and in any case asking OpenAI to not release was a huge mistake (unless it only did so to avoid a perception of bias against Anthropic). But Anthropic has been relentlessly making false insinuations
@natolambert Yes but you can't separate their policy from Anthropic's constant fearmongering
@teroterotero @natolambert Well, Macron won't allow that...
To all those calling for the open-ended politicization of infrastructural access: you got what you asked for.
Our new AI policy is that the White House decides ad hoc, for whatever reasons it likes, who does and does not get access to frontier intelligence. This seems rather maximally terrible.
@natolambert @JoshPurtell Do you think you can’t find most of the vulnerabilities using open models, even ones older than GLM? The cybersecurity risks are clearly overstated. And even if they aren’t, defense is much easier than offense, so withholding these models could actually hurt national security.
@AtaeiMe @natolambert Anthropic happened
@natolambert How is this not literally
@natolambert Why can't you understand that this technology is now capable enough to disrupt power structures? When you realize that, it's not a surprise at all that our institutions are starting to regulate more and more.
@natolambert Do you think Mythos legitimately spooked them?
@natolambert degrading? Google, Anthropic keep expanding US infra right now. companies' capex choices say more than admin statements.
@natolambert Bad policy you can adapt to. Chaos that shifts monthly? No one invests in that. The money just goes to Singapore and London.
@AtaeiMe @natolambert sacks out as AI czar, bessent has a weird hardon for AI regulation and seems to be calling a lot of the shots
@natolambert With the new OpenAI restrictions, 2 of the US’ top 2 AI companies are now throttled.
