David Sacks: Dario's AI safety push is a recipe for centralized control
David Sacks on X: Some thoughts on Dario's post
In a detailed response to Anthropic CEO Dario Amodei's post on AI regulation, David Sacks argues that Amodei misrepresents his critics, ignores the revolving door between AI companies and regulators, and that his proposed pre-release approval process would entrench incumbents like Anthropic while handicapping the US against China. Sacks contends that Anthropic's own business model depends on staying ahead of open models, making its safety advocacy self-serving, and warns that concentrating power in a federal bureaucracy and a few frontier labs is more dangerous than distributing AI capabilities.
Dario believes frontier AI is too powerful to distribute; we believe it is too powerful to centralize.
- speak_plainly
I agree with Sacks.
Amodei's conception of regulation is difficult to distinguish economically from some form of regulatory capture, protectionism, or rent-seeking in the face of limited infrastructure and resources and intense regional competition.
It doesn't make sense given that the US does not have a monopoly on AI development and regulation would put US companies at a disadvantage, especially given the US's hegemony is debatably potentially over. The idea of US using regulatory sway requires that the US can influence the entire ecosystem, from access to capital to semiconductor development to allied export controls. That world is probably over.
The US has lost its grip on all of these threads and China has picked up the slack. This is like trying to put the genie back in the bottle with one hand tied behind your back.
- Gormo
> Dario acknowledges that AI is structurally centralizing but attributes this mainly to chips and scaling laws.
Is that a correct idea to acknowledge in 2026, with the proliferation of open-weight models hitting an inflection point, and Nvidia and AMD both putting significant effort into developing local AI products at price points already accessible to prosumers and SMBs?
Is this "acknowledgement" actually a rear-guard attempt to protect centralization on the software side of the AI market via the exact sort of regulatory capture that he's hand-waving away, so he can use it to artificially prop up his inferencing-as-a-service business model?
- mikgp
“ I have repeatedly argued for strong antitrust enforcement to keep industries competitive, especially Big Tech.”
Has David Sacks actually called for stronger antitrust enforcement? Not that he has total control of this administration but it’s gone pretty hard in the other direction in my observation.