Elon Musk says that, “all things considered, I think California should probably pass the SB 1047 AI safety bill”, and saying so is “a tough call”
Always B — Be C — Ctalking your book Alex Konrad / @alexrkonrad : this will create some interesting divided loyalties in Silicon Valley 👀 Sean Durkin / @seandurkinsf : @DanHendrycks @elonmusk Elon's AI training facilities are in Texas, which would not be subject to the regulation. [image] Eliezer Yudkowsky / @esyudkowsky : @elonmusk Honestly I haven't studied SB 1047 hard enough — I don't think it saves the world by itself, so the main impact is political consequences on which I'm not expert. But given how it selectively hits big AI companies like yours, this looks to me like an act of integrity by you. 🫡 @basedbeffjezos : 1) Move X / X AI out of CA 2) Build AI superclusters in TX/TN 3) Support CA bill that nukes his competition 4) profit. High-dim ♟️? @daniellefong : my feeling has been and continues to be that i am lukewarm against SB 1047 at about the same level that Anthropic voiced support for it: “i am not certain of this. there are still some aspects of the bill which seem concerning or ambiguous. there are some arguments for it. @basedbeffjezos : @elonmusk AI's can be powerful but authoritarian govt with shadow control over powerful AI's is scarier. Sets an awful precedent at a national level + will open up the door for govt lawfare against companies that don't offer govt LLM backdoors. Like the Telegram situation but for AI. Gary Marcus / @garymarcus : I have many differences with Elon but am very glad to see this tweet.
@elonmuskElon Musk
Context & Ripple Effects
SB 1047 had already become a fault line between its backers and critics: Sen. Scott Wiener defended it against claims that it would disadvantage smaller rivals and open-source work in an earlier defense of the bill. A subsequent roundup documented organized opposition from researchers, companies and academic institutions across the AI ecosystem.
Musk's qualified endorsement matters because it adds a prominent AI builder to the pro-bill side while the discussion itself raises questions about whether the rule's geographic scope creates uneven exposure. Related coverage later showed that concerns about a development “chilling effect” remained central to the political decision. California's governor echoed those concerns
First-order effects
Musk publicly gives SB 1047 a high-profile, if qualified, endorsement, strengthening the bill's safety-focused coalition in the immediate policy debate.
His stance invites closer scrutiny of the claim that his AI training operations sit outside California's reach, making perceived regulatory asymmetry part of the response to his endorsement.
Second-order effects
Opponents can sharpen their argument that the bill burdens California-based developers unevenly, while proponents gain a rebuttal to the idea that major AI companies uniformly reject safety obligations.
The debate shifts from whether safety rules are needed to which developers, models and locations should be covered—an issue that affects how firms assess where to build and train systems.
Third-order effects
If prominent AI executives increasingly support some safeguards while contesting their scope, AI governance may evolve through negotiated thresholds and jurisdictional boundaries rather than a simple industry-versus-regulator split.
The lasting test for state AI rules will be whether they can claim legitimacy across large labs, smaller developers and open-source communities without reinforcing location-based regulatory arbitrage.
The trend: SB 1047 is one instance of AI governance moving toward rules whose credibility depends as much on who they cover as on the safety goals they declare.
Guy whose competitors are all in CA interested in passing bad law that would cripple their businesses. Huge if true. What's weird is that this guy is in the AI business at all after demanding everybody stop AI development because it's dangerous.
Only “unexpectedly” if one does not understand how Musk works. Musk moved much of his business concerns out of the State of California. Musk wants regulation to exist in California, so that he can exploit the lack of regulation elsewhere - as he does with all of his firms.
@elonmusk I am pretty sure Grok-2 wouldn't have caught up to SOTA models without open-source models and techniques SB-1047 will materially hurt xAI, so why support it?
Asked Sen. Wiener how confident he is that 1047 will be signed into law, he said he can't predict but Gov. Newsom “has made statements that are very consistent with my own thinking - that regulation could be appropriate here...” The latest: https://finance.yahoo.com/...
Not a huge surprise since bill author & champion @DanHendrycks now advises Elon. Can't imagine this moves any D votes for the bill though. At this point the bill either passes and is vetoed, or dies in #CALeg if Newsom threatens veto
Elon Musk says California should “probably pass” AI safety bill SB 1047. Anthropic has cautiously supported the bill, while OpenAI and many other prominent tech leaders have opposed.
Update: reflecting on amended SB 1047, I believe it is in an odd place that makes it close to a no-op. My fundamental position is that AGI progress has been stalled for many years, ARC Prize is evidence of this, and that we need new ideas to discover AGI. Unfortunately, AGI
Someone finished building a datacenter outside of CA for his closed models and now wants regulatory restrictions on his competition and open models. Follow the incentives.
@elonmusk Hey Elon, you know those consent orders that the @FTC is using to harass you? SB 1047 looks just like them, and could be misused the same way: [image]
Huge and somewhat surprising SB 1047 endorsement. Elon Musk has always been concerned about AI safety (super powerful systems) while extremely impatient with “AI safety” (censoring existing systems).
For anyone deploying AI, you know how much red tape there already is. Adding new levels of material liability for the underlying models only reduces the amount of progress and utility we get. Regulation should primarily be aimed at the applications of AI, less on the tech.
you gotta really appreciate the chutzpah of a guy who has spent the last decade effectively evading NHTSA bureaucrats on AV regs declaring that he's a long-time advocate of AI safety. 😂 Musk has also repeatedly called AI an existential threat to humanity while simultaneously
The unintended consequences of this bill in many years from now would likely be quite bad. The amount of risk that model providers take on would almost necessarily slow down the rate of AI progress and reduce competition overall. And as we know with regulation, it's often a one…
i dont have an opinion either way on this bill but it's pretty clear Musk can see some benefit in the AI race if this passes re: competitors headquartered and building infrastructure in CA and Musk hightailing it out of the state A — Always B — Be C — Ctalking your book
@elonmusk Honestly I haven't studied SB 1047 hard enough — I don't think it saves the world by itself, so the main impact is political consequences on which I'm not expert. But given how it selectively hits big AI companies like yours, this looks to me like an act of integrity by…
my feeling has been and continues to be that i am lukewarm against SB 1047 at about the same level that Anthropic voiced support for it: “i am not certain of this. there are still some aspects of the bill which seem concerning or ambiguous. there are some arguments for it.
@elonmusk AI's can be powerful but authoritarian govt with shadow control over powerful AI's is scarier. Sets an awful precedent at a national level + will open up the door for govt lawfare against companies that don't offer govt LLM backdoors. Like the Telegram situation but for…