Alan Rozenshtein

Assoc prof @UofMNLawSchool. Research dir. & senior ed. @lawfare. Nonres. @BrookingsInst. Term @CFR_org. Former @TheJusticeDept. Mr. @HNeprash.

[Alan Rozenshtein @ARozenshtein]

@ahall_research Even more depressing answers about law…

[Kevin Frazier @KevinTFrazier]

Don’t miss this @scaling_laws episode with @AlexBores.

[Alan Rozenshtein @ARozenshtein]

@curl_justin @jasminewsun I suspect it's mostly (1) laundered through (sometimes legitimate concerns about) (2).

[Jeffrey Sachs @JeffreyASachs]

Anti-con discrimination in academia is real. Part of it is a self-selection problem in the pipeline (ie "I have better things to do with my time than deal with a hostile grad program,") but part is also surely active discrimination in hiring, grad admissions, publication. 1/

[Alan Rozenshtein @ARozenshtein]

@Shalev_Roisman Sure but what is Canada supposed to do?

[Alan Rozenshtein @ARozenshtein]

It's wild how much everyone (including me) panics when Fable is unavailable even for a few minutes!

[Ilan Wurman @ilan_wurman]

Congrats to @MinnesotaLawRev for ranking very highly (#12) on this year’s combined impact score for law journals, under the Washington & Lee ranking. Very well deserved.

[Alan Rozenshtein @ARozenshtein]

@Miles_Brundage Miles.

[Kevin Frazier @KevinTFrazier]

.@ARozenshtein and I have had the privilege of talking with folks across the AI policy spectrum via @scaling_laws.

Alan is a heck of a lot smarter than me, but we always learn something from our guests.

We agree that this episode with @AndyMasley was especially informative.

Giddy up for a deep dive into all things data center.

[Alan Rozenshtein @ARozenshtein]

@Miles_Brundage Deeply meh

[Alex Stamos @alexstamos]

A lot to unpack here. Anthropic is burying some hard truths in careful political language. Some initial reads:

1) Anthropic verifies that none of the jailbreaks provided a capability beyond what many other models, including Chinese models, could do.

[Alan Rozenshtein @ARozenshtein]

In Chatrie, the Court did what it should have done in Carpenter: find that there was a Fourth Amendment search but be flexible on what counts as valid predication. https://t.co/tW0wPRYC6R

[Alan Rozenshtein @ARozenshtein]

@AndyJungTech @deanwball It’s actually not a decision at all, since the 9th circuit opinion was withdrawn when the case went en banc.

[Alan Rozenshtein @ARozenshtein]

It is time to get very serious about listeners' rights. That's the only way forward that durably limits government power while not making the frontier labs legally omnipotent.

[Dean W. Ball @deanwball]

The most important legal questions in AI right now all relate to the First Amendment. What are the best fact patterns to demonstrate that the creation, distribution, and *use* of frontier AI is a form of protected expression? Who, outside the labs, has standing to bring such suits? We need to move beyond 'code is speech' copium, and beyond the impulse to post into the void. Courts will be where the issues of the last two weeks ultimately get decided. It's not going to be easy, given the national security implications, but also, the underlying technology is a large *language* model, and this should count for quite a bit indeed. The best legal minds of our time should be stewing over these and many related questions.

[Alan Rozenshtein @ARozenshtein]

@deredleritt3r @lu_sichu Does Congress even remember how to do that anymore?

[Alan Rozenshtein @ARozenshtein]

@deredleritt3r @lu_sichu But on what authority? I don't see how export controls could apply to model weights that are literally already published.

[Jessica Tillipman @JTillipman]

The new AI executive order says its frontier-model framework is voluntary and goes out of its way to disclaim a licensing regime. But when your customer is the federal government, "voluntary" looks different.

Public-sector AI is not commercial AI, and the administration is increasingly governing the two in opposite directions.

My latest @lawfare piece, on how a "voluntary" framework becomes a condition of doing business with the government. Link 👇

[Alan Rozenshtein @ARozenshtein]

Here's the lawsuit filed to overturn the Mythos/Fable export control order. The strongest argument is that export control law simply doesn't cover API access. https://t.co/PcMNMKYBAJ

[Alan Rozenshtein @ARozenshtein]

https://t.co/0z4LfcTt3a

[Alan Rozenshtein @ARozenshtein]

.@robertwrighter has written a terrific new book about AI. I got to talk with him about it on @scaling_laws.