Russell Steinthal
steintr.bsky.social
Russell Steinthal
@steintr.bsky.social
Antitrust lawyer, computer nerd, long-suffering Mets/Jets fan

Watching NYC area weather via WatchedSky: 🌩️👀 xEqdwWB2srp6
Thanks, and thanks for all the informative posting and commentary this year. You were one of the reasons I just gave to AIC when I was looking for an immigration-related organization to support as part of my year-end giving. Keep up the good fight!
January 1, 2026 at 5:45 AM
And second, if they don't, what are the downstream effects of having even more massively unbalanced state tax rates coupled w/ benefits that change dramatically based on your domicile? I could imagine effects on business location, individual interstate mobility, etc. that could be quite significant.
December 31, 2025 at 4:06 AM
More broadly, this seems like a big gamble on at least two fronts. First, it would be a moral gamble with the health/welfare/education of residents of red states. Left to their own devices are those states really going to enact (and fund) equivalents to SNAP, Medicaid, Chip, etc.?
December 31, 2025 at 4:06 AM
Is it actually politically easier, esp. in this age of geographic polarization? Maybe at one point there were red state Sens who would vote for a new program only if it would be state-administered. But how many of those are there now? Today, it feels like a sop that doesn't get (m)any votes.
December 31, 2025 at 4:06 AM
If you were to take all of the federal cost-share/state-run programs and make them entirely state-funded, that would seem to require dramatically higher state tax bases. (I assume we're talking about Medicaid, SNAP, Title 1 Education funds, etc.)
December 31, 2025 at 3:39 AM
Unless this is coupled with dramatically rebalancing the federal/state tax balance, how would that work? Is the idea that a future Trump-like Administration would be less likely to cutoff direct federal payments to child care centers in MN than to the state child care agency?
December 31, 2025 at 3:39 AM
I assume that's a situation in which the employer will end up bearing the risk of a "late" postmark?
December 31, 2025 at 2:04 AM
Unless a state's rule specifically refers to postmark dates (which aren't necessarily reliable for meters anyway), couldn't you still do that?
December 31, 2025 at 1:48 AM
For both the FRCP and NY CPLR, service by mail is complete upon "mailing," which I would historically (pre-ECF) have attested to in an affidavit of service as to the date and location it was deposited into the mail, whatever the postmark date.
December 31, 2025 at 1:48 AM
(Again, consider the electric utility hypothetical from the Cox v. Sony argument.)
December 30, 2025 at 3:34 AM
I assume it's easier from a technical perspective to build those relatively fine-grained guardrails into the consumer-facing service (e.g. ChatGPT) than the underlying model, but would be curious what the SMEs think. My point for now is just that the layer of the stack matters.
December 30, 2025 at 3:34 AM
To take Connor's BAM safe harbor, is the BAM of keeping the GP model from creating deep fakes stopping them from creating any images? If you're a large corp def't with the burden of proving BAM to a jury, how confident could you be in an approach based on preventing, e.g., NSFW or < age 18 images?
December 30, 2025 at 3:34 AM
Those people are already directly liable for possession/creation of the images themselves, so it's less of an issue? I think the hardest cases in this area are the creators *of* the general-purpose models, as applied to downstream users of those models (as opposed to their own services).
December 30, 2025 at 3:34 AM
(FWIW, I don't think that a law that prohibited the use, development, sale, etc. of AI image editing or generation tools would raise 1A concerns --- or at least not as many --- but perhaps tellingly, that's not what people are agitating for now.)
December 29, 2025 at 9:18 PM
A CSAM-specific exception to that rule might well mean that tool manufacturers need to simply prohibit *all* AI image editing or video generation, lest they be hit with draconian penaties for individual user's misuse. Perhaps that's a good tradeoff, but we should be explicit about it?
December 29, 2025 at 9:18 PM
That accident of technological evolution is a problem b/c existing law makes it difficult --- for what IMHO are, in general, good reasons --- to hold the makers of general purpose tools liable when they are used for unlawful purposes. (See the recent SCOTUS argument in Cox v. Sony for an example.)
December 29, 2025 at 9:18 PM
I do think this is part of the (legal) problem. A hypothetical sticksomeoneelsesfaceonporn.com service would be relatively easy to go after. But the technology actually emerged as a general purpose, very powerful image editor as part of an even more general-purpose AI model.
December 29, 2025 at 9:18 PM
Ah, that makes sense. (I think i just conflated your points about the novelty both of the style and of the vote count.)
December 29, 2025 at 5:47 PM
Of course, in that case the (narrower) concurrence would control, at least to the extent any of these opinions have precedential force (which question itself is frequently assessed in a subsequent case).
December 29, 2025 at 5:21 PM
Nice article as always. I'm curious, however, why you believe the styling of the Kavanaugh concurrence as "in the judgment" necessarily means that there were 5 other votes for the order. Can't a concurrence be styled that way even where there is only a plurality opinion? (See, e.g. Baake.)
December 29, 2025 at 5:21 PM
(Of course you should first have a theory of how it's relevant to your case. As all lawyers (should) know, it's unethical to use discovery to harass, burden, oppress, delay, etc.)
December 28, 2025 at 3:23 AM
No, but if you have an RFA or rog to spare under your cap, you could give them a bit of grief by seeking an admission as to their legal name. With this administration and its lawyers, there is a non-zero chance they would either lie or provoke a motion to compel rather than just admit the truth.
December 28, 2025 at 3:23 AM
Not a military/naval expert, but does the relative compactness/isolation of the Black Sea affect the ability to generalize from that example?
December 25, 2025 at 4:17 AM
Chicago (and a decent chunk of the N.D. Ill.) is clearly within the 100-mile zone, so unless it was intended as pure dicta, that sounds right to me as a logical matter. As you say, look at what he wrote, not what he might have thought. (Although citing Whren, of all cases ... is a choice.)
December 23, 2025 at 9:16 PM
Sure. But it's also possible that someone independently sent it, either as a prank, or to emphasize his relationship with Trump, or for some entirely other reason. You'd presumably have to do other investigation to figure out who sent it, why, etc. that we're not going to get from the posted image.
December 23, 2025 at 3:48 PM