Ari Cohn

Everybody involved loses everything is the only acceptable result.

Rep. Dan Crenshaw says culture of misinformation fueled his primary loss: "The truth didn't matter" The fourth-term congressman, who lost decisively to state Rep. Steve Toth, said baseless attacks about his alleged insider trading and gun stances fueled the upset. BY GABBY BIRENBAUM MARCH 6, 2026, 4:25 P.M. CENTRAL REPUBLISH SHARE

"Fire hurts," complains arsonist

www.texastribune.org/2026/03/06/d...

I don't think the speech here would give rise to liability in any other circumstances

Like, my concern here is not that I think OpenAI (or any company) is a "good actor." My concern is that the knock-on effects of this kind of liability are *enormous* and hit in ways people don't expect (when it's too late), and that's why the courts have refused to go there.

bsky.app/profile/aric...

No, we haven't

I dunno have you seen the people running it?

We might not disagree on what guardrails are reasonable. Just on whether or not government authority can impose them.

The law and what's right are two separate inquiries.

Go after the devs for saying the outlandish shit they say to hype their products. I don't really have a problem with that. But it's a different thing with meaningfully different implications.

(Also not remotely what's happening in this case)

An LLM can't file things in court.

I think in itself that is an overstatement of those cases, because that's actually not what the facts were.

Paging @jmiers230.bsky.social

Like people haven't been trying for decades/nearly a century lol

I'm not just making shit up! There is a lot of established, well-developed jurisprudence on this stuff.

We had AIM away messages and we did just fine with those.

It's OK to be on the side of users. It's just not Ok to impose government authority to baby-proof the world.

No, really it's not. It's arguments that have literally been made and are not meaningfully different, doctrinally or logically.

And yes, I think it's stupid to get the advice of a lawyer,and then decide to listen to ChatGPT instead. That's inherently unreasonable activity.

"Cooler and sassier kids" should be a parenting metric

Actually maybe incorrect. It's a realistic possibility that I'd be too drunk to post.

There's no legal duty to protect against the effects of words and ideas

So are books and movies and video games and comic books and

Not incorrect

Ah to be in the throes of puberty in like 1999.

Also: bsky.app/profile/masn...

No reasonable person fires a lawyer because of an answer they specifically solicited from an LLM, and nobody should (or does) have to build an LLM to protect idiots (or others) from their own unreasonable actions.

Tort law does not exist to dipshit-proof life.

Blowhard ChatGPT bot posed as lawyer, convinced woman to fire her real attorney — while citing phony ‘case law’: suit The AI-powered chatbot urged Graciela Dela Torre, of Des Plaines, to reopen her already-settled disability case and file dozens of illegitimate motions against her employer, Nippon Life Insurance C…

At some point we're just going to have to be comfortable calling the terminally stupid what they are.

nypost.com/2026/03/06/u...

That's beyond the scope of anything I'm qualified to opine on/can give any smart ideas for

This went even grosser places than intended

That's an entirely different (and valid) question

By bullying Anthropic, the Pentagon is violating the First Amendment. Here’s why. Government pressure on Anthropic raises a core question: can Washington punish an AI company for refusing to redesign its systems for the Pentagon?

Whether you love AI or think it's overhyped satanic sludge, we should all agree on one thing: government should never be allowed to control it, or force developers to build it in ways that violate their conscience:

www.fire.org/news/bullyin...

bsky.app/profile/aric...

Will Corey Lewandowski still be Noem's chief staff?