In the US, copyright law is created to serve the public interest--not the interest of the writers. AI requires titanic amounts of data to train. If companies are required to identify and pay every author, then America simply won't have AI--but China will. How is that in the public interest?

Why? The question of whether AI training is fair use or not is still pending in court, but it clearly doesn't cost writers anything, unlike real piracy.

www.ropesgray.com/en/insights/...

But what can they do?

So what can we do to stop it? (I'm in your district, for what it's worth, and I vote for you every two years.) :-)

The important question is whether the difference was statistically significant or not. If it's just a two-sigma result (or less), then it would have been better not to mention it. If it's more than 5 or 6, then it's a matter of concern.

Nah. The right word is "Braunhemden" aka "brownshirts."

The problem was that he lost what should have been a very winnable debate. He kept losing track of what Trump was saying. He seemed completely lost. A debate is a tough format; I'm sure he didn't have similar problems in policy meetings, but it was clear he just wasn't up to campaigning for the job.

I dunno. Letting them talk's not all bad; they tend to hang themselves with their own words. And we really ought to be able to defend our positions, although I agree that giving a platform to people who do nothing but lie would be tedious

Also DeSantis would be very focused on DEI and (hopefully) not so much on firing all generals who won't swear allegiance to Trump over the Constitution.

Actually I think the expectation is that SLS/Orion/Gateway will all get cancelled completely, discarding even working hardware. The key thing is to dismiss the workforce, which costs something like $10 billion/year and redirect those funds to something useful.

I don't know about "imbeciles." I thought it was "suckers and losers." :-)

I wouldn't say that's "common sense." Common sense would say that people behave better when they can be held accountable. But the studies you cite show that expecting the general public to hold them accountable doesn't work. Having the site owners do it worked best.

Not sure I follow you. The links you provided showed that 1) if users were totally anonymous, the result was terrible. 2) When they were known to the site but not broadly, they established stable pseudonyms and behaved much better 3) when they were forced to use real names, they got worse than #2.

I think they key thing is that they can't be anonymous *to the site.* Otherwise there's no way to ban bad actors. But, as others have indicated, as long as everyone is "known" to the site, they're better behaved even though they're using pseudonyms with other users.

That's an interesting study! I was going to suggest that it might be enough that people be known to the site without requiring they use real names in public, but I ran past my 300 characters! :-) It's nice to see someone actually tested this.

Assuming we agree that carbon capture is never going to be a significant factor, I'm thinking warming is going to end up at about 2.5 degrees. I wonder if anyone has a good idea of what mitigation is going to cost?