But they were peer-reviewed RCT, no? Nothing but Gold Standard Research, I was told!

When Trump's staff complain about "unelected judges" why don't any reporters ever remind or correct them that ALL federal judges are unelected? That's like me complaining about an "unhatched human". All humans are unhatched. We're from live births. That'd be insane. But reporters never say anything.

FAKE Science

Brandolini's Law: The amount of energy needed to refute bullshit is an order of magnitude bigger than that needed to produce it. In the age of LLMs "order of magnitude" may be grossly underestimating it.

I saw a great example of Brandolini's law when it took a 45 minute youtube video to successfully debunk Musk's 3 second claim of "I was on the 2nd best team in the world at Quake" that he said on Rogan, which is part of the way he became so beloved by gamers.

And that's only the first-order refutation of the claim itself. Then there's the second-order effects of dealing with everywhere that claim was (and continues to be) repeated or rehashed.

Yep. Because not only does Brandolini's Law apply to the length of the refutation, but also to the number of instances which it must be used to become effective. And "ain't no one got time for all that."

Notus - Doing Journalism. CBS, NBC, ABC, WaPo, NYT, etc.. all being stenographers.

This is what good journalism is. Especially when the subject of the reporting is known to straight up lie under oath. Every word he says should be fact checked.

All of this is courtesy of Republicans in the House & Senate. Without them, Trump is nothing. Republicans in Congress have to be made aware that in upcoming special elections & in 2026 they'll be held responsible for the consequences of this insanity. (Assuming there are elections.)

Where is this interview that you've screenshotted from? Is there a separate story where they explain their methods?

I don't know if it is available online, but it was in an email sent to NOTUS newsletter subscribers

Thank you. I'm not a subscriber, so that makes sense I wouldn't have it.

I too would like to read that.

So 98% of the of the citations were real that's an A right?

Know you are being facetious, but will point out that this is clearly a cursory search: no attempt was made to check if the cited papers that were not fabricated said what the report says they said which I do not have a high degree of confidence was rigorous when drafting the paper.

There was some of that, it's the part where they reached out to cited authors and were told the citations were real but were mischaracterized. I'm sure that's what took most of the 5 days! But also I'm sure it was nowhere near complete, doing it for every reference would be an even bigger ordeal.

idk sounds like a lot of work it's just so much easier to put on a Huberman podcast while I'm walking in the sunshine

Just it's such a common thing with sloppy writing, and if something like that happened with me I would feel a deep shame these people are incapable of. (This is from If Books Could Kill; would link transcript but it always tries to stream the audio.)

yes I know the deep shame after once submitting a paper in which I cited a result for general manifolds that was only proven in Euclidean space, couldn't imagine the Mariana Trench of shame if I had cited something completely made up

Could this be taken as a hint that "AI" is a complete crock of shit and shouldn't be considered useful in any way?

Written by AI?