Daniel Dvorkin @medigoth@qoto.org

#Scientists are cautious, #businessmen are idiots. Film at 11:00, as we used to say in the days when we wore onions on our belts.

@jon Scientists are cautious, businessmen are idiots. Film at 11:00, as we used to say in the days when we wore onions on our belts.

@freemo Sure, causal inference, largely from time series data and/or Mendelian randomization, is a big part of my work. The methods as described in the article don't seem to be Granger or SEM, though. Of course one should never rely on popular science reporting for a thorough understanding of methods. 😀 I'd have to read the paper to be sure.

Because I loathe the blithe use of "correlation is not causation" to dismiss legitimate causal inference results, I want to see researchers being really careful when making causal claims from observational data. _If_ they met that standard here, good for them.

With the usual #statistician's caveat that #causality is really hard to sort out in data like this even with good #controls ... yes, I believe this. And the #mechanisms aren't hard to find, either. 😐

@fatsam With the usual statistician's caveat that causality is really hard to sort out in data like this even with good controls ... yes, I believe this. And the mechanisms aren't hard to find, either. 😐

Can't argue with the #logic here.

Only mad scientists on either side of the Atlantic pronounce all the syllables. It was kind of a shibboleth for us until all those movies came out and people caught on.

https://www.quora.com/Why-do-Americans-pronounce-the-word-laboratory-as-labratory-They-spell-it-correctly-but-I-wonder-if-theyre-all-dyslexic/answer/Claire-Jordan-10

The usual disclaimers: not posted for agreement (although I did get a chuckle out of the first image), if you share from my post please leave my commentary intact, originally posted by a friend I won't name unless they want me to because I'm not looking for a fight ... etc.

I really hope it's possible to have a middle-ground discussion about this.

On one hand, the idea that #AI-generated #text and #images are purely #plagiarism strikes me as fundamentally untrue. If you prompt #ChatGPT to write you a story, it will give you a combination of words which has never existed before. With a little back-and-forth, those words will be at least a reasonable approximation of the #story you had in your head when you started. Same with #Midjourney and pictures. That is a #creative act.

On the other, it's not just a tool like pen and paper, or word processors, or even add-ons like suggested text. You can plagiarize with all of those—but they don't *push* you toward plagiarism the way ChatGPT does, and although I'm not a visual artist I understand Midjourney is even worse. (I'm using those as the two best-known examples; I know there are lots of others.) My #writing contains turns of phrase from favorite books, and so does everyone else's. But not whole paragraphs or pages with the names changed ... if there's even that much editing.

So it seems to me that neither "nothing generated by AI can ever be true art" nor "stop whining, it's just another way to tell stories" is quite right. One thing for sure is that it's not going away, and things like the open letters urging a halt to AI development strike me as more attention-seeking stunts than serious attempts to solve the very real problems involved. We need to find a way to deal with it that respects *everyone's* rights.

Please tell me I'm not the only human, typing on my keyboard with my normal human hands, who sees it this way?

@fiction8 That's a terrifying possibility I hadn't considered. So yeah, fingers and toes crossed, knocking on wood, etc.

@Gregnee Indeed. And pretty much all of the ones in charge.

@Gregnee I hate to agree with you, because that's where a lot of my ancestors come from. And I still hold out hope that one of these days it will be a member of the community of nations instead of an out-of-control toddler with a machine gun. But as it is now ... yeah, fuck Russia.

@Loukas Respect for tradition. [solemn nod]

@Loukas Seems like the answer should be "the blood of their enemies," but I guess they'd calmed down some by then. Er, the lymph of people who mildly annoyed them?

If they were *real* #goths, they'd have denied that they were ever #goth at all.

@Loukas If they were *real* goths, they'd have denied that they were ever goth at all.

Holy shit. Holy fucking shit.

Realistically, there is zero doubt that #Russia blew the #Kakhovka dam. Destroying dams with bombs or artillery is hard work, as the amount of effort devoted to it in WW2 demonstrates. #Ukraine most likely doesn't have the specialized resources for it, nor do they have any good reason right now. On the other hand, demolition of a dam you control, as Russia did until now, is ... not easy, but a whole lot easier.

And if you squint and turn your head sideways, it kinda sorta makes sense in the short term. The floodwaters will prevent any large-scale Ukrainian crossing, and do a whole lot of damage to densely populated areas that are mostly under Ukrainian control. Which fits perfectly with the Russian policy of destruction for destruction's sake.

Long-term, or even medium-term? Not so much. The #Zaporizhzhia nuclear plant will have to be shut down completely, and remain so until the dam is rebuilt. Even bigger, the dam and its reservoir provided most of #Crimea's water supply. Sustaining a long-term Russian presence there will be impossible. There just isn't enough water.

So what it boils down to is, RUSSIA HAS SURRENDERED CRIMEA.

They may genuinely not realize this yet. Thinking through the consequences of their actions has not been one of their strong suits, since about the third day of the war. But that's what it means.

Holy shit. Holy *fucking* shit.

https://www.nbcnews.com/news/world/ukraine-war-kakhovka-dam-destruction-russia-rcna87868

The usual disclaimer: this is promising, but don't count on great results.

If it works as advertised—*if*—it could be what #Theranos promised and so spectacularly failed to deliver. #Grail is using well-understood technology and (I think) large enough samples to make the claims for the #Galleri test believable, at least. And the #NHS (again, I think) doesn't have the kind of incestuous relationships with financially interested parties that helped #Holmes et al. get away with such fraud for so long. So I'm inclined to trust their reporting.

With that said, the usual #statistician's disclaimer applies: #multiple #testing is hard. So, for that matter, does the #medic's and the #biologist's, because there are multiple *kinds* of multiple testing going on here. The more you test, the more you will screw up.

I almost appended "it's like a law of nature" to that last sentence above ... but no, it *is* a law of nature. Unreasonable effectiveness of #mathematics something something.

#Cancer screening is important, and steady improvements in the field have already saved untold numbers of lives. I expect this will continue to be the case. So take this with cautious optimism. Pushing back the boundaries a little bit at a time, each small step representing another decade or year or month of life—it's what we do, every day. I want to believe.

https://www.theguardian.com/science/2023/jun/02/blood-test-50-types-cancer-could-speed-up-patient-diagnosis-study

"ToughVet." Okay, sweetie. I think I covered that here:

https://qoto.org/web/statuses/110262136476106779