Bibliolater 📚 📜 🖋 @bibliolater@qoto.org

I got an email from someone asking for my "Essential Trigonometry for Calculus" cheat sheet, which I give to my calculus students. So, I thought I'd put it on GitHub for anyone to download (compiled PDF and source code). https://github.com/divisbyzero/Essential-Trig-for-Calculus

For all my academic colleagues who now work at the University of Bums on Seats - or if you are in the US and dont speak English, the University of Butts in Seats (they share BS in common)

https://futureu.education/uncategorized/commentary-by-a-j-lenardic-when-bums-on-seats-meets-misinformation/

@academicchatter #academics #universities #bumsonseats #buttsinseats #misinformation #pressreleases #advertising #sciencecommunication #science #astrobiology.

@bibliolater @ai I cannot for the life of me imagine why lawyers still do this. There was that Mata v. Avianca case back in 2023 where a lawyer got flamed by the court for doing it, and ok, I have a little sympathy for him because he probably sincerely thought that ChatGPT was a fancy natural-language search interface. Under those circumstances it was an understandable error, though it's NEVER ok to cite things without checking that the citation supports the proposition you're citing it for even if you're copying out of a treatise or another case, so the guy deserved what he got. However, one of the things he got was a vast amount of mainstream media exposure, and at this point this has come up so often that there's really no excuse for not knowing how bad an idea it is. There's a guy in France who's trying to keep track of all the times it's happened; his database is up to 139 entries, and I'm sure there are a lot of cases he doesn't know about.

https://www.damiencharlotin.com/hallucinations/

Towering tornadoes of plasma the size of planets dance over the Sun's surface in this timelapse from the Solar Dynamics Observatory spacecraft.

Credit: NASA GSFC/SDO

"AI companies claim their tools couldn't exist without training on copyrighted material. It turns out, they could — it's just really hard. To prove it, AI researchers trained a new model that's less powerful but much more ethical. That's because the LLM's dataset uses only public domain and openly licensed material."

tl;dr: If you use public domain data (i.e. you don't steal from authors and creators) you can train a LLM just as good as what was cutting edge a couple of years ago. What makes it difficult is curating the data, but once the data has been curated once, in principle everyone can use it without having to go through the painful part.
So the whole "we have to violate copyright and steal intellectual property" is (as everybody already knew) total BS.

https://www.engadget.com/ai/it-turns-out-you-can-train-ai-models-without-copyrighted-material-174016619.html?src=rss

Race, ethnicity don’t match genetic ancestry, according to a large U.S. study
https://www.science.org/content/article/race-ethnicity-don-t-match-genetic-ancestry-according-large-u-s-study
Data from the All of Us program confirm what many geneticists have long promoted

Full paper here https://www.sciencedirect.com/science/article/abs/pii/S0002929725001739 (paywalled)