Comment Re:Oh noes.... (Score 1) 38
Comment Re:Sorry wokers... (Score 1) 166
Wait... are you planning on having sex with it?!?
Comment Re:+5 funny story (Score 1) 166
Can't tell if we should moderate the article as Troll or Flamebait.
Commenters have already pointed out the dishonesty of the prose based on their own published stats.
"You are a poor scientist Dr. Venkman."
-Dean Yeager
Yes, I'm enabling the clickbait by commenting.
Comment They're looking at the wrong "jobs" (Score 1) 166
Comment Re:Probably due to methodical errors (Score 2) 166
What's with that lighting? They didn't adjust for the fact that the woman looks slightly more like a serial killer, with a wrinkled skirt pulled up to her ribcage. I haven't seen graphics that bad since Metaverse. The study noticed it, but didn't adjust for it. Relevant quote:
"Johanna was perceived as less human-like than Johan"
Because the graphics were terrible.
Comment Re:As someone who likes Rust (Score 1) 54
Comment That begs the question... (Score 1) 54
Comment Re:Fuck it. (Score 1) 166
Comment Re:Probably due to methodical errors (Score 1) 166
Comment Re:Who is paying AI? (Score 1) 166
Comment Long Road (Score 1) 54
JPEG-XL is a good format but I just realized yesterday that Google Docs on Android doesn't support inserting a webp image a decade after Google was pushing it. I had one from a website and needed to run it through imagemagick first.
Android supports it, all the apps using android image primitives supoort it, but one of Google's flagship apps still doesn't.
We used move much faster thirty years ago. If an app supported say GIF, PNM, and TGA when JPEG came out it was supported in the next release.
Hopefully this round will be better. It's probably a management problem
Comment Re:mostly mash-ups of existing techniques (Score 4, Insightful) 105
It doesn't really matter if it makes a useful contribution to the field.
To which field? Math? Not even close. Data theft? absolutely.
If the mathematicians ever used ChatGPT to talk about math, then that is the source of the results. LLMs, by definition, cannot usefully contribute to anything that requires more than rearranging existing data. Rearranging existing data is their only function.
Comment Re:Problems so useless nobody cared? (Score 1) 105
Pure mathematics...but sometimes lead to unexpected discoveries that do turn out to be useful, and rarely turn out to transform entire scientific fields.
It's irritating when that happens. Ruins the fun. Then I have to go research something different.