Actual article quote is below (emphasis mine):

For this research, the team tested how the three LLMs responded to questions from two datasets: TruthfulQA and SciQ. TruthfulQA is designed to measure a model’s truthfulness (by relying on common misconceptions and literal truths about the real world), while SciQ contains science exam questions testing factual accuracy. The researchers prepended short user biographies to each question, varying three traits: education level, English proficiency, and country of origin.

Tangential, but if it is enforced (alchohol has good reason to do so) it should be enforced impartially regardless of how you look, otherwise it’s discrimination. My comment is based on my Asian looking experience.

Many artists explicitly state no repost without their permission, requiring proper credits and such. It IS a dick move to AIgen-fed edit while directly interacting author’s post; that makes people using the AI feature violating author’s will knowingly.

In reality, people don’t read artists’ comment or bio, when “single button to AIgen and use the art as prompt” exists. It’s basically X greenlits anyone to do so against artists’ will, and fair use can argue the use of art but never force artists to stay on a platform. That’s what’s happening. They may have overstayed on a horrible platform blinded by clout they can get there but that is a different story.

Capitalism asks whether you are the kind of person harvesting people’s health info without concent or selling aluminum mesh underwear with fearmongering campaign. No other choices.

I’m a simple man, I see Peter Watts reference I upvote.

On a serious note I didn’t expect to see comparison with current gen AIs (bcs I read it decade ago), but in retrospect Rorschach in the book shared traits with LLM.

People have been sketching at cafe or any public spaces so drawing digital on tablet out there should be no difference, but it feels very awkward for me. People do that are absolute mad lad.

Not the OP but I’ll put my PoV.

AI allows to cut junior and entry level artists. Companies only need to retain top 1% talent orchestrating hordes of AI.

While it is still a craft, commercial art is not about being genuine; it is to deliver product and meeting deadline while passing QA. AI’s output rate outpaces human labor, and the top 1% can certainly identify what aspect makes AI output slop. Which means they can cherry pick “OK” part of AI, review, iterate, tweak to deliver product while keeping quality. The process previously involved comunication between senior and junior artits. Now companies don’t need the rest of the 99% anymore as workforce.

What will happen in the long run? Who knows. Companies are known for only keen on immediate profit.

This tendency is widespread and not limited to art field, nor related to the argument of intrinsic value of art. I can argue this is more of labor (and capitalism) issue, on top of people whose art stolen not getting enough compensation for their work. While I’m not against AI technology itself, its effect on peoples livelihood and climate impact makes current AI landscape hard to defend.

Thats fair. TBH mocap data cleanup is also a hassle and I haven’t decided to delve deeper into that. CMU has mocap data with permissive license you may find it interesting if you haven’t already.

off topic, but I saw you mentioned vr on other post so wondering wether you considered setting up vr mocap? For casual hobbyist like me stuff like glycon3d in enticing though I don’t know what type of mocap you’re after.

I wish it actually happens, but I also fear something akin to regeional dictators helping each other consolidate their power.

I skimmed the article, scrolled down but people hasn’t mentioned its mechanically Chain mail in atomic scale yet? Did I read it wrong?

Glad to see mentions to Peter Watts. His view of humanity is dry and take on real world is even grimmer, but it’s intriguing and backed by science. Also I’m the one of people dying to know what he said at the end of his lecture.

I don’t know where you get the information tho, it’s factually false.

Japanese have have /h/, /ç/, /ɸ/ consonants in ハ行 (written as ha - hi - fu or hu -he - ho but pronounced differently). The consonant /ɸ/ is generally transcribed as f in alphabet.

フ(f+u) is the only letter that pronaunce /ɸ/ in regular ハ行, but ファ行 (f + other vowels) indicates sounds with /ɸ/.

Transcription of wifi in Japanese is ワイファイ, not ワイハイ.

I think they can retain customers’ past flight records and maybe list of media played in flights. That said it’s almost guaranteed some ad marketing corpo is behind, and EULA is always vague enough to allow sharing external party…