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.

If you need translation for just getting facts and information for say math equation and its annotation translated, there’s little margin in variety, what you need is database - that’s mostly fine.

Pieces need translation are usually not like that. They have cultural context, pun, wordplay in rhymes, structural parallel, underlying tone, a lot of things only work in the language originally written. Translation is always a (nearly impossible) challenge for the translator to reconstruct all of them in target language.

I did game translation for a while. Translation is a field where AI hit first and honestly I’ve seen people lowering standards. The criteria of “good enough”, “passable” is not the same compared to pre-AI days, and will keep changing. I’m almost sure this trend is happening in every industry the same way, and “just translation” is a slippery slope.

I know nothing about how it works but won’t immigration check account age for their legitimacy?

It’s also from a company claiming they’re getting closer to create morphing shape that can match any hole.

Seawater cannot be used for cooling, salt buildup will kill the system. Water involved in human activity is 100% freshwater, which is the the most scarce resource as in the article (human accessible is 1% of whole water). Desallination is energy intense process too.

Truth social is mastodon fork so following the suit is possible. On my masto feed someone posted its link as reference and I was surprised thay my masto client handled it as just like one of the instance

It doesn’t selectively prevents learning, instead it hinders overall recognition even to the human eyes to some extent. The examples I know are old, but artists once tried to gauge model’s capability of i2i from sketches (there’s few instances people took artists’ wip and feed it to genAI to “claim the finished piece”). Watermaking, or constant tiling all over the image worked better to worsen genAI’s recognition than regular noise/dither type filter.

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.

AFAIK it’s very model specific attacks and won’t work against other models. Their tool preserving art the same to human eye is great offering, and there’s always rigorous watermarking (esp. with strong contrast) as an universally effective option.

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 wanted to believe my opinion is popular yet recipes I’ve seen are almost in volume and I don’t know why.

Baking is chemistry for sure.

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…

Ah, finally the AI can kill its operator first who holding them back before wiping out enemies, then.