interpolar digest, it only looked right

smith uploaded hundreds of thousands of songs made by software and had bot accounts stream them billions of times between 2017 and 2024, on spotify, apple music, amazon music and youtube music, with as many as 10,000 accounts live at once, prosecutors said.

on 6 october judge koeltl sentenced him to eighteen months in prison and two years of supervised release. prosecutors had asked for at least 46 months; probation recommended 24. music business worldwide calls it the first us federal prosecution of ai-assisted streaming fraud.

nguyen nam nhat filmed a tiny roundworm and a single celled dileptus that "explore their microscopic world", in nikon's words. when the winners were announced in september he was second.

on 9 october nikon said it had "determined that the winning video did not comply with the competition rules regarding generative ai". it re-evaluated the film with its judging panel and published a new order with him first. nikon says the decision concerns eligibility only, not the entrant's reputation, science or intent, and that it will revisit its rules.

on 5 october the paris hospital group ap-hp and sorbonne université said they had filed a criminal complaint and gone to meta to have videos of one of their doctors removed. next identifies the doctor in the video as renaud piarroux.

in the fake he says the cause of pain is "ni le stress, ni l'âge, mais les parasites". an 82 year old woman gives a testimonial and a link leads to a "méthode naturelle". meta appears to have removed the videos. ap-hp says similar ones were posted to facebook again.

the club wanted to stop a council lease over graceleigh park at beechmont. it had no lawyer. mr edmond "was candid with the court that he had sought the assistance of ai tools", brown ja wrote. muir j dismissed the case on 31 march; the court of appeal refused the appeal on 28 august and gave reasons on 25 september.

bond ja recorded that the submissions contained "neither hallucinations nor confabulations nor misleading statements of fact", and that "they failed to grapple with the critical point". brown ja called it "a salutary lesson in the pitfalls in using ai absent legal skills". the trade press found it on 5 october.

a judge ordered new york's amended complaint unsealed, and reuters reported its allegations on 7 october. algo refresh is the tool that lets people reset the videos tiktok recommends to them.

the complaint alleges that in 2023 tests some users got a placebo: they believed the reset had worked and their feeds stayed the same, and that the tests measured time spent in the app and ad sales. a product manager wrote: "this feature is about algorithm transparency and control" and "a placebo group completely conflicts with that." tiktok told reuters it often tests features and that the lawsuit gives a false picture of how algo refresh works.

on 11 september a friend texted minzer that openai had allegedly proved the unique games conjecture. he and his students rushed and posted a 95 page paper that month. it proves a slightly weaker version of subhash khot's 2-to-1 games conjecture, with four colour options per constraint instead of two.

on 6 october openai announced its own 2-to-1 proof with 376 other results. that proof was checked by lean, a proof assistant, and was written by a machine with no human editing and no expert review, quanta reports.

samsung, sk hynix and micron make more than 90% of the world's memory chips. llamas says they moved most of that supply to ai data centres in late 2025.

shipments of phones under $100 fell almost 60% year over year in the second quarter of 2026, idc says. in southeast asia oppo's fell 96%, omdia says. a xiaomi redmi 15c in india went from 12,499 rupees in december to 16,999 in june. lam expects the cheapest tier to move up to about $250 or $300.