habr (@habr@zhub.link)
Три аккаунта ChatGPT и один прокси: как перестать следить за лимитами
У меня три ChatGPT-аккаунта. На каждом свои лимиты, своя история, свои причины существовать. И на каждом они кончаются в самый неподходящий момент. Переключаться вручную то ещё удовольствие. Особенно когда работаешь через Codex CLI или OpenCode и хочешь просто писать код, а не следить за тем, на каком аккаунте сейчас крутится запрос. Именно под эту боль и написан codex-lb.
https://habr.com/ru/articles/1002258/
#chatgpt #openai #proxy #load_balancer #codex #rate_limit #fastapi #opencode #api_keys #docker
Show Original Post
hhardy01 (@hhardy01@mastodon.social)
New short video from my stories, art, RPG campaign and persistent world of "The Starry Messenger."
The Specter crew foils an attempted ambush. One of the would-be ambushers, Stonehead the Rockmaw, is captured, and the main characters and their Proximian allies investigate.
#aiart #SORA #ChatGPT #GoogleVideo #video #sciencefiction #cyberpunk
Interplanetary Geographic: Scarlet Desert Ambush teaser
https://drive.google.com/file/d/1FaYlaRgXVMNw4yO1N16wjEjLlMqRBtt7/view?usp=sharing
Show Original Post
r (@r@fed.brid.gy)
Accountability, justice, truth and systemic reform #ChatGPT Llewelyn Pritchard 23 February 2026
Show Original Post
xameer (@xameer@mathstodon.xyz)
#Claude need be taken a bit more seriously, any shitty #ai from #Gemini to #perplexity to #chatgpt can act as an agent but Claude is smarter and that implies gravity , this can actually take out many mediocre devs' livelihood
Although the only honest reason of paging anyone anything was that they were good enough for their task which was indispensable at that time , but this shouldn't be the only reason , many social arrangements may collapse and millions won't know what to do , this is different from human #competition which was unfair and gamed too tbh
This won't be #recession or #layoffs
Show Original Post
posts (@posts@aisparkup.com)
LLM이 만든 비밀번호가 위험한 이유, 100비트처럼 보이지만 실제론 27비트
AI가 만든 비밀번호가 강해 보여도 실제로는 예측 가능한 패턴 덩어리라는 Irregular 연구. Claude 생성 비밀번호의 실제 엔트로피가 이론치의 27%에 불과한 이유를 설명합니다.https://aisparkup.com/posts/9480
Show Original Post
forcebrass (@forcebrass@mastodon.social)
The complete moron that #ChatGPT 's Sam Altman is just defended the insane energy use of his models by saying that it "also takes a lot of energy to train a human". Hope this will stoke your fire if the #fuckai energy was not high enough as of yet in 2026.
Article link in which you also find the video:
Show Original Post
MarieClaudeS (@MarieClaudeS@mastodon.top)
Hypothèses/L’équipe de rédaction d’Academia (2 février 2025).
Pourquoi je n’utilise pas ChatGPT, Florence Maraninchi (enseignante-chercheuse en informatique)
https://mcinformactions.net/pourquoi-je-n-utilise-pas-chatgpt-florence-maraninchi
#IA #OpenAI #ChatGPT
Show Original Post
MarieClaudeS (@MarieClaudeS@mastodon.top)
Clubic
Microsoft lâche OpenAI : le divorce tech qui risque de tout faire basculer pour ChatGPT
https://mcinformactions.net/microsoft-lache-openai-le-divorce-tech-qui-risque-de-tout-faire-basculer-pour
#IA #OpenAI #ChatGPT #Microsoft
Show Original Post
yuicho (@yuicho@mstdn.yuicho.net)
え？？？？
お前実際にDavinciResolveを触ってテストするなんてことできるん…？？？
Show Original Post
craigduncan (@craigduncan@mastodon.au)
It's in their defence of legal liability that we learn AI companies vlaims that their models are built to influence user behaviour.
"OpenAI said it trains ChatGPT to discourage imminent real-world harm when it identifies dangerous situations and to refuse assistance for illegal activities."
This kind of deliberate psychological influence, whether successful or not, isn't regulated is it?
Source :The Standard
https://www.thestandard.com.hk/world-news/article/325000/
Show Original Post
dagger (@dagger@chaos.social)
Trage dieses Shirt.
Interessant, was #ChatGPT meint, was auf dem Bild zu sehen ist „Nahaufnahme von Text auf einem Stoff in Schwarz mit einem weißen Hintergrund.“.
Idee beim #39C3 gesehen.
Show Original Post