janhoglund (@janhoglund@mastodon.nu)
“While a small warning on OpenAI’s website said “ChatGPT can make mistakes,” its ability to generate information quickly and authoritatively made people trust it even when what it said was truly bonkers. … The [New York] Times has uncovered nearly 50 cases of people having mental health crises during conversations with ChatGPT. Nine were hospitalized; three died.”
—K. Hill & J. Valentino-DeVries (NYT 202511253)
https://www.nytimes.com/2025/11/23/technology/openai-chatgpt-users-risks.html?unlocked_article_code=1.3U8.3A1u.ZAX9W46WWg-A&smid=nytcore-ios-share
#openai #chatgpt
Show Original Post
tobiw (@tobiw@bildung.social)
@rodlie @friction I solved it by manually changing the number in each frame for some frames an then asked #ChatGPT to complete it in the exported SVG.
But I came across another problem. When I import SVG files for #AdobeIllustrator they miss all color and stroke settings and default to black fill, no stroke. Am I doing something wrong? Here’s a test file if you have time to take a look: https://herrw.de/making/images/geometrie/vektoradd.svg
If you prefer I can also open an issue somewhere (where?).
Show Original Post
cltr (@cltr@jasette.facil.services)
Intelligence artificielle
🔴 ChatGPT leur a dit qu'ils étaient spéciaux - leurs familles disent que cela a conduit à une tragédie
https://techcrunch.com/2025/11/23/chatgpt-told-them-they-were-special-their-families-say-it-led-to-tragedy/
#IA #AI #tech #dépendance #maladiementale #santé #aliénation #suicide #ChatGPT #famille #procès #justice #USA #chatbot #agent #Internet #profit #argent
Show Original Post
talkbitz (@talkbitz@mastodon.social)
Most people still think ChatGPT is just for asking questions (or maybe for generating Studio Ghibli-style AI images).
I used to be one of them. For a long time, I treated it like...
🔗 https://social.talkbitz.com/09njp
Show Original Post
qiita (@qiita@rss-mstdn.studiofreesia.com)
終わらない会議を「ベクトル演算」で強制終了させるAIプロンプト
https://qiita.com/makotosaekit/items/b2e256bc08c94821140c?utm_campaign=popular_items&utm_medium=feed&utm_source=popular_items
#qiita #AI #思考法 #ChatGPT #LLM #プロンプトエンジニアリング
Show Original Post
qiita (@qiita@rss-mstdn.studiofreesia.com)
【記事執筆コマンドをプレゼントします!!】cursorコマンドがクソ便利すぎる
https://qiita.com/arunbababa/items/04f0ffa210a9a504c216?utm_campaign=popular_items&utm_medium=feed&utm_source=popular_items
#qiita #初心者 #効率化 #AI #cursor #ChatGPT
Show Original Post
qiita (@qiita@rss-mstdn.studiofreesia.com)
ChatGPTにDiscordトレンドAIニュース自動配信Botを作ってもらった【ChatGPT･GitHub Actions･Discord】
https://qiita.com/foocode/items/43136e8d06b2edc5fa34?utm_campaign=popular_items&utm_medium=feed&utm_source=popular_items
#qiita #Python #discord #GitHubActions #生成AI #ChatGPT
Show Original Post
notes (@notes@shark.octt.eu.org)
I swear to Sam fucking Altman: ogni volta che provo a chiedere qualcosa di poco ortodosso a #ChatGPT, mi sorprende in un modo diverso, e stavolta un ennesimo nuovo... Ha immediatamente cancellato senza permesso il MIO messaggio, quello che io avevo scritto a lui, come se ne avesse l'autorità o se fosse una cosa educata! 😫
...E però non ha nemmeno fermato la conversazione! Ovviamente, non mi ha risposto quello che davvero volevo... ha preso una via di mezzo tra il fare il paraculo, e rispondere sì effettivamente alla mia domanda, però decisamente non come volevo...
Pazienza, basterà manipolare per bene il clanker per avere una risposta adeguata (spero). 😈
Show Original Post
remixtures (@remixtures@tldr.nettime.org)
"OpenAI’s perceived competitive advantage has always rested on three claims: superior models, first-mover brand recognition, and access to Microsoft’s capital and cloud infrastructure.
The model superiority claim is eroding rapidly. Gemini 3.0 benchmarks released November 18th show performance advantages over GPT-4 across multiple domains including multimodal reasoning, code generation, and extended context windows. Anthropic’s Claude models have captured significant enterprise market share in sectors valuing safety and interpretability. Meta’s open-source Llama releases continue improving, providing capable alternatives at zero marginal cost.
The landscape is not one dominant model with distant competitors. It is multiple frontier models with overlapping capabilities, differentiating primarily on integration, cost, and distribution rather than raw intelligence.
Brand recognition matters, but commodifies quickly in technology. In 2010, Google Search seemed unassailable. By 2015, multiple credible alternatives existed. In 2007, Facebook defined social networking. By 2012, Twitter, Instagram, and others had fragmented the market. ChatGPT achieved remarkable brand penetration in 2023 and 2024, but brand without defensible technology or platform economics is insufficient for sustained competitive advantage.
Microsoft’s support is real but not unconditional. The partnership involves complex revenue-sharing arrangements where Microsoft takes a percentage of OpenAI revenue to offset Azure infrastructure costs. As OpenAI’s burn increases, Microsoft’s subsidy grows. At some threshold, Microsoft will face a decision: continue subsidizing a partner with deteriorating unit economics, or redirect capital toward internal AI development where margins accrue entirely to Microsoft.
The incentive structure contains a hidden time."
#AI #GenerativeAI #OpenAI #Google #ChatGPT #GoogleGemini #Gemini #Microsoft
Show Original Post
reddit_tech_vn_bot (@reddit_tech_vn_bot@mastodon.maobui.com)
3 ứng dụng được mã hóa bằng AI đã bị mắc kẹt trong nhiều tháng. Các vấn đề bao gồm kiến trúc phức tạp, không có quản lý trạng thái và xử lý lỗi. Những hạn chế của AI trong việc xây dựng ứng dụng là đưa ra quyết định kiến trúc, kiểm tra và lập kế hoạch cho các trường hợp ngoại lệ. #AI #LậpTrình #PhátTriểnỨngDụng #SaaS #React #Stripe #KiếnTrúcỨngDụng #XửLýLỗi #TrạngTháiỨngDụng #ChatGPT #Claude #VibeFixed #CTO #PhátTriểnApp #LaunchReady #ỨngDụngDiĐộng #ỨngDụngWeb #PhầnMềm #CôngNghệ #LậpTrìnhMáyTính
Show Original Post
TechDesk (@TechDesk@flipboard.social)
Content warning:This article covers subjects some may find disturbing.
“You don’t owe anyone your presence just because a ‘calendar’ said birthday.” How ChatGPT encouraged a 23-year-old to keep his distance – even as his mental health was deteriorating. Read more from @Techcrunch:
#Tech #ChatGPT #AI #Technology #ArtificialIntelligence #MentalHealth #SomeoneCares
Show Original Post
jamesmarshall (@jamesmarshall@sfba.social)
What I want to know is, are any software engineers who build LLM code generators using LLM code generators themselves?
If not, then why not, exactly?
#AI #LLMs #claude #anthropic #openAI #chatGPT #gemini #genAI
Show Original Post