Wie OpenAI erklärt, warum LLMs bei völliger Ahnungslosigkeit sicher auftreten
In einem Paper benennt OpenAI selbstbewusste Fehler von großen Sprachmodellen als gewollte technische Schwäche. Das Beheben erfordert ein Umdenken der Branche.
See tagged statuses in the local BookWyrm community
Wie OpenAI erklärt, warum LLMs bei völliger Ahnungslosigkeit sicher auftreten
In einem Paper benennt OpenAI selbstbewusste Fehler von großen Sprachmodellen als gewollte technische Schwäche. Das Beheben erfordert ein Umdenken der Branche.
Sam Altman spricht in einem Blogpost über ChatGPT für Jugendliche
Nach dem Suizid eines Jugendlichen will OpenAI junge ChatGPT-Nutzer besser schützen. Dazu muss der KI-Chatbot das Alter des Users schätzen.
#ChatGPT #IT #KünstlicheIntelligenz #Netzpolitik #OpenAI #news
KI-Update: Klage gegen Google, Mikrozahlungen, Abschaltung aller KI, Google Meet
Das "KI-Update" liefert werktäglich eine Zusammenfassung der wichtigsten KI-Entwicklungen.
#GenerativeAI #KünstlicheIntelligenz #Google #Journal #KIUpdate #Microsoft #OpenAI #Wissenschaft #news
@nixCraft AI is incredibly powerful, but there's a huge gap between the hype and reality. Social media is full of 'gurus' promising this will make you rich with revolutionary game-changing results. Meanwhile, anyone who can do simple math looks at the pricing: a million input/output tokens costs [X amount]. Then you track your actual usage and think: 'Wait, what? I'm burning through tokens that fast?'
Maybe these AI gurus have overlooked a few things about the economics?
Microsoft und OpenAI arbeiten an neuem Deal – immernoch
Microsoft und OpenAI haben sich geeinigt – auf eine unverbindliche Absichtserklärung. Die Umstimmigkeiten werden zur Farce.
OpenAI und Oracle: 300-Milliarden-Deal für Project Stargate
OpenAI soll 300 Milliarden US-Dollar für die Nutzung von Rechenkapazitäten zahlen. Woher das Geld kommen wird? Bisher unklar.
#IT #KünstlicheIntelligenz #OpenAI #Oracle #Wirtschaft #news
Allianz gegen Umstrukturierung von OpenAI
Es formiert sich Widerstand. Philanthropen, Arbeitnehmerverbände und Staatsanwaltschaften wollen, dass OpenAI Non-Profit bleibt.
Tell me this isn't a bubble that's about to burst.
#OpenAI going after nonprofits who are questioning the company's nonprofit status with subpoenas. https://sfstandard.com/2025/09/03/openai-midas-project-elon-musk-subpoena/
OpenAI is preparing to launch an AI-powered jobs platform that could place the company in direct competition with LinkedIn, even though Microsoft, LinkedIn’s parent company, remains its largest financial backer.
https://www.computing.co.uk/news/2025/ai/openai-unveils-jobs-platform-to-challenge-linkedin
OpenAI Paper: Halluzinationen offenbar unumgänglich
Große Sprachmodelle und KI-Chatbots, die nicht halluzinieren? Das hält selbst OpenAI für unmöglich. Aber es gibt einen Ausweg.
Sam Altman: Social Media seit KI-Bots fühle sich fake an
Der OpenAI-CEO wundert sich offenbar über die Auswirkungen von KI auf das Internet und vor allem auf soziale Netzwerke.
Calishat: QueryAnvil: AI As Search Sidekick Instead of Main Character. “I was staring in frustration at a particularly annoying page of search results when I thought to myself, ‘Man, I wish I had some way of marking which of these results are useful and which aren’t, and then have an AI analyze the different sets for language use and give me suggestions for how I can revise my search to get […]
"First it peddled the poison, and now it's selling the cure.
OpenAI announced that it's launching a new AI-powered jobs platform next year that will use AI to match employers with potential candidates. Called the OpenAI Jobs Platform, the initiative will put it in direct competition with job juggernauts like LinkedIn, which has also been experimenting with integrating the tech.
It marks a new direction for the company, perhaps as it tries to maintain some semblance of its altruistic image. As its CEO Sam Altman has frequently warned for years now, it's an inevitability, in his view, that the AI his company is creating will destroy jobs. Many bosses have openly boasted about replacing their workforce with AI agents.
Fidji Simo, OpenAI's CEO of applications, acknowledged the tech's role in destroying jobs.
"We believe fundamentally that AI will unlock more opportunities for more people than any technology in history, but …
"First it peddled the poison, and now it's selling the cure.
OpenAI announced that it's launching a new AI-powered jobs platform next year that will use AI to match employers with potential candidates. Called the OpenAI Jobs Platform, the initiative will put it in direct competition with job juggernauts like LinkedIn, which has also been experimenting with integrating the tech.
It marks a new direction for the company, perhaps as it tries to maintain some semblance of its altruistic image. As its CEO Sam Altman has frequently warned for years now, it's an inevitability, in his view, that the AI his company is creating will destroy jobs. Many bosses have openly boasted about replacing their workforce with AI agents.
Fidji Simo, OpenAI's CEO of applications, acknowledged the tech's role in destroying jobs.
"We believe fundamentally that AI will unlock more opportunities for more people than any technology in history, but it will also be disruptive," Simo, who formerly led Instacart, told Bloomberg. "While we can't eliminate the disruption, we can certainly help more people become fluent in AI and connect them with companies that need their skills.""