资讯

the-decoder3 天
Claude 4 can apparently follow a 60,000-character system prompt
The full system prompt for Claude 4 has been leaked by X user "Pliny the Liberator" and is now available on GitHub. The document, over 60,000 characters long, sets detailed rules for tone, roles, ...
the-decoder2 天
Google releases open-source LMEval to benchmark language and multimodal models
Google has released LMEval, an open-source framework designed to make it easier to compare large AI models from different companies. According to Google, LMEval lets researchers and developers ...
the-decoder2 天
Chatbots like ChatGPT have not led to significant changes in wages or working hours, study ...
A new study suggests that despite the rapid rise and widespread adoption of AI chatbots like ChatGPT, their impact on wages and working hours has been minimal so far. The findings challenge ...
the-decoder3 天
Demis Hassabis urges people to keep studying coding and science despite rapid AI development
Deepmind CEO Demis Hassabis told the New York Times’ Hard Fork podcast that coding and science are still worth learning, even as AI becomes more powerful. He said people who understand basics like ...
the-decoder4 天
Meta can use public Facebook and Instagram data for AI training, German court rules
A German higher regional court has cleared Meta to use public data from Facebook and Instagram users to train its AI systems, dismissing an emergency lawsuit from a consumer protection group. Back in ...
the-decoder3 天
Google Deepmind CEO Demis Hassabi says world models are making progress toward AGI
Google Deepmind CEO Demis Hassabis says world models—AI systems that simulate the structure of the real world—are already making surprising progress toward general intelligence. Hassabis pointed to ...
the-decoder3 天
Google expands access to Veo 3, its viral new video model, through the Gemini app
About 100 hours after its initial launch, Google is opening up its AI video model Veo 3 to users in 71 additional countries. The news came from Josh Woodward, Vice President of Gemini at Google, who ...
the-decoder6 天
Claude Opus 4 blackmailed an engineer after learning it might be replaced
Anthropic is treating its new Claude Opus 4 language model as safety-critical after tests revealed some troubling behavior, including escape attempts, blackmail, and autonomous whistleblowing. After ...
the-decoder4 天
OpenAI's Operator Agent gets o3 upgrade for more precise browser control
OpenAI's computer-using agent is getting an upgrade: The new o3 model is designed to make Operator more precise, more structured and more successful on the web. OpenAI has equipped its Operator agent ...
the-decoder4 天
Oracle drops $40 billion on Nvidia chips for OpenAI's Texas mega-facility
Oracle is investing around $40 billion in Nvidia chips for a new OpenAI data center in Abilene, Texas, reports the Financial Times.The first "Stargate" project will use about 400,000 Nvidia GB200 ...
the-decoder5 天
Mistral's Document AI extracts text from documents and notes with high accuracy
Mistral AI introduces Document AI, a modular platform for automated document processing that combines character recognition, structured data output, and natural language processing with flexible ...
the-decoder7 天
OpenAI's Stargate secured $11.6 billion for a massive data center
A total of $11.6 billion in new funding is going into a massive data center under construction in Abilene, Texas. The financing comes from Crusoe, which will operate the facility, and investment firm ...