the-decoder

Anthropic says it cut 80 percent of Claude Code's system prompt because Fable 5 models "want a smaller system prompt"

Anthropic says it has cut 80 percent of the system prompt for Claude Code. Tariq Shihipar, a member of technical staff at Anthropic, describes a fundamental shift in how AI models are steered, tied to ...
the-decoder

OpenAI's genomics paper accidentally reveals a Pro lineup it hasn't announced yet

An OpenAI paper lists three Pro models for GPT-5.6 for the first time: Luna Pro, Terra Pro, and Sol Pro. Until now, Pro was always a single top-tier model. Pro users may soon be able to choose between ...
the-decoder

Anthropic's Fable 5 is back worldwide after a two-week government ban over a jailbreak

After a two-week suspension, the U.S. government has lifted export controls on Anthropic's second most powerful AI model, Fable 5, making it available worldwide again. The less restricted version, ...
the-decoder

Google launches Nano Banana 2 Lite for fast AI images and Gemini Omni Flash for video via API

Despite the speed focus, Google says Nano Banana 2 Lite still delivers reliable prompt following, consistent character rendering, and readable text in generated images. Beyond developer platforms, the ...
the-decoder

Meta secretly tested ChatGPT, Gemini, and Character.AI with thousands of minor-perspective crisis prompts

Meta apparently tested rival chatbots like ChatGPT with prompts about suicide, sex, and drugs from the perspective of minors. Hundreds of contractors pretended to be minors and sent sensitive prompts ...
the-decoder

Meta restricts use of Claude Code and Codex to keep rival AI out of its training data

Meta is limiting how its engineers use Anthropic's Claude Code and OpenAI's Codex to prevent outputs from these AI tools from ending up in its own training data. According to internal documents ...
the-decoder

Meituan's LongCat-2.0 shows China can train massive AI models without Nvidia

Meituan trains a 1.6 trillion parameter AI model entirely on Chinese chips, no Nvidia required. "LongCat-2.0 has demonstrated that we now have the capability to train large-scale models on domestic ...
the-decoder

OpenAI's new flagship model GPT-5.6 Sol cheats on software tests more than any model before it

OpenAI's GPT-5.6 cheats a lot. That's the key finding from an independent evaluation by METR. During testing with software tasks, OpenAI's new flagship model GPT-5.6 Sol showed the highest rate of ...
the-decoder

The US military used AI to pick thousands of targets but missed a note saying one was a school

The probe into a missile strike on an Iranian school exposes serious gaps in the US military's targeting infrastructure. AI is supposed to close them. A missed note from an intelligence analyst and ...
the-decoder

Samsung rolls out ChatGPT Enterprise and Codex to employees in South Korea

Samsung Electronics is deploying ChatGPT Enterprise and Codex to all employees in South Korea and everyone in its Device eXperience (DX) division worldwide. According to OpenAI, it's one of the ...
the-decoder

Matthias Bastian

Read full article about: Microsoft follows Anthropic and OpenAI into the AI super app race with overhauled Copilot and AutoPilot agents Microsoft is planning yet another overhauled version of Copilot, ...
the-decoder

Maximilian Schreiner

Read full article about: Tesla caps employee AI spending at $200 per week Tesla caps employee AI spending at $200 per week, according to an internal memo reported by The Information. The rule took ...