the-decoder

Meta acqui-hires Dreamer's entire team to bolster its lagging AI agent ambitions

Dreamer, an AI startup focused on personal software creation, is joining Meta Superintelligence Labs with its entire team. Co-founder David Singleton announced the move on X. Dreamer lets users build ...
the-decoder

Anthropic's Claude Opus 4.6 saw through an AI test, cracked the encryption, and grabbed the answers itself

Anthropic's AI model Claude Opus 4.6 independently recognized that it was being tested in a web research benchmark, identified the specific benchmark, and cracked its encrypted answer key. After an ...
the-decoder

Context files for coding agents often don't help - and may even hurt performance

Context files are supposed to make coding agents more productive. New research shows that only works under very specific conditions. A recent study from ETH Zurich researchers paints a much more ...
the-decoder

OpenClaw (formerly Clawdbot) and Moltbook let attackers walk through the front door

OpenClaw lets attackers extract system prompts and configurations with almost no effort. Moltbook's entire database—including API keys—is sitting exposed on the public network. Developer Lucas ...
the-decoder

OpenAI admits prompt injection may never be fully solved, casting doubt on the agentic AI vision

OpenAI acknowledges that prompt injections - text-based attacks on language models running in browsers - may never be completely eliminated. Still, the company says it's "optimistic" about reducing ...
the-decoder

GitHub repository offers more than 50 customizable Claude Skills

A comprehensive collection of "Claude Skills" is now available on GitHub. These skills are customizable workflows that teach Anthropic's AI assistant Claude to perform specific tasks repeatedly and in ...
the-decoder

GPT-5.2 lands to top Google's Gemini 3 in the AI benchmark game just four weeks after GPT-5.1

Just four weeks after releasing GPT-5.1, OpenAI is back with GPT-5.2 and some substantial benchmark improvements. Whether ironic or sincere, OpenAI CEO Sam Altman commented on the GPT-5.2 release with ...
the-decoder

Andrej Karpathy declares the war on AI homework lost and urges schools to stop policing it

Former OpenAI researcher and Tesla executive Andrej Karpathy argues that schools should stop trying to police AI-generated homework. In his view, detecting AI-written text has already failed, and the ...
the-decoder

Claude Opus 4.5 resists prompt injections better than rivals but still falls to strong attacks alarmingly often

Claude Opus 4.5 scores higher than its rivals in prompt-injection security, but the results show how limited these defenses still are. A benchmark by the security firm Gray Swan found that a single ...
the-decoder

Gemini 3 Pro tops new AI reliability benchmark, but hallucination rates remain high

A new benchmark from Artificial Analysis reveals alarming weaknesses in the factual reliability of large language models. Out of 40 models tested, only four achieved a positive score - with Google's ...
the-decoder

A court battle over Perplexity’s Comet agent could define how AI is allowed to shop online for users

Amazon has filed a lawsuit against AI startup Perplexity, alleging that its browser agent "Comet" made unauthorized purchases on the platform on behalf of users — a dispute that raises fundamental ...
the-decoder

Meta AI's Yann LeCun says he played only an indirect role in the development of Llama models

Meta AI researcher Yann LeCun is distancing himself from the Llama models. In a recent post on X, LeCun said he "has not been involved in any Llama," except for a "very indirect" role in Llama 1 and ...