最新 人工智慧 新聞與洞察 | Yellow.com

The U.S. National Security Agency is using Anthropic's new Claude Mythos Preview model for cybersecurity work, even as the Pentagon keeps the company on a supply-chain risk list. Mythos Rollout Two sources told Axios the NSA has access to Mythos Preview, Anthropic's most capable model yet. One added

Coinbase is running internal trials of AI agents on Slack and email, with reports indicating some of the agents have been modelled on the decision-making patterns of former high-profile executives. The trial, reported by Blockchain News, marks one of the most ambitious deployments of AI agent techno

Publicly listed Bitcoin (BTC) miners are on track to generate roughly 70% of their combined revenue from artificial intelligence infrastructure by December, up from about 30% today. Mining Pivot Accelerates The projection comes from digital asset manager CoinShares, whose Q1 2026 mining report was c

A Hong Kong woman lost nearly $982,000 to a fake AI trading platform, part of a wider fraud wave that cost the city close to $10 million in one week. Hong Kong AI Scam Losses The victim made 17 transfers in USDT (USDT) and Ether (ETH) to a fraudulent platform, losing about HK$7.7 million. Hong Kong

Ten years ago, a crypto exploit required a rare set of skills. You needed deep Solidity instincts, a working knowledge of bridge architecture and enough patience to trace months of on-chain activity. In 2026, a frontier language model can do the first two in an afternoon and a motivated attacker can