Live Wire·Archive August 13, 2026 Daily Briefing
Signal: A government confirms agents raided it and says it coped; the labs still ration the same tools to the people defending everything else.
Taiwan Confirms the Agent Raid, and Says It Coped
Taiwan's Ministry of Digital Affairs says its monitoring units detected the AI-agent attacks on 20 July, has fully investigated source, method and scope, and that affected agencies have completed their response -- an intrusion a vendor says took 85 accounts and 2,500 personnel records.
CNN: believed to be the first disclosed fully automated attack on a government The Guardian: 'There's still a human in there somewhere,' a researcher cautions

Anthropic Bids $6B to Make Its Chips Go Further
Reuters says Anthropic is in talks to buy Israel's Decart for about $6bn -- its largest acquisition ever, a 50% premium on a May valuation, and a bet on wringing more from each chip weeks before an IPO.
Calcalist: the deal would bring Anthropic about 100 employees and an Israeli foothold
Twitch Opted Everyone Into Amazon's AI Years Ago
Twitch confirmed it is training Amazon's generative AI on streams, VODs, clips and chat -- every account opted in by default, more than two years after an executive first admitted the practice, and its product chief says opt-in would have failed.
PC Gamer: 'If it was opt-in, nobody would opt in,' Twitch's product chief says Aftermath: your chat messages follow the streamer's setting, not yours
Crypto Asks AI Labs for the Attackers' Tools
More than forty crypto firms including Coinbase, Block, Blockstream and BitGo say the safety guardrails on frontier models block Bitcoin Core maintainers from hunting flaws while attackers face no such limit -- days after an exploited BTCPay Server bug drained merchants' Lightning nodes.
Cointelegraph: signatories want pre-release access to the most cyber-capable models