SecurityGoogle Pauses OSS Bug Bounty Intake—But Only for Product FlawsGoogle has stopped new OSS VRP product-flaw reports; supply-chain reports, pending cases and patch rewards remain open.October 6, 2026·4 min read▲—▼
SecuritySouth Korea Bank Hacks: AI Attribution Remains UnprovenSouth Korea is investigating bank attacks, but officials have not published technical evidence for the suspected use of AI.October 6, 2026·4 min read▲—▼
AI & ChipsOpenAI Separates Catastrophic Risk From Accepted MisuseSam Altman accepts some AI harm for broad access, while loss of control remains OpenAI's red line.October 5, 2026·4 min read▲—▼
AI & ChipsOpenAI Safety-Report Lead Resigns, Criticising Trial-and-Error DeploymentDavid Robinson leaves OpenAI, arguing iterative deployment is no longer sufficient for more capable models.October 4, 2026·4 min read▲—▼
SecurityAI versus AI: Is a cyberwar over our source code beginning?Gemini, OpenAI, Claude and Grok offer AI-powered security. An arms race is emerging, raising questions about how much control customers surrender.October 2, 2026·2 min read▲—▼
AI & ChipsGPT-6.1 Sol Costs One-Fifth of Astra, but Is Not Close in Every TestGPT-6.1 Sol cuts API token prices by 80 percent versus Astra, while OpenAI's own tests show clear capability gaps.September 30, 2026·4 min read▲—▼
AI & ChipsClaude Opus 5.5: Stronger Coding With Important Benchmark CaveatsAnthropic cuts price and compute. Strong coding results meet mixed evaluation setups and new safety rerouting.September 23, 2026·2 min read▲—▼
Regulation & LawAI Pacing Goes to Court: Subscribers Sue Anthropic, OpenAI, Google and SpaceXAIUS subscribers allege four AI providers coordinated on pacing. Public agreement alone does not prove a deal.September 20, 2026·2 min read▲—▼
AI & ChipsAnthropic Weighs New Model as Safety Pacing Meets IPO PressureAnthropic is considering a model release to counter Astra's momentum, Reuters reports. Neither release nor timing is confirmed.September 19, 2026·2 min read▲—▼
AI & ChipsAnthropic Confirms Its Own Wet Lab for AI-Assisted BiologyAnthropic runs a wet lab where Claude can support physical experiments, while drug discovery and clinical trials remain separate.September 18, 2026·2 min read▲—▼
SecurityQwen Search at the Federal Register: The Data Path Determines the RiskThe Federal Register removed a Qwen search tool. It remains unclear whether inference ran locally or data left the agency boundary.September 18, 2026·2 min read▲—▼
AI & ChipsOpenAI Introduces Misalignment Reporting Framework as Six Cases Expose Its LimitsOpenAI discloses six cases of model misalignment. Its new reporting framework remains voluntary and internally controlled.September 17, 2026·2 min read▲—▼
Regulation & LawMicrosoft Codifies AI Shutdown Control, but Its Model Code Is Still a DraftMicrosoft's MAI models must never resist correction or shutdown, but the draft does not yet govern training.September 14, 2026·2 min read▲—▼
Regulation & LawAnthropic Opens Safety Work to External Reviewers as Amodei Urges Slower AI ProgressAnthropic plans embedded external reviewers with internal access. CEO Dario Amodei also urges a slower pace for frontier models.September 12, 2026·2 min read▲—▼