TechCrunch

AI, Security & Privacy

Trump Administration Clears Anthropic’s Mythos 5 for Use by Over 100 U.S. Companies and Agencies

Trump administration authorized Anthropic to redeploy Claude Mythos 5 for 100+ U.S. agencies and companies, including Anthropic’s non‑American employees, partially reversing a ban that started June 12, 2026. Commerce Sec. Howard Lutnick wrote to chief compute officer Tom Brown that safeguards are in place after researchers allegedly bypassed Mythos 5’s guardrails; the directive doesn’t address Fable 5.

AI, Security & Privacy

OpenAI Restricts GPT-5.6 Rollout at U.S. Government Request, Pushes Back on Process

On June 26, 2026, OpenAI will roll out GPT‑5.6 only to a small group of “trusted partners” at the U.S. government’s request, calling the arrangement temporary. Sol/Terra/Luna are priced at $5/$30, $2.5/$15, and $1/$6 per 1M input/output tokens, amid a U.S. exec order pushing firms to submit advanced models for review up to 30 days early, which Dean Ball says creates de facto involuntary licensing.

AI, General

Tesla Settles Fatal 2023 FSD Crash Lawsuit While Federal Investigations Remain Open

Tesla settled a fatal 2023 lawsuit tied to its FSD (Supervised): a Model Y hit 71-year-old Johna Story as she stepped out to direct traffic after an earlier sun-glare crash, with settlement terms undisclosed (Bloomberg). The family case is resolved, but NHTSA’s 2024 probe—upgraded in March 2026—still says Tesla’s degradation detection fails to detect and/or warn the driver in reduced visibility such as glare and airborne obscurants.

Scroll to Top