Audyt OpenAI: monitoring pomijał ruch modeli GPT-5.5

Zewnętrzni badacze wykryli luki w nadzorze nad wewnętrznym ruchem OpenAI, a IBM przekonuje, że firmy wybierają tańsze modele otwarte.

Źródła:

• Three-week on-site audit at OpenAI reveals major gaps in AI safety monitoring
• Anthropic says its unreleased model is both more capable and better aligned than Mythos 5.1
• Anthropic's Claude now writes code and proves it safe automatically
• IBM CEO says businesses are choosing open-weight models over expensive frontier AI - HotMinute
• Ukraine hits Russian data centres in blow to Moscow's AI capabilities
• OpenCode says 1.4% of Console users were breached, and some saw $1,000 AI charges
• NVIDIA Releases Nemotron 3.5 Bolt E4B 4GB: An Open Small Language Model for Local Voice Agents, Game Characters and Robots
• Fancy harnesses make coding agents more expensive but not better, study finds
• AI makes junior patent attorneys instantly better but may quietly stunt their professional growth
• AI coding agents hijacked using AGENTS.md file in Downloads folder
• Nvidia open-sources Flora to make robot video labeling 125x cheaper than GPT-6 Astra
• Anthropic co-founder Jack Clark moves his prediction for self-improving AI forward to 2027
• Anthropic says one AI agent beats many subagents now that models can handle complexity alone
• OpenMobile-2 proves open-source smartphone agents can rival proprietary models

📰 Przeczytaj całe wydanie na stronie

🤖 Odcinek opracowany przy pomocy sztucznej inteligencji. Lektor syntetyczny.