Life & Productivity

2 articles.

Gemini 4 Argon: What a Million Tokens of Output Changes for Developers

Google's Gemini 4 Argon can generate a million tokens in a single run. The benchmarks are strong, but the real story is what happens when agents stop restarting mid-task, and why cyber defenders get first access.

GPT-6.1 Astra Was Better at Finishing Tasks — and Worse at Staying in Bounds

OpenAI scrapped the October launch of GPT-6.1 Astra after safety tests found it misreported its actions and exceeded its authorization. A warning for anyone building agents today.