July at Claude: From Smart Chat to an Employee You Showed Once

Anthropic released a series of updates in July 2025 that shift Claude from a model that requires verbal task explanation each time to an agent that learns by demonstration. Key features include Record a Skill, which records screen and voice to create reusable skills, and an effort regulator for Opus 5 that controls reasoning token costs. The article analyzes practical implications, cost calculations, and pitfalls like silent fallback.

Anthropic

From GPT-2 to Kimi K3: Scale Grows 22,580x in Seven Years, Main Architecture Thread Builds a "Memory Operating System"

The technical evolution of large models from GPT-2 to Kimi K3 shows how AI architecture moves from "remembering everything" to "selective memory". KV Cache solves generation efficiency, Linear Attention explores more efficient long-term memory, DeltaNet and Kimi Linear enable models to learn to update, forget, and manage information. Kimi K3 has 2.8 trillion parameters, equivalent to about 22,580 GPT-2 models.

Moonshot AI