Web infrastructure company TinyFish is making its search offering free. The article breaks down the two bills an AI agent faces when it reads the web: the per-request fee charged by search or fetch APIs, and the token cost of feeding page content into a model, plus reasoning and generated answers.
Four agents running on Google's Gemini 4 Argon slew the Ender Dragon in a fully autonomous Minecraft run, a first. Without assigned roles or strategies, they coordinated among themselves, wrote 2,783 programs, swapped 229 chat messages and died 605 times, for a total cost of $1,620.
The new OpenClaw release makes its update command finish and recover on its own across devices, file systems and platforms, instead of leaving a broken install. Interrupted turns, subagent work and deferred tasks now resume after a Gateway restart, so conversations are no longer lost.
Anthropic is working with law firms on norms for AI agents that would negotiate prices, settle complaints or improve contracts for users. It is also recruiting lawyers and humanities scholars as 'Conceptual Reasoning Fellows' to strengthen Claude where good training data is scarce.
Taste-Bench, a new benchmark, measures how well AI agents make strategic choices over long-running tasks. Even the top performer, GPT-5.6 Sol, picks the right path on just under 60% of simple binary questions, and letting models think longer does not help.
Clubic explains the Model Context Protocol, a common standard that lets chatbots, assistants and agents connect to everyday software. Through it, AI tools can find documents, check calendars or create tasks by reaching data and tools directly.