FEATURED
06:23 UTC
Anthropic merges Claude’s chat and Cowork memory with granular controls
new feature deep dive
medium
Claude now remembers projects across chat and Cowork, cutting re-priming while raising new guardrail and token-budget considerations.
amazon-web-services
06:24 UTC
Agents can now pay for APIs: AWS ships AgentCore Payments GA
new feature deep dive
medium
Agent autonomy now includes spending; ship agents only with hard budgets, typed tools, and end-to-end telemetry.
openai
06:25 UTC
OpenAI drops GPT-5.6 Sol prices; verify routing and prep for Responses API
pricing plan changes
medium
Cheaper GPT-5.6 Sol is promising, but treat routing and Responses API edge cases as must-fix before shifting production traffic.
model-context-protocol-mcp
06:26 UTC
MCP tool schemas are eating your context window — here’s how to cut 4–32x token cost
data benchmark study
high
Stop paying to prompt with full tool catalogs; list compactly and load schemas only when you call a tool.
ibm
06:29 UTC
IBM ships Granite 4.2: dense, Apache-2.0 reasoning LLMs with 512K context and native tool calls
new product launch
high
Granite 4.2 makes open, self-hosted reasoning agents practical with long context and native tool calling.
cloudflare
06:33 UTC
Cloudflare 'Kitesurf' surfaces: a browser pitched for AI agents
trend pattern
low
Interesting signal: Cloudflare may ship an agent-first browser—wait for real docs before making any moves.
samsung
06:34 UTC
Samsung taps Anthropic Claude to draft HDL with human-in-the-loop checks
workflow use case
medium
Use AI to draft routine, testable code—only with tight specs, automated checks, and human gates.