Claude subscriptions will no longer cover usage on third-party tools like OpenClaw.

#1 𝕏

Peter Yang reports that Claude subscriptions will no longer cover usage on third-party tools like OpenClaw, highlighting how both Anthropic and OpenAI currently lose money on power users.

#2 𝕏

Anthropic introduces a “diff” research tool that applies software-development diff principles to open-weight AI models, pinpointing unique behavioral differences between them.

#3 𝕏

Claude rolled out Microsoft 365 connectors on every plan, letting you bring Outlook, OneDrive, and SharePoint emails, docs, and files directly into conversations.

#4 𝕏

Boris Cherny explains that a surge in Claude usage from third-party tools has outstripped the design of existing subscriptions, so capacity is being managed carefully and prioritized for product and API customers.

#5 𝕏

xAI launched a new Quality mode in the prompt bar for its Imagine tool on web and mobile, while retaining the previous model under “Speed”; try it now at grok.com/imagine and share your images by tagging @imagine for feedback.

#6 𝕏

Boris Cherny argues that subscription plans are optimized for specific usage patterns rather than raw token counts. He notes that building at scale involves trade-offs that prioritize certain use cases over others.

#7 📝 Simon Willison

Vulnerability Research Is Cooked - Thomas Ptacek argues that frontier LLMs and coding agents will rapidly and dramatically change vulnerability research, automating exploit discovery by pattern matching known bug classes and searching for reachability/exploitability.

#8 𝕏

Lenny Rachitsky shares Simon Willison’s insight that November 2025 was the inflection point for AI coding, unleashing autonomous coding agents benchmarked by Pelican Benchmark and driving red/green TDD with “thin templates.

#9 𝕏

Teresa Torres positions Banani’s AI as an “autopilot” for designers—features like variant generation and adjustable control-vs.-AI toggles let designers stay in the driver’s seat but hand off tasks whenever they choose.

#10 𝕏

Santiago highlights ZooClaw’s launch of hyper-specific, Claw-themed agents and cites a study showing that specialized agents—despite using the same model—outperform general-purpose ones by a mile.

#11 in

Claire Vo argues that AI has outpaced legacy PM scaffolding—tickets, docs, roadmaps and PRDs—and that merely layering AI onto outdated foundations won’t cut it.

#12 in

Marc Baselga argues PMs should absolutely have agentic coding tools (e.g., Claude Code, Cursor) to prototype, query the codebase, and turn specs into working artifacts—yet granting them direct push access to production remains a far more complex debate.

#13 𝕏

Rowan Cheung reports researchers taught a robot to play tennis—mastering forehands, backhands, and shuffles—by training solely on clips of a few amateur players, achieving lightning-fast, coordinated strokes.

#14 𝕏

Anthropic unveiled a difference-only auditing technique for AI models that isolates feature changes to streamline audits, though it can be oversensitive and flag analogous features as distinct.

#15 𝕏

Lenny Rachitsky highlights @simonw’s “dark factory” pattern as the next leap in AI software engineering—teams no longer write or even look at their own code.

#16 in

Peter Yang reports that Anthropic has blocked third-party harnesses like OpenClaw on Claude subscriptions, underscoring how both Anthropic and OpenAI are losing money on $100–200/month unlimited plans—and he predicts they’ll hike prices and tighten margins once they go public...

#17 📝 Simon Willison

Willy Tarreau - Willy Tarreau reports a surge in kernel security reports driven by AI tools: from a few per week to multiple per day, sometimes duplicates, requiring more maintainer effort.