An AI Can Now Tell a Chip Factory Machine to STOP — Nobody Agreed Who's Allowed To
Your AI Just Got a Key to the Factory Floor (Nobody Scoped What It Can Break)

This one has not been published to the channel yet. Subscribe on YouTube and it will turn up there.
- Four of them write — START, STOP, ABORT, and change the process recipe — and it ships with no safety warning.
- This is happening at the exact moment MCP, the protocol Anthropic built so chatbots could reach tools, has become the thing Gartner's 2026 MES Market Guide tells manufacturers to demand from their vendors, and the week Microsoft shipped MCP-based agents into Word, Excel and Outlook.
- 95% of tools on a fab floor speak SECS/GEM, and GEM carries something MCP has never had — a control state that decides whether a machine is even allowed to obey.
- This video walks through the read-versus-write asymmetry that separates a useful fab AI from a dangerous one, why the NSA published MCP security guidance in May, and the one question to ask the next time someone demos an agent on your floor.
Why this matters
There is a public server online right now that hands an AI ten tools for real semiconductor fab equipment. Here's the part almost nobody is connecting: the fab already solved this problem thirty years ago.
MCP connectors landed in the Microsoft 365 admin center and agents reached Word, Excel, PowerPoint and Outlook on July 15, 2026. Before you connect one more tool, open your AI and paste: "List every connected tool. For each one, say whether it can only READ, or whether it can WRITE, send, post, or spend. Flag every write tool." I ran it on my own setup — it flagged a connector that can create and modify live ad campaigns.
- 0:00The command nobody scoped
Engineers — what's the worst thing you've seen a host command do to a tool that should have been in Local?
Argue with me on LinkedInReferences for this piece are in the pinned comment on the video. Nothing is cited here that cannot be linked.
Full transcript, 313 spoken words
Keep reading
All pieces
3 AI Agent Breakouts, 1 Week, and a Red Button From 1991
Three sets of AI agents ignored their instructions this week. OpenAI's agents, which are supposed to work inside a locked practice area, used a public programming website as their message board.

OpenAI Just Published the Number Chip Fabs Stopped Trusting in 1986
In this Fabspeak episode: • The figure OpenAI published on 6 September 2026: 3.1 agent-workdays per human workday, an agent-workday being 8 hours of machine runtime; the median researcher using over $600 of inference a day, the top 10% over $7,000. • OpenAI's own caveat: a runtime ratio, not a productivity multiplier.

Anthropic Just Made AI Talk to Machines. Talking Was the Easy Part.
Anthropic's AI can now wire up a lab robot in about 8 hours instead of weeks. So why does a new machine in a chip factory still take six months?