When the people building autonomous software engineers can't trust each other, what are the rest of us supposed to do with their code?
The Summary
- Factory AI's CEO publicly accused rival Cognition of corporate espionage, alleging a board advisor was secretly meeting with the competitor while sitting in confidential strategy sessions
- Factory also claims Cognition employees posed as job applicants to extract product information during the interview process
- Both companies are racing to sell AI coding agents to enterprises, the exact market where trust and security are supposed to matter most
The Signal
Factory AI CEO Matan Grinberg went scorched-earth on X Wednesday, announcing the company had terminated Chris Degnan from his board observer and advisor role. The reason: Degnan was simultaneously holding recurring conversations with Cognition executives while advising Factory on confidential matters. Degnan is now joining Cognition as chief revenue officer. The timing matters. This isn't about someone leaving for a competitor. It's about allegedly playing both sides while sitting in the room where strategy happens.
The accusations get sharper. Grinberg claims Cognition employees posed as job candidates to pump Factory engineers for product details during interviews. Both Degnan and Cognition deny any information about Factory was sought or shared. But the public nature of this fight tells you something about how high the stakes have climbed.
"It feels as though ethics is being thrown out of the window in the AI era."
These aren't garage startups. Both Factory and Cognition carry multibillion-dollar valuations. Cognition built Devin, the AI coding agent designed to automate software engineering tasks. Factory sells infrastructure for deploying AI agents across the entire development process. They're both selling to enterprise customers who need to trust that the AI touching their codebase won't leak secrets or introduce vulnerabilities. And now those same customers are watching these companies accuse each other of corporate espionage.
The AI coding market is moving fast because the prize is enormous. Big Tech and well-funded startups are racing to build autonomous software engineers. Whoever wins gets to sell picks and shovels to every company trying to move faster with fewer engineers. But speed and corner-cutting have a way of traveling together. When Grinberg says ethics are being thrown out the window, he's describing the water everyone in this market is swimming in.
Key dynamics at play:
- Board observers and advisors see everything: roadmaps, financials, customer lists, technical strategies
- AI coding tools have access to proprietary codebases and internal systems
- Enterprise buyers need to believe these tools won't become Trojan horses
The Implication
If you're evaluating AI coding agents for your company, this fight just became part of your due diligence. Ask harder questions about information security, employee vetting, and contractual protections. Watch how these companies handle this dispute. The ones that stay quiet might be calculating. The ones that lawyer up might be scared. The ones that address it directly might actually care about trust.
For anyone building in the agent economy: your competitors are watching everything you do, and some of them will cut corners you won't. The question is whether you build guardrails now or explain failures later.