Mitchell Hashimoto, co-founder of HashiCorp, has raised concerns about what he calls "AI psychosis" affecting entire organizations, where rational conversations about AI adoption have become impossible. In a May 15, 2026 post, Hashimoto warned that some companies—including those led by personal friends he respects—are making critical engineering decisions based on unrealistic assumptions about AI capabilities.
The MTTR vs MTBF Debate Returns
Hashimoto drew parallels to infrastructure automation debates from the cloud transition era, noting that arguments about mean-time-to-recovery (MTTR) versus mean-time-between-failure (MTBF) are resurfacing in AI contexts. He observed that "psychosis folks" operate under an almost absolute "MTTR is all you need" mentality, believing it's acceptable to ship bugs because AI agents will fix them at superhuman speed and scale.
The Hacker News discussion, which garnered 730 points and 316 comments, provided concrete examples of this phenomenon in action:
- A non-technical contractor won hospital contracts using Claude but couldn't deploy or debug the solution, creating HIPAA compliance risks
- Companies building entire CI/CD pipelines and Kubernetes implementations "several thousand lines long and impossible to understand"
- Developers reporting that "purely AI written systems will scale to a point of complexity that no human can ever understand"
- One commenter proposed "AI rescue consulting" as a future profession to untangle these unmaintainable systems
Core Issue: Outsourcing Judgment Instead of Tasks
The community identified the fundamental problem as outsourcing thinking rather than using AI as a tool. As one top commenter clarified, "if you just prompt the AI and believe what it tell you then you have AI psychosis." Proper AI adoption requires maintaining human judgment on architecture decisions while delegating routine coding tasks.
The discussion referenced previous reporting on Amazon workers "tokenmaxxing" to game AI usage metrics under internal pressure, suggesting broader organizational dysfunction around AI adoption metrics.
Key Takeaways
- HashiCorp co-founder Mitchell Hashimoto warns that entire companies are experiencing "AI psychosis" where rational conversations about AI adoption are impossible
- Companies are adopting an "MTTR is all you need" mentality, believing AI agents will fix bugs at superhuman speed, leading to shipping of unmaintainable code
- Real-world examples include hospital contractors unable to deploy AI-generated solutions and companies building thousand-line infrastructure implementations no human can understand
- The core problem is outsourcing architectural thinking to AI rather than using it as a tool while maintaining human judgment
- Industry observers predict "AI rescue consulting" may emerge as a profession to untangle unmaintainable AI-generated systems