Mitchell Hashimoto, the creator of HashiCorp tools like Vagrant, Terraform, and Packer, recently dropped a bombshell take that has the tech world buzzing. In a post that quickly racked up over 2,100 points on Hacker News, he declared:
"I believe there are entire companies right now under AI psychosis."
The statement is blunt, provocative, and cuts straight to the heart of a growing unease in the industry. Hashimoto isn't talking about AI itself being dangerous. He's talking about the collective delusion sweeping through startups and enterprises alike — the feverish belief that slapping "AI" on a product is a substitute for building something real.
What Does "AI Psychosis" Look Like?
Hashimoto describes a pattern he's observed across the startup ecosystem. Companies that have no clear product-market fit, no sustainable revenue model, and no real engineering moat are pivoting hard into AI. Not because they have a genuine insight about how to apply machine learning, but because investors demand it and the hype cycle rewards it.
The symptoms include:
- AI-washing everything. Products that were perfectly fine as SaaS tools suddenly rebrand as "AI-powered" with a thin ChatGPT wrapper.
- Abandoning core value props. Companies that had a working product ditch their roadmap to chase the latest LLM trend.
- Fake complexity. Teams that can't explain what their model does or why it's better, but lean on buzzwords like "agents," "RAG pipelines," and "fine-tuning" to sound sophisticated.
- Zero defensibility. When the underlying model is just an API call to OpenAI or Anthropic, what exactly is the moat?
The Hacker News Reaction
The thread exploded with over 1,200 comments, and the sentiment was overwhelmingly in agreement. Developers and founders shared war stories of watching sensible companies lose their minds.
One commenter noted: "I've seen startups burn through $10M+ building 'AI' features that are literally just a prompt template wrapped in a Stripe subscription. The emperor has no clothes."
Another pointed out the irony: "The companies most loudly claiming to be AI-native are often the ones with the least understanding of what AI can actually do."
The Deeper Problem
Hashimoto's critique goes deeper than just calling out hype. He's pointing at a structural issue in how venture capital and startup culture operate. When every pitch deck needs an AI angle, you end up with companies that are optimized for fundraising, not for building.
The real danger isn't that AI is overhyped. It's that the hype is distorting incentives. Engineers are being pushed to build things that don't need to exist. Capital is being misallocated. And genuine innovation in AI is being drowned out by noise.
The Takeaway
Hashimoto's warning is a healthy reality check for anyone building in AI right now. The question every founder and product builder should ask themselves is simple: Would your product still be valuable if you stripped away the AI label?
If the answer is no, you might be suffering from AI psychosis too.
The cure? Focus on real user problems. Build things that work. And maybe, just maybe, take the AI buzzwords out of your pitch deck for a week and see if your product still makes sense.
What do you think? Is AI psychosis real, or is this just healthy skepticism? Drop your thoughts in the comments.
