What's this about?
A former member of the team says Google internally built "basically ChatGPT one year before it came out" — a chat product called LMChat — but Google was too nervous to release it, and DeepMind was blocked from shipping products that could disrupt Google's core business. "I think about this a lot," he wrote. The confession went viral as first-hand confirmation of one of tech's most consequential misses.
What exactly is "the innovator's dilemma" here?
It's the classic trap where a market leader can't ship a breakthrough because the breakthrough threatens its own cash cow. A chatbot that answers questions directly undermines search ads — Google's revenue engine — so the safer-looking move was to sit on it. Startups with nothing to lose (like OpenAI) have no such brake. The irony: not shipping didn't prevent the disruption; it just handed the lead to someone else.
What's the community saying?
- swyx: bookmark this for the next VC who asks "what if the incumbent builds this?" — having the tech means nothing if the org can't ship it.
- People connect it to Google's transformer-inventors-who-all-left pattern: invention happened there, productization happened elsewhere.
- A sympathetic minority notes the nervousness wasn't irrational — early LLMs hallucinated freely, and Google faced reputational/regulatory stakes a startup didn't.