They had it, and deliberately held it back. That's the whole story.
Before ChatGPT launched in November 2022, Google was widely considered the global leader in AI research and had developed LaMDA, a language model reportedly more capable than the initial ChatGPT. When employees asked at an all-hands meeting whether this was a missed opportunity, Pichai and Jeff Dean said the company had comparable capabilities but that moving quickly represented a major "reputational risk" , precisely because Google was so much larger than OpenAI.
There was a second, less-discussed reason: the business model. Google makes money when you click ads next to search results. Fitting ads into conversational answers is both technically awkward and bad UX. So Google had little financial incentive to move users away from a search model generating billions annually toward an unproven one.
This is a textbook case of the Innovator's Dilemma, a company's current success becoming the main obstacle to adopting innovation that threatens it.
The human cost was real too: two of LaMDA's core engineers left Google in late 2021 after executives repeatedly blocked attempts to release the chatbot publicly. They founded what's now Character.AI.
Longer breakdown here, including how the dynamic has reversed in 2026: [My Article]
Top comments (0)