OpenAI launched custom GPTs as the next app store. Google looked at the same feature and made the opposite bet. Lisa Huang led the Gems team at Google. She told me the story of what happened when OpenAI dropped custom GPTs in the middle of their development cycle. The team had to decide: chase OpenAI's framing or go a different direction. OpenAI positioned GPTs as a third-party ecosystem. GPT Store. Monetization tools. The pitch was "this is the new app economy." Google's read was different. The instructions in a Gem are easily copied. The custom knowledge could potentially be prompted out. There wasn't a defensible moat for third-party creators. So the ecosystem thesis didn't hold. Instead they focused Gems entirely on personal productivity. Your Gems. Your team's Gems. Tools you build for yourself and share internally. No marketplace. No monetization layer. Lisa's framing on the episode: a "me too" feature doesn't always perform as well as going back to first principles on what you're actually trying to accomplish. And the data so far validates that read. The GPT Store hasn't become a meaningful revenue channel for most creators. The custom GPTs that actually get used are the ones people built for themselves, not the ones they found in a store. This is a product strategy lesson worth studying. Same technology. Same feature set. Two completely different product visions based on different reads of the defensibility question. Google bet on utility over ecosystem. That bet looks correct.