No. And there are a LOT of assumptions baked into this about people's passive engagement with models, early model flaws being projected indefinitely into the future, and the effectiveness of AI.