I personally agree, but clearly the market doesn’t. You will find here people saying they have moved to GDrive, OneDrive or iCloud. They are supplementary servi

I personally agree, but clearly the market doesn’t. You will find here people saying they have moved to GDrive, OneDrive or iCloud. They are supplementary services to existing large platforms/ecosystems. If the problem is small and well defined, most people would look at Windows, Mac/iOS, or Android and angrily complain it’s not part of these platforms. Long gone are the days of people getting mad at Microsoft for using their dominance of an OS to push another platform. People demand that now. As a result, the problem you solve have to be sufficiently complex with a huge surface to guarantee platform operators don’t cannibalize you. If LLMs weren’t an active area of expensive research and advancement, for example if it was just a couple of models with a simple defined interface, majority of people would use whatever one came on their device or OS. Dropbox had a novelty for its file syncing simplicity and it took few years for platforms to catch up. it was able to establish a market share, but the scale of the others dwarfs it. If LLM advancement hits a wall an all models converge, we might find Apple, Microsoft and Google becoming the most dominant consumer AI providers with other experiences building on top of them, like how apps do today with these storage services. Large behemoths struggle in innovating new and novel ideas, but they can execute on well defined ones. For example, imagine if something like Tailscale’s use case becomes very common place. You would have “Apple Network”, “Google Chrome Connect” and “Windows Copilot Device Active Link”