People pretend to care about privacy, but a comfortable majority won't inconvenience themselves more than a minor way to get much more of it.
The data portability thing is spot on. People will get better at learning how to store their data, and apps will get better at pulling it in and learning faster. It's a solvable hurdle that people will find partial solves for.
-
The much bigger thing is 'does it solve my problems substantially better', is it relatively cheap, and is it easy to use. That last one is the biggest hurdle right now, outside of us geeks. If you require that users need to learn prompt engineering to really leverage your app well, you're pushing away at least >80% of the potential market.
People pretend to care about privacy, but a comfortable majority won't inconvenience themselves more than a minor way to get much more of it.
The data portability thing is spot on. People will get better at learning how to store their data, and apps will get better at pulling it in and learning faster. It's a solvable hurdle that people will find partial solves for.
-
The much bigger thing is 'does it solve my problems substantially better', is it relatively cheap, and is it easy to use. That last one is the biggest hurdle right now, outside of us geeks. If you require that users need to learn prompt engineering to really leverage your app well, you're pushing away at least >80% of the potential market.
"Memory" feature is overvalued; personalization won't create a lasting moat in AI models.