But here’s the really funky bit. If you ask Claude how it got the correct answer of 95, it will apparently tell you, “I added the ones (6+9=15), carried the 1, then added the 10s (3+5+1=9), resulting in 95.” But that actually only reflects common answers in its training data as to how the sum might be completed, as opposed to what it actually did.
This is not surprising. LLMs are not designed to have any introspection capabilities.
Introspection could probably be tacked onto existing architectures in a few different ways, but as far as I know nobody's done it yet. It will be interesting to see how that might change LLM behavior.
To elaborate on this a little, you can use Flatseal to specify which directories a Flatpak app can have access to directly. For example, in a music player that stores the path of your music library, you'd want to use Flatseal to be sure it has direct access to that folder. This is similar to GrapheneOS's storage scopes.
Aside from that, apps can also call on a file picker that lets you choose any file/folder on your system, and flatpak then creates a virtual path to bridge to that file/folder without exposing the entire rest of the filesystem. This is nice for one-time open/save commands, but doesn't work for apps that need persistent access to a specific directory like in the music player example. This is similar to Android's file provider API.
I don't recall off the top of my head what flatpak apps have access to by default. Some subset of the home folder, I think?