Thanks for the info. I have not really tested Seedvault myself so this is all good to know.

Ironically, one of the main reasons I switched to GrapheneOS was because Google’s backups were so frustrating and I was hoping Seedvault would be more comprehensive.

What’s wrong with Seedvault?

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.

I refer you to #7 on Bruce Tognazzini’s evergreen top ten list of design bugs.

https://www.asktog.com/Bughouse/10MostWantedDesignBugs.html