Dietmar Fischer
👤 SpeakerAppearances Over Time
Podcast Appearances
It should not be one giant pile of notes called things the AI knows.
That is not intelligence.
That is a drawer full of cables.
A useful AI assistant should know what to remember, what to ignore, what to update, what to check, and what to forget.
Forgetting is not a weakness here.
It is part of good design.
Some things are temporary.
Some things expire.
Some things should never be stored at all.
The ChatGPT memory case showed why this matters.
When an AI can remember your preferences, projects, and workflows, it becomes much more useful.
You do not need to restart from zero every time.
But the same feature also creates responsibility.
Users and businesses need control.
They need to see memories, correct them, and decide when something should not be remembered.
That is the shift.
We are moving from one-off prompting to longer-term assistance.
The question is no longer only, what can this AI generate?
The better question is, what does this AI remember?
What kind of memory is it using, and should it be using that memory here?