You can't just remember every conversation and ... something useful.
Indiscriminate memory buries you. What you need is "monumental" memories, which is also known as documentation. AI can be very effective in helping you manage that. That should be kept tight, just like code. You want to keep enough of a record of how you got to where you are that you are equipped to be effective in future tasks.
I would say that your suggestion is basically digital hoarding. Lets say there are 3 kinds of people, the person with OCD who keeps his house almost barren in the name of tidiness, the person who visits his possessions periodically with a critical eye, deciding what to keep and what to throw away, and the hoarder. When it comes time to solve some household problem, say hang a picture, which one of them will have what they need to do it (a hammer and a nail, maybe a little bit of wire) and be able to find it?
I think you might consider keeping the session history for a while, maybe a month or 3 months or 6 months, but have something that automatically deletes them after that.
While I generally agree for general human living, I've seen impressive "digital hoarder" success using AI because it changes the dimensions of reasonableness.
Imagine your hoarder scenario except that the hoarder has unlimited storage space and a team of librarian/warehouse workers to go find and fetch the items needed for the wall hanging task.
AI does not have unlimited storage space and digging through tens of thousands of keyword hits will cost you an unlimited amount of money as your history grows and grows.
I was not really thinking in terms of reusing all knowledge at once.
It is useful to be able to see what was the reasoning behind a single PR to discriminate for effort. The difference between copy+paste+yes+yes+yes+open-pr and asking n relevant questions about the code is what tends to matter.
Also, suppose after a few months you need to look at that code again. Having the possibility to go back and look at what was the human/agent conversation about it helps clarifying which tradeoffs where being considered.
Sticking to PRs (and meat proxy), a discussion over which reasoning led to that code should help a team pick and choose reliable collaborative patterns.
Overall, I don’t think this is a panacea, simply that not storing it feels wrong, to me.
I think it might be more useful to store a summary of the key decisions, but with mistakes, AI fumbles, iterating on test failures, and all the noise boiled out. If you kept, maybe as part of the commit log a summary of the AI discussion leading to that commit I could set that as very valuable.
I went there, and walked back to full sessions, mostly to be able to resume chats.
As a general principle, why would i want to lose precision? A senior might care about seeing a junior conversation to provide advice. A colleague might spot another one making as different assumptions about X, and have a chat about it. AI grepped file XX instead of XY and made a wrong decision based on that.
An engineering manager might care about completely different metrics.
But to your point, think it's fair to say that there isn't a single correct answer to this: different people like different stuff and at eventually standards will emerge (or not).
Just have devs commit (or provide however you like) their ai coding sessions.
It then becomes obvious if someone has, or has not, reviewed the code, because then they discussed it.
Bonus points: on very complex tasks, you can more or less resume the conversation, via session files.
Same for schoolwork, research papers, and a lot more.
The conversation is not something to throw away: it is arguably as important as the code in the pr.