how your ai finds what you need in your memory, how to ask so it finds the right thing, and what it means when it finds nothing.
the fundamentals · step 4 of 5
how your ai finds what you need in your memory, how to ask so it finds the right thing, and what it means when it finds nothing.
a memory is only useful if the right piece of it shows up at the right moment. you don't have to open anything or copy anything. you just ask.
in any connected ai, ask the way you'd ask a colleague who was there:

the ai hands your question, word for word, to innernet. innernet doesn't send back your whole memory. it sends the few passages that answer it, a handful of things that were close, and a list of every page so nothing is out of reach. then the ai answers from that.
this is why it stays fast and cheap however big your memory grows: a memory with a thousand pages answers in the same small package as one with ten.
sometimes the honest answer is your memory has nothing on this. innernet says so plainly, and a good ai passes that on instead of guessing. that's a feature: it means everything else it tells you really came from your memory.
if you think it should have known, save it now: remember that the refund policy is 14 days. it's there for the next question.
a new thing reaches innernet in a moment, but netti folds it into the pages a little later. until then it waits in the project's inbox folder, and it's already included when you ask. you never have to wait for netti to be able to ask.
if your ai can show apps inside the chat, a small innernet line appears under its answers, saying what it just found or kept.
.](/guides/shots/ask-ledger.webp?v=f361c8ea)
one question is worth making a habit of, at the start of any session:

when a session ends, a good ai leaves a handoff: what got done, what's next. the next session, in any tool, reads it first. you'll find them in the project's handoffs folder.
the technical side, including exactly what comes back for each question, is retrieval in the docs.
next: branch a memory.
every chapter
innernet is memory your ai tools read live: every project, every decision, every skill, in one place, connected once.