Currency
When you change your mind, does your AI know?
You reverse a decision — a new constraint, a better idea, a lesson learned. The question that decides whether your AI helps or misleads is simple: does the thing feeding it context retire the old call, or keep it sitting there looking current?
Real work reverses calls constantly, and most context stores only ever add: a memory keeps every mention, a file gets appended to. So the old position and the new one end up side by side with no marker of which stands. What fixes that isn’t more retention — it’s supersession, the move Unl is built around: the old call retires, its history kept, and the current one is what’s served.
Changing your mind is normal; carrying both is the problem
Reversing a call is not an edge case — it’s most of what thinking is. The trouble is a store that only adds: it holds your old position and your new one together, and nothing in it says which one you actually stand behind now. Two answers to the same question, both looking equally live.
Retention is a virtue for an archive, a liability mid-decision
Never forgetting is exactly what you want from a log, and for an autonomous agent trawling its own history it’s the right property. It’s the wrong one when you ask where things stand and the model surfaces the call you reversed last month — confidently, because nothing told it that one retired. The mismatch isn’t the retention; it’s the environment, where a person mid-decision wants the position that holds, not every position they ever held.
Supersession is the missing move
The human version of changing your mind isn’t deletion and it isn’t accumulation — it’s supersession. The old call retires, its history kept; the new one stands. That’s a gesture: you say this replaces that, and the store tracks which one is current. A static file and a similarity-based memory both lack that move, which is why both can hand you a decision you’ve outgrown.
Why the old call is more dangerous than no call
A missing decision makes the model ask, or flag that it’s unsure. A stale decision dressed as current makes it act — confidently, on something you’ve already moved past. Outdated-but-authoritative does quieter damage than absent, because nothing prompts anyone to check it.
What supersession looks like
Unl makes changing your mind a first-class move: supersede the call and the old one retires with its reasoning kept, while the current one is what’s served when it bears. You stay in command — held on your word, retired on your word — so your context stays a single standing position rather than a pile of everything you’ve ever thought. Tuned for Claude, Claude Code, ChatGPT & Cursor at launch, extending across the AI ecosystem. Connects anywhere MCP does.
Changing your mind is normal; the danger is a store that keeps the old call looking current — what you need isn’t more retention but supersession, the old decision retired and the new one served, so the answer comes back measured against what you already decided, and why.
Reads through Unl arrive with measured context — in the presence of the decisions you’ve already settled. The reach lane is live: one box, paste anything. If it speaks MCP, Unl can reach it. Readings arrive unprompted, the data beside the criterion; Unl is a courier, not a warehouse, and keeps only your keys and the frame.
Read further
Questions people ask
Does AI memory update when I change my mind?
Usually it adds rather than replaces: the new mention is stored alongside the old one, and both can resurface. Without a supersession step, nothing marks which decision stands now, so the model can serve the call you reversed.
Why is an outdated decision worse than no decision?
A missing decision makes the model ask or flag uncertainty. An outdated one dressed as current makes it act confidently on something you've moved past — so a stale call does quieter damage than an absent one.
What is supersession?
It's the move where a new decision retires the old one: you say this replaces that, the previous call steps down with its history kept, and the current one is what's served. It's how a position stays singular as your thinking moves, instead of accumulating.
Do I lose the old decision when I supersede it?
No — it retires, it isn't deleted. The reasoning and the history stay, marked as superseded, so you can see how the call evolved. What changes is which one is treated as current and served when it bears.
What this is
Think inside your AI world — you stay in command
Unlimitless (Unl to friends) holds what you've settled, reads what your tools are showing, and catches what's changed out in the world — and hands your AI whatever bears on the work, the moment it's needed, without you asking. The right thing, in front of the model, unprompted, with you in command of the call. So you keep moving toward what you set out to build, on top of everything you've already decided.
It plugs into Claude, Claude Code, ChatGPT and Cursor as an MCP connector. Quick to connect, in a couple of steps.
Unlimitless is open now to invited Alpha. Apply for the Beta waitlist to come in ahead of the full launch:
Alpha is invite-only · free at launch.