I began utilizing AI as an exterior reminiscence. One way or the other, it made me really feel extra human — which feels prefer it ought to be a much bigger contradiction than it’s, however right here we’re.
For the final couple of months, my life has felt like pure chaos. Not dramatic chaos, essentially. Not the type the place one big factor occurs and every part collapses round it. Extra just like the quieter, extra exhausting type, the place there are too many open tabs in your precise life, and regardless of what number of you shut, 5 extra seem, smugly, like they’ve been ready.
Work. Writing. Analysis. Private issues. Challenges I give myself, as a result of apparently those life palms out without spending a dime weren’t sufficient. Journey planning. Extra journey planning. Three totally different journeys, by some means current in my head on the similar time, none of them totally booked. Residence searching. Life choices. Selections that aren’t technically pressing however really feel existential anyway. And, beneath all of it, the acquainted background nervousness of being 27 and nonetheless not having found out one single steady factor about what I’m doing with my life — which is a really dignified method of claiming I’ve a colour-coded calendar and no concept what I’m doing.
In some unspecified time in the future, I informed myself: give me a break.
However a break isn’t actually doable when even an enormous planner isn’t sufficient to assemble all of your day by day quests.
I personal a number of planners. Stunning ones. I just like the fantasy that if I purchase the fitting pocket book, my life will lastly turn into legible.
Plot twist: it by no means does. It simply turns into a really well-organised document of how ungovernable I’m.
So I turned to the one factor I’m really good at
So I realised I wanted to make use of expertise on myself, which, to be truthful, is one thing I’m good at. I used Claude Cowork to retailer my very own reminiscence — fairly actually.
Yesterday, I gathered all the newest issues I’ve: on-line, offline, inside my laptop computer, inside my chats with ChatGPT and Claude, emails, even some messages I’ll admit to having saved, my very own reflections, and the goals I write down after I get up and nonetheless, in opposition to all proof, consider imply one thing.
I saved all of it — all my plans, all of me — on Claude. Actually, I saved myself in AI and created a Git repository for it. Non-public one, after all. However nonetheless.
A grown grownup, version-controlling his personal character.
I created a residing file that can get up to date the following time one thing occurs in my life. One thing I’ll share with AI — and belief me, it should occur, as a result of one thing at all times does, often three issues without delay, often whereas I’m making an attempt to e book a flight.
It feels mechanical. It’s additionally a bit of scary
Regardless that it feels mechanical to have an exterior drive for reminiscence, and actually a bit scary too, I really feel precisely the best way my Google Drive ought to really feel each time I clear it of undesirable recordsdata — which, if I’m trustworthy with myself, occurs about as usually as I really use these planners.
So it’s an excellent feeling. To date, not less than. Ask me once more after the AI has learn my dream journal.
There’s one thing to that discomfort value taking significantly, although, as a result of I don’t assume it’s simply me being dramatic a couple of file construction. Reminiscence appears like essentially the most intimate factor we now have — not simply data, however texture, temper, the model of you that existed when one thing occurred. Handing items of that to a bit of software program felt, at first, like turning a life into recordsdata. Which, in equity, is strictly what I did, on goal, at 11 pm, with a barely manic sense of accomplishment.
However then I realised one thing a bit of deflating: I had already accomplished that. My life was already scattered throughout Google Docs, Notes app pages, screenshots, and paperwork named one thing like “new ideas closing closing perhaps.” I hadn’t been protecting my reminiscence pure and untouched by expertise, like some form of monk. I had simply been protecting it badly, and calling the badness “being a free spirit.”
Philosophers, unhelpfully, received right here earlier than I did.
In 1998, Andy Clark and David Chalmers proposed the prolonged thoughts thesis — the concept cognition doesn’t cease on the cranium, utilizing the instance of a person named Otto whose pocket book does the remembering his mind can’t. My Git repository isn’t Otto’s pocket book and I wouldn’t have his situation, however the underlying level held up higher than my ego wished it to: the thoughts was by no means as sealed off as I appreciated to think about. It’s simply that most individuals externalise their reminiscence badly and unofficially, and I made a decision to do it badly and formally, with model historical past.
A container, not a alternative
I wish to be clear that this isn’t the “AI will resolve my life” mind-set.
I don’t consider that.
Actually, I discover that form of considering a bit of harmful, and in addition a bit of insulting to how arduous my life has labored to be difficult.
I don’t desire a machine deciding who I’m, what I ought to do, or what my life means. I simply wanted a spot to place issues, as a result of the place I used to be utilizing — my very own head, largely out of stubbornness — was full.
There’s a psychological idea for this that I solely seemed up after the actual fact, which tells you every part about how this venture really began: transactive reminiscence, described by Daniel Wegner within the Nineteen Eighties from learning {couples} who cease every making an attempt to recollect every part and as a substitute break up the job — one remembers birthdays, the opposite remembers the place the passports are. Apparently, I’ve now assigned this position to a chatbot, which both means I’ve optimised my life or given up on discovering a associate who’s good with passports.
Presumably each.
What really modified
The unusual half isn’t that AI helped me bear in mind extra. The unusual half is that when I ended utilizing my mind purely as storage, my precise considering received higher. Considering, it seems, isn’t the identical as hoarding data, regardless of what number of years I spent working as if it have been.
When every part was inside my head without delay, I used to be not essentially considering deeply. I used to be simply drowning quietly in my very own archive whereas presenting, outwardly, as a functioning individual with opinions. There’s precise analysis behind this — cognitive offloading is the time period for utilizing an exterior device to cut back the psychological load of holding data your self, and up to date work modelling it as a value-based resolution discovered folks offload extra precisely when the price of holding every part themselves is highest — when reminiscence load is biggest and inside storage is most strained. Reader, the price was excessive. The price had been excessive for months. I used to be just too busy holding every part to note.
As soon as issues had someplace to go, I may really take into consideration them. I observed a fear from three weeks in the past was the identical fear, sporting a special costume. I observed a call I saved circling was one I had already half-made in a chat I’d forgotten existed.
None of that’s the AI being intelligent. That’s simply what occurs once you’re now not the one submitting cupboard.
After all, I nonetheless assume this requires warning, and I say that largely to reassure myself.
Privateness issues. Boundaries matter. Not each feeling must turn into knowledge, and a few issues ought to most likely keep messy, on precept, even for somebody as dedicated to methods as I apparently am. However I’m not utilizing AI as a result of I wish to cease being an individual.
I’m utilizing it as a result of being an individual has turn into unusually administratively demanding, and nobody warned me that turning 27 would really feel much less like a milestone and extra like being appointed, with out consent, as my very own venture supervisor.
Possibly what I constructed isn’t a second mind — a phrase I’ve by no means appreciated, because it implies the self will be optimised right into a clear, environment friendly duplicate. What I constructed is nearer to a second reminiscence room, someplace I can go away a field, label it badly, and are available again to later.
AI can maintain the file. I’m nonetheless the one who has to recognise myself in it, dream journal and all.
