Archival Selves

What happens when you pay off all your intention debts?

If you’re like me, you’re already past the first hypomanic transition across the event horizon of Claude-Code-powered frenzied bespoke-personal-project execution paralysis. The flywheel has spun up, and you’re using up session token budgets as fast as they become available, and perhaps even into spending more (I’ve spent $50 beyond my Pro account limits so far). You’re probably deep into orientation debt, with fraying mental models of why you’re doing what you’re doing. Are you neck-deep in random acts of Claude-Coding, or is there more going on with you?

You’re probably wondering what comes next, and whether there is any larger logic to the frenzy. Is it just going to be one damn bespoke personal project after another from here on out? Or are there further levels we haven’t glimpsed yet? It’s worth pausing to take stock of where we are right now before attempting an answer.

Showing off your portfolio of bespoke Claude Code projects and looking at others’ portfolios is a new social activity that has already acquired the quality of campy tedium we associate with people in the 70s subjecting each other to slide shows of unremarkable vacations. Or people in the 80s and 90s inflicting VHS home videos on each other. As a medium, the Claude Code bespoke personal project (CCBPP?) is much more expressive, but the actual variety of CCBPPs coming into view is much lower than what the medium is clearly capable of. What should be an unruly wilderness bursting with diversity is turning out to be a landscape of Ballardian neoliberal mimetic life-script banality.

I’m no exception. My portfolio is as home-movie-banal as any other. Our collective challenge now is to get past this almost monocultural stage to the explosive wilderness and divergence stage that has clearly been unlocked. But it will take some work to get to that starting line. We’re all busy with backlogs at the moment.

The current banality goes deeper than most people simply being poor narrators of their personal journeys. Most people don’t have storyworthy life journeys to work with. So personal projects born of such lives reflect the poverty of the source material.

Archival selves. Made with my Bucket Art model on titles.

“Only when the tide goes out do you learn who has been swimming naked,” as Warren Buffett said. Mutatis mutandis, when a powerful narrative technology comes in, you see who’s been living without stories.

It’s not just new cliche of “notion obsidian to-do workflows” (the “not x but y” tic of AI-in-the-loop humans). There’s a much deeper poverty and banality to people’s lives being revealed, as they pave their life paths with AI-bespokification. And we can’t blame ourselves, really. The 20th century/early 21st late modern world turned people’s lives into degenerate caricatures of human potential expression. The more “successful” your life by normal scripts, the duller it looks when paved and made legible with AI bricks. The very potential for bespokification reveals the stark uniformity of people’s lives.

I suspect a lot of people are discovering the depressing truth that beneath gnarly superficial differences in their life logs and data exhaust, which requires bespoke code to clean up and parse, they are living lives rather similar to everyone else’s.

At least the young can be forgiven the uniformity. They haven’t yet had time enough for their base identities to stabilize, and they haven’t yet logged enough life to possess the banal raw material for “unique” self-presentations. But if you’re (say) 30+, you have some raw material to work with. If you’re 50+ like me, you have a lot of material to work with; a whole life-act’s worth.

Looking at my own Claude Code portfolio, it is striking the degree to which it is only “interesting” in direct proportion to my failure to execute the normie neoliberal life script. All my interesting projects are derived from adaptations to script failures.

Stepping back, it is even more fascinating the extent to which all my projects are rooted in my past, in things I’ve already partly done or tried to do (banal or not), rather than in the future, in things I hope to do.

A quick inventory (I won’t inflict screenshots or details on you). Of my 30-odd non-trivial projects, all evolving briskly at the rate of my Claude usage limits, probably 27 are based on my past.

• I have a couple of dozen book projects in flight based on series from my blog archives (which I count as 3-4 meta-projects at Claude Code level, based on transform pipeline similarities).

• I have a major project going to port my WordPress sites to static archival sites. One is done but not yet deployed (Breaking Smart), while the other one needs some serious re-architecting as a museum site (Ribbonfarm).

• I have another major project to transform my Roam graph for a future set of books (my Clockless Clock project refactored into a 3-volume trilogy that will take a decade to write, with Tempo retconned as a prequel, with the whole renamed Configurancy) into an Obsidian vault and a pipeline to cast that notebook-like material into chapter scaffoldings.

• I also have 3-4 technical research projects (in control theory and robotics) based on unfinished ideas I couldn’t pursue during my postdoc 20 years ago because I had reached the limits of my own knowledge and skills.

• I have a few administrative projects too. My big messy folder of 600+ PDFs is now neatly organized into a fully tagged and searchable library, with scripts for tagging, indexing and filing away any new PDFs I drop in there, and another for popping up a random PDF for me to read when I’m bored. I plan to do something similar to my photos (literal 70s vacation slide show descendent) I have several personal dashboards going.

All of this is moving along at a brisk canter. None of it is blocked. Claude Code unblocks everything at dirt cheap prices. You’ve already seen some output (the Twitter book and the Art of Gig Volume 3 book). You’ll see more starting a few weeks — I’m spending some time setting up some larger-scale factory-like scaffolding.

Amazingly, I don’t feel stuck with any of these projects. I know what needs to be done, and roughly how it should be done from a technical perspective (I have enough techno-managerial experience for that), and am doing it. This is a new experience for me, as I’m sure it is for most of you. I’ve spent most of my life feeling mostly stuck on most fronts. I simply did not have the knowledge, skills, and financial resources required to feel generally unstuck by default rather than stuck.

This is a radical new human condition. Only a tiny minority have experienced it so far, but it will soon become much more widespread (not universal though — the barrier to entry is higher than that).

What is notable is the complete absence of live, progressing projects that need to start from blank canvases and starter creative visions/attacks. I do have ideas for several such projects, and have set up empty folders for them, but only non-blank-canvas projects have gotten going. Claude Code has a bias for legacy projects that have a lot of starter raw material.

The entire manifest of projects constituting my Claude Code flywheel, I have come to realize, has to do with paying off intention debt, processing psychological baggage and incompletions I’ve been carrying around for years to decades, and dealing with a great deal that was only blocked by lack of grinder energy and raw execution leverage.

And it looks like it will all get done. To the point where I no longer have any intention debt left. An unprecedented personal-life singularity on the horizon, and within reach. And I’m not alone here. I see a bunch of people racing towards their own debt-freedom horizons. Byung-Chul Han is going to hate it, but we’re all treating life as a project and actually starting to finish it.

What happens when we all get there?

If you thought the initial mass hypomania and derealization we’re witnessing right now is an astounding sight, wait till most of us clear our aging, rotting intention backlogs and sit staring at blank canvasses for the first time in years or even decades. When we are faced with a life with more empty room than baggage to fill it with.

That will take a few months to a year, and only a fraction of those getting started now will likely actually clear their backlogs enough to experience the emptiness. It does take some discipline, psychological courage, and budget to keep going; Claude Code unblocks a lot but not everything.

What happens, I think, will have a lot to do with how we’ve cleared our backlogs of intention debts. Because the generativity of the blank canvas of the future will be framed by the choices we make in archiving the past.

Starting to clear my backlog already feels like starting to craft an archival self. Sachin has been writing some fascinating essays treating LLMs as representing archival time, and if extend that logic to all our slates of Claude Code projects, I think we’re all creating archival selves.

This isn’t going to be equally natural for everyone of course. You have to be between major chapters or acts of your life, in some sort of a liminal passage, for the idea of an archival self to make sense. It is definitely natural for me. I’m almost a decade into the liminal passage between my personal Acts I and II (yeah, yeah, I procrastinate a lot).

What will this archival self be like?

As I noted in the opening, the harsh truth is that the raw material of the archival self isn’t going to be that inspiring for most of us. But what potential it does have can be either poorly expressed or well-expressed. And whether the creation of the archival self feels like paying off psyche debts, or refinancing it, depends on how much thought and introspective rigor you put into the archiving. And how complete-able it is of course. Not all of us carry around baggage that’s easy to get rid of.

There are layers of analysis available here.

The first, and most obvious, layer is the layer of concrete artifacts you produce with AI assistance that constitute your archived self. In my case, it looks like it will take the form of a couple of archival websites, and a dozen new books, plus a few stalled or mothballed writing and technical projects resurrected and refinanced (in terms of intentionality and unstuckness, not capital). A second-order artifact ambition for me, since so much of my archival self comprises written text, is casting the archival self into a kind of oracular ghost of my own past. A model trained on my archives that I can talk to, as a memory prosthetic. I imagine others may also be interested in talking to my Act I self, but I plan to design it mainly for myself.

This first layer of the archival self is already an unsettling idea. A set of artifacts forming a cast-off, almost-alive ghost of my past that haunts my present and future.

The second layer has to do with the meaning of the archival self. Is the archival self merely a site for nostalgic wanderings down memory lane? A deeper source for future activities? I don’t know. Some projects that are “archival” to start with may become reanimated with new intentions. Others may feel like decisive amputations. I mostly have a pretty healthy relationship with my past. I don’t think there’s a whole lot of unprocessed trauma or deeply repressed intentions or baggage down there. I have no particular desire to fully amputate my archival self from my current and future selves.

But it is already obvious that for a lot of people, this second layer of the meaning of the archival self will involve some gut-wrenching pain and trauma processing. Claude-Coding them into an archive will feel like aggressive therapy. To the point that I suspect many people will abandon projects because the baggage is too painful to process. It will feel like some sort of past-present-future temporal dysphoria, embodied by personal projects.

Then there is the third layer. How the paying off of psyche debts creates entirely new frames for the future. We’ve all experienced minor versions of this. Back when I was a dedicated GTDer, I frequently experienced the catharsis of doing the big sweep of commitments required to initialize (or re-initialize after a derailing) a GTD workflow. But that kind of purely manual processing of your life’s inbox can never get truly deep, or dig fully into the foundations. You need AI assistance to go that deep.

I suspect getting to a proper AI assisted archival self will be to a GTD-sweep catharsis as an ayahuasca trip is to a few bong hits.

And finally, there is a fourth layer — creativity. Creating an archival self is not just a grinding process of parsing the archives of your life into banal vacation home movies unless you want it to be. There is both room and need for creative editorial decision-making. You are bringing a kind of print-like fixity to a currently fluid sense of your own past. The cost of this fixity is clear — you will curtail your own future abilities to rewrite your past. But the benefit of having a stabilized past will depend on the creativity with which the fixity is engineered into it. In creating an archival self, you are, to some degree, creating a work of fiction that is more or less true to the archival memory territory it rests on. But you are also creating a perspective and an orientation within that archival memory.

This fourth layer is hard to think about. I’ve started thinking about it as creating a ground-truth canvas for a future memoir (whether or not I write one). The process of creating an archival self is about creating a canonical self-authorship reference. Who knows, if it is set up well enough, it might even be able to actually write the memoirs, not just ground it.

That’s a four-layer stack emerging under your random acts of Crazy Claude Coding: Artifacts, Meanings, Future Frames, Orientation and Authorship.

And you don’t have to plan for this to happen. Your archival self is emerging whether or not you consciously intend it to or not, simply as a function of Claude Code being better at paying off the debts of your past than at scaffolding the possibilities of your future.

I’m probably about 30-40% of the way into archiving my Act I self. I think it will take about a year or two to get to almost 100% (assuming Claude Code remains available at similar or improving price/performance).

And then? It will be interesting times.