I'll Be in London For a Week
Also, a few thoughts on what I'm thinking and writing about
Housekeeping:
I’ll be in London from May 29-June 8, attending two conferences:
The former is going to be me networking like a fiend. I am trying to get my larger narrative on tech extensity & the erosion of friction on the ‘things to care about’ radar. EAG is a good community for this sort of thing. I’m also going to the Foresight Institute for similar reasons, though probably with less overt networking. But who knows! If you’re at either event and we haven’t already chatted, please let me know!
Between that period, I’ll probably be camping out at LISA during the day for some of my week in London, because I have awesome friends, and I enjoy being in the same gravitational space as smart people.1
If you would like to meet up, talk about AI, gradual disempowerment ,power concentration, privacy, black holes, beer, or hell, anything else, hit me up (carey@priva.cat) or hit the message button.
Separately, I realized after writing the ‘Did AI Kill Publius’ piece that nobody read, that hiding behind a pseudonym that all the LLMs already “know” is me is kind of a pointless waste of time. So, hi. If you’ve been following Privacat and thought ‘That Cat person is very interesting. I like Cat, even if she’s a little angry —> Cat is me, and my real name is Carey.
My articles are very long. I’m trying to fix that.
So, I’ve been laser-focused on trying to distill my tech extensity argument into something that is shorter than five blog posts that are 12,000 words each.
And importantly, I’m trying to strike the right balance between ‘this is important’ and ‘we’re all fucked’ on the see-saw of concern. I actually have ideas for what might work, if we can convince the powerful to shift their focus a smidge.
But I get it. Nobody has time to read long, depressing blog posts. So, I did the unthinkable: I managed to get my thesis to less than 1k words and to not have it sound totally doomer-y. You can find the whole thing here: https://consult.priva.cat/tech-extensity.
It’s broken into a few sections, but this is still a work-in-progress. Please read it and poke all the holes. Seriously. Tell me why my argument sucks. I can take it. Anyway, if 994 words is still too long, here’s a summary.
The problem
Current AI policy & governance efforts keep proposing the same set of tools to address harms that come from AI and power concentration — fines, audits, transparency, human-in-the-loop, etc. But these solutions are wrong, because they assume that de jure laws will continue to have leverage over companies who are amassing nation-state level power by building tools (AI, compute, infra, surveillance, communications) that make us all increasingly more dependent on them.
This is what I define as tech extensity: when tech firms and the tools and services they provide become so embedded in the complex systems that society is dependent on, that they become impossible to remove or constrain.
What extensity is — and isn’t
In this section, I distinguish why tech extensity might look like monopoly, but it isn’t. And why extensive systems & firms encompass more than just vendor lock-in and regulatory capture.
Why this matters for AI governance now
Here, I touch on three related concepts:
The AI/tech scale-up craze isn’t producing new monopolies. It’s producing new substrates — compute, foundation models, satellite + sensor infrastructure, surveillance & machine-human integrations.
AI & new tech are eroding the friction points that have historically constrained most concentrated power grabs & collective disempowerment.
Finally, the outcome of all of this, if nothing changes is billions of disempowered people. This would suck.
Tech extensity is a unit of analysis, not just a theory
Treating extensity as the unit of analysis (rather than, say, “market share” or “AI risk”) changes the policy focus, because
It predicts which interventions don’t work. Procedural mitigations (post-hoc transparency, compliance theatre) cannot bind a system when compliance is gameable or ignorable, and enforcement is easily locked up in litigation.
It identifies what does work. Structural remedies that reduce interdependence (e.g., enforcing structural separation, bolstering institutional power, limiting power concentration, etc.) are interventions whose shape better matches the problem.
It is testable. I think whether an entity is becoming too big to regulate is quantifiable. I’m still working on the exact math here, but the end state is that a policymaker will be able to tally a score, run some predictions, and target interventions precisely (think: more Digital Markets Act, and less EU AI Act).
It offers a European delivery vehicle. The EU is the one jurisdiction with both the regulatory instincts and the dependency exposure (80% of the EU’s tech stack is foreign-owned) to take this question seriously. Building effective and enforceable approaches with actual consequences before this brief window closes is crucial
I close with my next steps. Expect to see more posts exploring extensity, power concentration, the negative consequences of unconstrained AI/corporate dependency, and how we can fight back. Also, I really want to find folks to collaborate with and ways to communicate this message more effectively (hence going to these conferences).
If you want to explore topics around tech extensity, ungovernability, gradual disempowerment, and power dynamics, please drop me a message. Or check out my fancy new consultancy page.
I would like to remain very far away from a black hole or a world where a handful of billionaires own the world though.