Welcome!. This is my first post on this site, and that small fact is worth acknowledging. I have started blogs before — started them with genuine enthusiasm, written one or two posts with care, and then watched them quietly go dark as the gap between having something to say and actually sitting down to say it grew wider than I could manage. Life moves fast, good ideas don’t always survive the commute from brain to keyboard, and maintaining any kind of cadence requires a discipline I have historically not had.

What feels different this time is that the tooling has changed. The distance between “I should write something about this” and “there is a published post” is shorter than it has ever been. I am not suggesting that the writing is done for me — I find I still very much want to do that myself — but the scaffolding, the formatting, the friction of getting something from a rough thought into a real page has all but evaporated. So this is my attempt at a more regular cadence, helped along by the AI tools that are quietly reorganizing how I spend my time.

Which, as it happens, is exactly what I want to write about today.

---

I have been thinking about what this AI moment actually means, not in the overheated way the discourse usually goes, but more slowly, the way you think about something while the coffee is still hot and nobody is demanding anything of you yet.

The headline version of the story, the one that makes people anxious, goes something like this: large enterprises are shedding tech roles, entry-level positions especially, and there are fewer of those jobs today than there were a year ago. If you are a recent CS graduate who was counting on a certain kind of hiring environment, the ground has shifted underneath you in a real way.

The same tools that are letting enterprises run leaner are also handing an extraordinary capability set to individuals. Web development, back-end infrastructure, marketing copy, bookkeeping, legal boilerplate, customer support automation — all of it is now within reach of a single motivated person with a laptop and a few API keys. The barrier to starting something, to actually building and shipping a product, has dropped faster in the last two years than in the previous twenty. That is not a minor footnote to the job-loss story. It is a story of its own.

So you end up with this strange, simultaneous dynamic: enterprises are shrinking their technical footprint while, at the same time, the same tools allow people to credibly compete with the enterprise capability sets with these new tools. The marketplace is not just contracting in one place; it is expanding in another, and the expansion is happening much closer to the edge — in individuals, small teams, and ventures that a few years ago would never have existed at all.

And here is the part that I think gets underappreciated when people talk about incumbents weathering this transition. The large enterprises do have advantages that matter — market share, brand recognition, distribution, capital. Those are real. But they are also dragging an enormous anchor. The systems that power most large enterprises were built in a different era, before any of these AI-native abstractions existed. They are a tangle of legacy code, acquired technical debt, and organizational inertia, and business continuity demands they maintain those systems even as they are actively obsolete. You cannot easily rebuild a fifteen-year-old enterprise product from scratch, no matter how much you might want to. The new entrant faces no such constraint. They can build on the modern stack, natively, from day one, and that architectural advantage is not trivial.

This is where I find a useful metaphor, one that has been rattling around in my head for a few weeks now.

Think about what happened to media and publishing when Facebook and Twitter arrived. Suddenly every person had their own nano-publishing company. You could reach an audience, cultivate a following, build a brand, and produce a content operation that would have required a full editorial staff a decade earlier. A whole commercial ecosystem sprang up around it. The incumbents — the newspapers, the magazines, the cable news channels — had the audience, the reputation, the infrastructure. But they also had the cost structure, the institutional habits, the physical plants and union contracts and legacy CMS systems. And the new entrants, the individual creators and substacks and newsletters, could be nimble in a way that the old guard simply could not.

Instagram and TikTok did the same thing for broadcasting. Suddenly every person had a nano-broadcasting company. The production quality gap that once kept individuals off-screen collapsed, and an enormous amount of attention migrated toward individual voices and away from institutional ones. Again: the incumbents had reach, but the new entrants had speed, authenticity, and a structural cost advantage that was very hard to compete with.

So now I am sitting here on a Sunday morning wondering: what happens when Anthropic hands everybody their own nano-software company?

I do not think this is a rhetorical question. I think it is the actual question, and I am genuinely uncertain how to answer it. If the publishing analogy holds, we should expect a long tail of small software products that would never have been economically viable to build before — highly specific tools for niche problems, built and maintained by one or two people, reaching audiences too small to have ever attracted a proper venture-backed company but still large enough to sustain something. We should expect incumbents to lose share in categories where they were already coasting on switching costs rather than actual product quality. We should expect the interesting new software, increasingly, to come from outside the traditional institutions.

But the analogy also has limits. Software is not content. There are network effects in software that do not exist in a newsletter. Enterprise software especially has switching costs that are almost gravitational. And the kind of AI-assisted building that is possible right now still has meaningful ceilings — it is very good at standard patterns and gets complicated at the edges.

Still. The direction feels clear even if the magnitude is not. The asymmetry that used to define the technology industry — where building anything real required a team and capital and time that only a certain kind of organization could muster — is eroding. And erosions of asymmetry tend to produce more interesting markets over time, even when they produce painful dislocations first.

I find myself, on balance, more optimistic about this than anxious. Not because the disruption is painless — it clearly is not — but because a world where more people can build things that matter, on their own terms, seems like a net improvement over one where the ability to build was concentrated in a handful of institutions and the career paths that served them.

More to think about. Happy to be wrong about any of it.