Solving Problems in Abundance

An essay

Bread is cheap everywhere now. In Tokyo there are bakeries where people queue ninety minutes for a single loaf of milk bread.

Both of those things are true at the same time. That is the central weirdness of living inside a culture of abundance. The thing that used to be hard, calories, became easy, so it stopped being the point. The point moved.

You can see this pattern across every domain that has crossed the abundance line. Food is everywhere, so food becomes craft. Fabric is cheap, so fashion runs on six-month cycles of micro-aesthetics that change before the last cycle has finished. Recorded music is free, so the parasocial economy around individual musicians is louder than the top 40 was in 1995. Even attention, which is supposedly the new scarce thing, has been fractalized into a thousand sub-niches that no media planner from 2005 would recognize.

When everything is abundant, what stays scarce is curation. Setup. Care. The slight spin on the familiar. The thing that says someone, somewhere, decided this loaf was worth queueing for.

Code is not going to zero

It has become fashionable to say software is going to zero. It is, in the same way bread went to zero, which is to say for the trivial cases. CRUD apps spin up in minutes. Landing pages are a prompt and a deploy. The category of “I needed this thing built and now it is built” has compressed by something like three orders of magnitude in two years, and is going to compress by another order before anyone really notices.

That is not the same thing as “code is finished.”

The world has nowhere near enough code, and nowhere near the right kind. We are not building software for a finished planet. We are building it for a planet whose actual hard problems have not been touched. A space elevator is a software problem before it is a tether problem. A Dyson swarm is a software problem at every layer, from autonomous fabrication on Mercury to the planetary grid that takes the output. Moon bases need closed-loop life support that nobody has shipped. Real mass robotics, the humanoid-in-every-warehouse kind, needs simulators and safety policies that current frameworks gesture at and do not deliver. Pick any frontier and the codebase is mostly negative space.

We are very far from peak technology. Brian Potter’s Origins of Efficiency (Stripe Press, 2026) walks through how production has gotten cheaper across centuries: penicillin, steel, semiconductors, ammonia. The trick that runs through every chapter is that nothing about the resulting cheapness was inevitable. The cost curves did not bend on their own. Efficiency happened because someone fought a specific friction and kept fighting it. Abundance is the running total of fights already won. The fights that remain are bigger than the ones we have finished.

The new scarcity is curated signal

If software is plentiful, the scarce thing one rung up the stack is signal. Taste. Judgment. A high-quality opinion about which problems are worth solving, defended by someone who has actually been near the problem. A piece of work shaped by an actual model of what is broken in the world.

Most of the demand for AI right now is for the opposite of that. Most users want the cheap version, in volume. The slop. The slop is comforting and easy, and it is what most of human consumption has always been. People in 1850 read penny dreadfuls. People in 1950 watched television. People in 2026 watch six-second loops of someone reacting to another six-second loop. Entertainment is a load-bearing human need, not a flaw. There is a giant business in providing it, and someone is going to win that business. It will not be most of the people reading this.

The other layer is smaller and stranger. It is the people who notice that something is broken and cannot stop thinking about it. The frustrated ones. The ones who got rejected by their insurance company and started reading legal code. The ones who waited three years for a planning permit and built an agent swarm to harass the city. The ones who tried to learn calculus from a popular MOOC and realized the entire genre of online education is, structurally, daycare for adults.

Their attention is becoming the most valuable signal in the system. Not because they are smarter. Because they are pointed at something specific and refuse to let it go.

The strivers split into two camps

Watch the people around you who claim they are building. They sort, very cleanly, into two groups.

The first has their attention locked onto a specific broken thing. They would be working on it whether or not anyone paid them to. They can describe the thing they are fixing in one sentence, and the sentence has names and dates in it. Their problem set has not changed in months. What has changed is the depth of their understanding of it.

The second group is tab-switching through demos. They posted a thread last Tuesday about which model is best this week. They are running a vibe check on a new tool. From the outside they look very busy. They will look very busy for years. Abundance is generous to them. There is always one more shiny thing to spin up, one more launch to react to. Time quietly evaporates inside the noise.

The first group ships. The second group performs. The line between them is whether the problem owns the person, or the person is shopping for problems.

We are very far to the right of the curve

David Deutsch, in The Beginnings of Infinity, makes the case that knowledge creation has no upper bound. There is no problem in physics, biology, or governance that is permanently unsolvable, given enough good explanations and enough time. Most people do not build that way for one simple reason: the world is good enough most of the time to not feel like it needs fixing.

Most people in the developed world will live their entire lives inside a version of civilization that already works. The water comes out of the tap. The plane lands. The phone rings. The food is in the fridge. The gradient of “this could be ten times better” is invisible from inside it, the same way it would have been hard to convince a 17th century farmer that he needed penicillin when he had not yet seen a child die of strep.

That is fine. The job of civilization is partly to be invisible. But it produces a strange selection effect on builders. The people who do see the gradient tend to be the ones for whom one specific thing has gone wrong in a way that gave them taste. A medical condition nobody was treating. An industry that wasted their decade. A piece of software they use every day that, on inspection, is held together with duct tape and the will of two open source maintainers in Lithuania.

Those grievances are the renewable resource of progress. Abundance does not eliminate them. It buries them under a thick layer of comfort, and the people who can still feel them through the padding are the ones with something to do.

How to solve problems in abundance

A few practical notes from inside the curve.

Start with friction you have personally hated. Not friction in the abstract. Specific friction, with a date attached, that ruined a specific Tuesday for you. The reason this matters in the abundance regime is that there are ten thousand plausible problems to work on, and your only filter against the noise is having actual stakes in the answer. The market for “AI for X” is saturated with people who have heard X is an industry and have never spent a day inside it. The market for “I have been inside X for ten years and the thing that drove me crazy was Y” is structurally less crowded.

Skip the comfort layer. Other people will win that. Win the layer above, where the unit of competition is taste and density of insight. Pick a sub-domain so specific that you can be the world expert on it within a year. Originality at this point is going deeper into one particular thing than the slop has the patience to follow.

Build instruments, not just products. Abundance generates noise. The thing that holds value is the apparatus that filters the noise: dashboards, calibrated forecasts, scoring rubrics with skin in the game, anything that lets you tell what is real from what is hype. Prediction markets are an instance of this. So is any tool that drags a vibes-based field into measurable territory. The supply of vibes is infinite. Someone has to do the metric work.

Get fluent with agents. This is mechanical but it matters. The people who solve abundance-era problems will not be writing line-by-line code, and they will not be typing prompts into a chatbot either. They will be running fleets of agents at problems with taste about which agents and which problems. If you have not yet built a small thing where five agents argue with each other and one of them ships a PR, do that this week. The intuition compounds fast.

Be willing to look uncool. The valuable problems left in 2026 are mostly unglamorous. Tax code. Local zoning. Permit timelines. Battery supply chains. Concrete. Intermodal freight. The reason they are still problems is that they have rejected the last twenty years of attempts. They will reject most of the next twenty too. The ones who break them open are the ones willing to spend two years bored on something nobody will tweet about.

Pair seriously with someone faster than you. Ambient ambition is the best resource you can put yourself near. The thing the great periods of building all share, from the Royal Society to PayPal, is a small set of people who saw each other every week and refused to let each other coast.

The shape of the next decade

The question “how do you solve problems in abundance” mostly answers itself once you stop treating abundance as the destination. Abundance is the floor that lets you stand up and look at what is left to build. The floor is real. The view from it is enormous.

Most people in this generation will spend their whole lives inside a civilization that works well enough not to need them. Most will not need it to be otherwise. A few will. The few are who the next century is going to be made of.

If you can feel the gradient through the padding, you have already done the hard part. Pick the friction you cannot stop thinking about, and start.