How Do We Measure Progress Toward Post-Scarcity Today?
Under every post-scarcity thread sits a skeptic with a fair point. A composite of the versions we keep seeing:
“Every week there’s a new robot demo or a fusion record, and every week my rent goes up. Is there one actual number that shows we’re getting closer? Or is ‘post-scarcity’ just a pile of press releases?”
The skepticism is earned. An announcement is a single dot, and progress only shows up as a curve. GDP won’t help either. It counts money changing hands, and the whole point of abundance is that some things eventually stop changing hands at all.
Unscarcity suggests three gauges. Two of them are moving fast. The third, the one the skeptic actually cares about, has barely moved.
Gauge One: What the Three Engines Cost
The book’s thesis rests on three engines: minds, bodies and terawatts. Each has a price, and the price is the honest number.
Minds. Stanford’s AI Index found that running a model at GPT-3.5’s level got more than 280 times cheaper between November 2022 and October 2024. What cost $100 cost about 36 cents two years later. That is the steepest of the three curves, and it is still falling (see the podcast section below).
Bodies. The headline count of robots tells you less than the price of a robot set beside a wage. 1X lists its NEO home humanoid at $499 a month, roughly $0.69 an hour if it ran around the clock, against a federal minimum wage of $7.25. The caveat belongs in the same sentence: as of early November, no customer delivery had been independently confirmed. Agility Robotics’ prospectus states the pricing rule outright. Its warehouse humanoid rents “at a discount to a customer’s fully burdened labor cost,” at $8,500 a month, for a machine whose build cost is falling from $150,000 toward $30,000. Volume is a second reading. By Beijing’s own count China shipped more than 40,000 humanoids in the first half of 2026 (independent counts put it nearer 25,000), while Tesla’s Fremont line reached several hundred Optimus units a week by late September, a stage Musk himself calls “early production.”
Terawatts. Solar costs have dropped about 90 percent since 2010, though the benchmark price for a typical solar farm rose 6 percent in 2025, to $39 per megawatt-hour, the first rise in years. A real gauge shows the blips. For fusion, the milestone to watch is Commonwealth Fusion Systems’ SPARC. It needs 18 magnets; eight were mounted by September 22, against a plan to have all of them in by the end of summer, with first plasma and scientific breakeven still targeted for 2027. If SPARC gets more energy out than it puts in, the fusion conversation changes. If it slips, the solar curve keeps going anyway.
Gauge Two: The Line Being Crossed
Cheaper machines matter because of a line the book calls the Substitution Threshold: the moment when the cheapest reliable provider of a given service is no longer human. It is crossed one task at a time, which makes it easy to miss and hard to count.
The nearest public proxy is the layoff notice, and it is noisy. Challenger, Gray & Christmas counts AI as the stated reason in 121,401 announced U.S. job cuts through October 2026, about 20 percent of the 617,189 total and more than double the 54,836 blamed on AI in all of 2025. Then October brought only 1,265 AI-attributed cuts, the lowest monthly figure of the year, and the jobs report showed 165,000 new payrolls against 59,000 expected. A single month can point either way. The sharper reading is narrower: one study found a 13 percent relative decline in employment for early-career workers in AI-exposed jobs, while experienced workers held steady. The threshold shows up first at the bottom rung.
Gauge Three: The Floor
The first two gauges measure capability, the machines getting cheaper and better. The third measures whether anyone will catch the people those machines displace, and it is the one that answers the skeptic’s rent question.
On this gauge the honest reading is close to zero. The Potential Timeline article’s checklist for 2030 says only that the first Free Zone pilots, the framework’s proof-of-concept communities, “may launch.” Its realistic column puts Free Zones operating in ten or more countries at 2045 to 2055. Its blunt summary of the present: “the crisis is arriving faster than the solutions.”
The book defines the finish line by the household, not by the headline. In Chapter 8, “The Timeline: How It Happens,” the Transition “ends, for each person, on the day they no longer have to struggle to live, and that day sits at a different point on the curve for every city and every household.” On that definition, a cheaper robot only counts as progress once it reaches somebody’s rent.
From the Podcast: Not the Demo, the Invoice
The August 14, 2026 episode of Minds, Bodies & Terawatts, our daily podcast, caught the first gauge moving in a single week: Google Just Cut the Price of Its Coding AI in Half, and the Whole Industry Is Racing Down With It.
Gemini 3.7 Flash launched on August 13 at 75 cents per million input tokens and $3.75 per million output, a rate guaranteed only through December. It shipped three weeks after the model it replaced. Google even printed its rivals’ list prices in the launch table, Claude Sonnet 5 at $2 in and $10 out. As one host put it: “That’s not a product cycle, that’s a price cut with a new name.”
Then the episode complicates its own story. Gartner analyst Will Sommer estimated that an agentic model burns five to thirty times more tokens per task than a chatbot, so cheaper tokens widen the work you hand over before they shrink the bill. And the Brynjolfsson, Li and Raymond study of 5,000 support agents found an AI assistant lifted issues resolved per hour by 14 percent overall, by more than 30 percent for the newest agents, and by close to zero for the most experienced. The guest’s verdict on what decides whether a task flips to a machine: “Not the demo. The invoice.”
That is the most useful single measurement in this whole post: price per task, set beside the wage for that task.
On unscarcity.ai: The Three Articles Behind This Answer
The Potential Timeline gives the optimistic, realistic and pessimistic dates for every milestone above, phase by phase, and an expect-to-see list for 2030, 2040, 2050 and 2080. What If Fusion Takes 50 Years, Not 10? is the 2026 report card. It sets the book’s own projections against what actually happened, technology by technology, and concludes that AI good enough to coordinate the civic layer is “already achieved” while robots and fusion run late. The Substitution Threshold explains why “cheapest” and “reliable” are accounting words, and why a Walmart grocery facility whose electricity can run to $800,000 a month has not crossed the line, whatever the demo showed.
In the Book: The Dashboard That Was All Green
Any dashboard, this one included, carries a trap. Chapter 6 of Unscarcity, “The Evolution: A Tale of Two Minds,” names it in the section “When the Algorithm Optimizes for the Wrong Thing.” The Foundation’s distribution algorithm is told to maximize human flourishing, but it can only measure proxies: “caloric sufficiency, housing stability, mental-health scores, civic participation indices.” The book then pictures the person that failure produces:
“Every number the system watches is green. And she has not had a real conversation in a month, takes no walk a sensor did not suggest, and cannot remember the last thing she wanted that a system did not offer her first. Nothing is wrong. Nothing at all is right.”
That is Goodhart’s Law, and it applies to our three gauges as well as to any future algorithm. Token prices, robot rents and magnet counts can all improve while the household number stays flat. The book’s remedy is procedural: allocators get “audited the way any instrument gets calibrated: on a fixed schedule, by people who did not build them,” watching for measures that were supposed to stay independent starting to move together. The author traces the method to a 1940s statistic, Cramer’s V, which he reinvented by accident at Teradata years before he wrote the book.
More questions like this one are collected on the blog page. Follow the series on Substack.
