The news item that opened the segment was small enough to sound like a marketing gimmick. On the Moonshots panel, host Peter Diamandis reported that AI tokens are turning into a kind of consumer currency in China: banks hand them out as credit card rewards, China Telecom sells access to 142 AI models the way it sells a mobile data plan, and restaurants give you compute credits after a meal. He read out figures he said had been reported that week, putting China's daily token consumption at a 5,000-fold increase in two and a half years. "Airline miles for thinking," he called the perks.
A token, in this sense, is simply a unit of AI usage: the chunks of text a model reads and writes, metered and billed. Giving them away as loyalty points means giving away a quantity of thinking.
What followed was a disagreement about two different ways the gains from that thinking might reach ordinary people. One is access — everyone gets intelligence. The other is ownership — everyone gets a share of the machines doing the work.
"Universal basic tokens"
Alex picked up the reward-points story and extended it. He began by clearing away the obvious confusion: the tokens at issue here have nothing to do with crypto tokens, which he described as essentially irrelevant in China. These are "ones that embody individual units of superintelligence and have nothing to do with sort of zero-sum financial exchange."
The confusion is real enough that another panelist had run the experiment by accident. He had made a T-shirt reading "less talking more tokens" and worn it on an earlier episode. About a third of people thought he meant crypto, he said; another third thought he meant bong hits.
Alex's forecast is that credit card points are the first step toward something deliberate. In what he called "token socialism," the instruments of state power — in China the party, elsewhere some Western near-equivalent — begin issuing intelligence as a redistribution, whether through reward schemes, welfare payments or another mechanism. He gave it a name: universal basic tokens, UBT. He expects the pattern to "blanket all of humanity."
He also thinks it is the load-bearing form of abundance. Healthcare, utility pricing, food, shelter, education: in his account these are all downstream of making compute universally available. Emad Mostaque said South Korea had announced universal AI for its people through a consortium, and that this idea sits at the core of the Champion initiative his organization is pursuing — locally owned national AI institutions, now recruiting founding councils. Rather than campaigning to ban AI, he said, Bernie Sanders should be arguing for "tokens for everybody. Make it a universal right."
Mostaque's book overview, The Last Economy, sets out the same idea as universal access to intelligence: a baseline of computation, a personally sovereign AI and shared open knowledge, with institutions redesigned around abundant machine capability.
The objection: cheaper tokens, bigger bills
The panel's stock phrase for where prices are heading is "too cheap to meter," and one panelist refused to let it stand. Is electricity too cheap to meter? "No, obviously not. It's expensive. We just use money."
His point was not that costs stay high. He accepted that the price of intelligence is falling by a factor of a hundred, and possibly a million. But what you can do with 5,000 or 500,000 or 5 million agents — AI systems that carry out tasks on your behalf — is, in his words, "mind-blowing." Nobody who has that capability wants fewer agents. "So the budgets are going to be real." His conclusion: "I think the use cases go to infinity at the same rate that the costs come down."
That leaves free tokens as something closer to a head start than a permanent entitlement. If someone gives you free compute, he said, grab it and savor it, because it gives you a chance to get on the map and get ahead.
Mostaque framed the stakes competitively: on his account, the average Chinese person and their AI will be smarter than the average American and their AI, and the average Chinese person's robot more physically capable — unless Americans and Europeans change their mindset. A panelist put the gap in public attitudes at 80-plus percent pro-AI in China against 80-plus percent against in the United States, describing it as a point the show had made before rather than a survey result.
From access to ownership
Asked about the future of jobs, Mostaque said the models have reached a level where the digital workforce could be replicated within a year, with physical work following. Jobs will not be created fast enough, he argued, but abundance can be. His answer was not more tokens; it was equity.
He had already proposed the Champion structure — a dollar pre-money, everyone in, with children owning the equity of the robots and other assets. Now it came with an addition. "But actually we can do one better," he said. The government should run a massive infrastructure program to build 100 million robots in America, "and they should be owned by the people." If you want an end purchaser for all those machines, make the people the owners; the same build-out, he said, would upgrade American infrastructure. He cited Elon Musk's remark at the G20 that a robot can do the work of five humans.
Who holds the deeds
Alex objected immediately. He does not think it advisable, at least in America, for the government to own 100 million robots. Told that the owners would be the people rather than the government, he was unmoved: how do the people own 100 million robots except through some centralized state, which "to my maybe overly American ear, smells like, wait for it, like luxury automated communism." He is not in favor. "I would like to see every American owning a thousand robots. I don't think they necessarily need to be socialized or communally owned."
His reason for caring is that he expects no occupation to be spared. Every profession as currently constructed is "cooked," he said, and what varies is sequencing — which work gets automated first, which in his view is what social policy has to respond to.
Mostaque then narrowed the distance. The government does not need to own the robots, he said: sovereign wealth funds are the Champion idea, "or just the government can underwrite the robots that the people can own." Underwriting means the state stands behind the financing while the shares sit with citizens. "I think the main thing is you have to get the ownership of the robots to the people somehow."
That is where the two mechanisms part company. Universal basic tokens would give everyone the use of intelligence, on terms someone else sets and can withdraw. Equity in the robots would give people a claim on what the machines earn, and a say in who directs them — which, as both proposals concede, still depends on somebody writing the structure before the robots are built.