The Madness: We Are Paying Them to Take Everything We Know - and We Get Nothing Back The Madness: We Are Paying Them to Take Everything We Know - and We Get Nothing Back

Before I make the argument, and it is an argument that will make some people angry, I want everyone on level ground. I am going to put four ideas in front of you: AI job loss by 2027, Universal Basic Income, a Universal Basic Fee, and a sovereign wealth tax. If we do not define them clearly, the full picture does not land.

First, clear the air: four things to understand

1) The 2027 prediction is not science fiction

Serious analysts are warning that the next wave of automation targets cognitive work, not just repetitive physical tasks. The range of forecasts differs, but the direction is consistent: large-scale displacement is plausible, and the timeline is short.

Hold this clearly: a major AI-driven labour shock is plausible around 2027, and it targets the thinking class who believed they were safe.

2) UBI is the popular answer

Universal Basic Income says: if machines take jobs, pay people an unconditional baseline. This may be necessary in a transition. But the framing matters. The word income implies earnings. Under UBI, the recipient can feel positioned as a dependent rather than an owner.

3) Sovereign wealth tax is the proven nation-scale model

Norway taxed extraction of a national resource, built a sovereign wealth fund, and compounds the returns for long-term citizen benefit. The principle is straightforward: if a national resource is extracted, the nation shares the proceeds.

4) Universal Basic Fee reframes the contract

A Universal Basic Fee (UBF) reframes payment from welfare to royalty. Same cheque, different logic: not a subsidy for being unnecessary, but compensation for value contributed to the system.

UBI says: here is support because you are displaced. UBF says: here is your fee because your intelligence capital was extracted and monetised.

The question that changes everything

When you put your hardest strategic problem into an AI system, who pays whom? You pay the subscription, then supply high-grade judgement, process knowledge, and decision logic. The platform compounds from that interaction and sells increasingly capable outputs back into the market.

This is why the convenience feels so good and the economics feel so wrong at the same time.

We are not only being harvested. We are often paying for the machine that performs the harvesting.

The con hidden inside convenience

Every mature industry pays for core inputs. AI platforms carry large fixed costs, but the most strategically valuable input, human intelligence at scale, is frequently priced as zero at the capture layer.

This is not just another tool cycle

The standard comfort line is that each automation wave destroys some jobs and creates others. The challenge now is velocity and scope. If AI compresses retraining windows faster than human adaptation cycles, the old transition assumption weakens dramatically.

When the machine can learn new cognitive tasks faster than populations can retrain, the old historical analogy starts to fail.

The bubble and the pin

Capex has surged faster than realised returns. Several enterprise studies still report weak measurable ROI across early AI deployments. Layer on legal risk repricing and you have a fragile valuation narrative if courts continue assigning hard monetary penalties to unauthorised training inputs.

The legal repricing has already started

A key shift has already occurred: courts and settlements are beginning to place explicit monetary value on misuse of training material. Whether each case succeeds on specifics, the macro effect is the same: the assumption that the intelligence input is free is increasingly difficult to sustain.

The cost of goods sold was never truly zero. It was deferred, externalised, and now increasingly litigated.

Why lawsuits alone do not solve the whole problem

Litigation is retrospective, expensive, and selective. It protects claimants with clear rights chains. It does not structurally compensate the broader population whose aggregate knowledge contributions are diffuse and non-registrable.

The Norway model, applied to intelligence extraction

The sovereign answer is policy architecture, not courtroom roulette:

• Define national intelligence extraction as a taxable base.
• Collect levy into a ring-fenced sovereign fund.
• Distribute returns as a Universal Basic Fee.
• Retain principal for intergenerational compounding.

The harder question no fund can answer

Tax and distribution can solve financial stability. They do not solve existential purpose. If work's role as identity, status, belonging, and contribution decouples from income, societies must intentionally rebuild meaning architectures at scale.

We are obsessing over how people will eat in an AI economy while underinvesting in why they will get up.

The likely growth domains are deeply human: care, mentorship, community leadership, intergenerational support, and lived presence. These are not side activities in an AI economy. They become core civilisational infrastructure.

The madness in one frame

We built tools from human knowledge, monetised the outputs, and externalised the extraction cost. Courts are now repricing that cost. Labour displacement risk is rising. The policy response cannot be hand-waving.

Tax the extraction. Build the fund. Pay the fee. Then confront purpose directly.

Now I want to hear from you. Did this shift how you see AI value flows? Is a sovereign wealth tax plus Universal Basic Fee practical policy, or overreach? And the deeper question: in a machine-thinking world, where will purpose come from for you, your team, and your children?

Tell me I am wrong. Tell me I am right. But do not tell me the intelligence inside these systems was free.

- Carlo