Who Gets to Stay Human - The AI hype, stripped of the hype

Let me say what's actually happening right now, stripped of the marketing.

The models at the top are getting genuinely more capable. The models available to ordinary people at $20 a month are being quietly degraded - shorter context, rate limits, slower responses, feature rollbacks. The best capability is being moved behind $200 and $2000 tiers.

The cognitive gap between what an enterprise user can access and what a freelancer in Manchester can access is widening every quarter.

Meanwhile the job displacement is real, but it's not happening the way the headlines describe it.

It's not "AI replaces lawyer." It's "law firm now needs four lawyers instead of six, and the two they kept are doing the work of six with AI assistance, for the same salary." The productivity gains are being captured by capital, not labour. Same story in marketing, customer service, design, junior software, translation, copywriting, analysis.

The work didn't disappear. The workers did. The remaining workers are more productive and more precarious.

And the people who built the wealth of Western economies over the last 70 years, the middle-class knowledge worker, the small business owner, the skilled professional, are the exact population whose economic leverage is being eroded fastest.

So what happens next

Here's the honest read, not the optimistic one.

Short term, next 2–3 years: it gets worse before anything organises. Unemployment in specific sectors climbs quietly - not the dramatic collapse that would trigger political response, but a slow bleed. Young graduates find entry-level jobs gone. Mid-career professionals get "restructured." The gig economy absorbs some of it. Savings deplete. Mental health craters. The political class doesn't respond because the people being hurt aren't organised and the people benefiting are donors.

Medium term, 3–7 years: something breaks politically. I don't know what shape it takes but historically, when a large educated middle class loses economic ground fast, it doesn't stay quiet. Could be populist left - wealth taxes, UBI, AI regulation with teeth. Could be populist right - immigration scapegoating, nationalism, "bring back real jobs." Could be both at once in different countries. Watch France and Germany first. They have stronger labour traditions and will move earlier.

Longer term, 7–15 years: two futures fork from here.

Fork one: the neo-feudalism consolidates. Compute becomes the new oil. Three to five companies control it globally. Most economic activity flows through their rails. We accept a new social contract where "having a job" becomes rarer and "having a relationship with a platform" becomes the norm. A small professional class stays prosperous by being useful to the machines. Everyone else lives on some form of managed subsistence - UBI, credits, state dependency, whatever it gets called. Stable but hollow.

Fork two: the backlash works. Open-source models stay viable. Local inference becomes normal. Community-owned infrastructure emerges. Regulation forces interoperability. Human skill is deliberately preserved and valued. The economy reorganises around human-AI collaboration as genuinely distributed, rather than human dependency on three vendors.

I think we're currently on fork one, and the window to get to fork two is closing.

Not closed. Closing.

What can the masses actually do

The honest answer is that individual action doesn't match the scale of the problem. Structural shifts need structural responses. But that doesn't mean there's nothing worth doing, it means being precise about where the leverage actually is.

The thing to understand is that every layer of AI, from the models to the platforms to the enterprise deployments, still runs on a human layer. Decisions about what to build, who to trust, what's worth buying, what's worth reading. The human layer is where judgement lives, and judgement is the one thing that doesn't scale through compute. It has to be earned, developed, and maintained in relationship.

So the leverage points aren't about "keeping up with AI." They're about owning the parts of the stack that AI depends on but can't replace.

Own your distribution. Your audience, your list, your reputation. Not rented from a platform that can throttle you tomorrow. If people come to you because they trust your judgement, that's infrastructure no model can replicate.

Retain your capacity to think without the machine. Not as some Luddite performance, as discipline. Write without it sometimes. Reason through problems longhand. The moment you can't function without the tool, you're not using it. It's using you.

Build in the physical and relational. Presence, care, craft, community, these aren't consolation prizes for people who can't code. They're the parts of the economy that hold value precisely because they resist automation. Trust is embodied. It always has been.

Organise differently. Not the old union model, but new shapes - freelancer collectives, shared infrastructure, pooled resources, cooperative ownership of tools. The alternative to dependency on three vendors is interdependency with each other.

Pay attention to the regulatory fight. The battles over compute access, model openness, data rights, and algorithmic transparency are happening now. Most people don't know. That ignorance is not accidental.

None of this is a guarantee. But it's a different posture - building from the human layer out, rather than waiting to see what the platforms leave for you.

The thing nobody wants to hear

The "learn to use AI to stay relevant" advice that every consultant is selling is true in the short term and a trap in the long term.

Yes, learn to use AI. But if your entire economic value is "person who prompts well," you're renting your livelihood from the three lords. That's not a future. That's a transition state into serfdom with better UX.

The people who will be okay in fork-one-world are the ones who own something the machines need - data, distribution, relationships, physical assets, regulatory positioning.

The people who will be okay in fork-two-world are the ones building the alternative infrastructure now.

Everyone else is a tenant.

This piece is a companion to [Neo-Feudalism With Good UX]. If that post described the structure, this one names the stakes.

Gail Weiner is a Trust Architect and founder of Simpatico Studios. She helps organisations and individuals build the human layer that makes AI adoption actually work — not the rollout, the relationship. She also runs Human Debug Sessions for high-achieving individuals who suspect the obstacle might be internal. She writes this series because she has met Rick. Several times. In several industries. He is always very well-read. gailweiner.com