AI in 2026: Everyone's Talking Bollocks Except the People Actually Building It

Right. Let's have a proper look at this AI madness, shall we?

March 2026 and everywhere I turn it's the same story. China's got a five-year plan. NVIDIA's got a conference. The UN's got a report. The NHS is "implementing solutions." Investopedia's got "ongoing concerns."

Know what I mean? It's all performance.

China's doing what China does — state-driven, top-down, everybody march in the same direction. Fair play to them for committing, I suppose. At least they're not pretending. They're saying "we're doing this" and then they do it. Whether it's good for anyone is a different question, but there's no bullshit about whether they're serious. That's something.

NVIDIA though. GTC 2026. Jensen Huang up there with the "global technology leaders" showcasing the "Age of AI." Are you having a laugh? It's a dog and pony show. Massive screens. Flashy demos. Everyone nodding like they understand what's actually happening. Most of them don't. They're there to be seen, not to learn. It's corporate theatre and we're all paying for the tickets.

The UN report on working conditions — now that's interesting. Actually proper interesting. Because while all these CEOs are wanking on about productivity and efficiency, real people are losing their jobs. Or their jobs are changing so much they don't recognise them anymore. That's not a "concern." That's someone's livelihood. That's someone trying to feed their kids. And some suit in Geneva is writing a report about it like it's a fucking academic exercise.

NHS putting AI in healthcare. Look. I get it. Anything that helps doctors do their jobs better, anything that gets people treated faster — that's biblical. That's actually good. But I've seen how these things roll out. It's never as simple as "implement AI, sorted." It's months of meetings. Procurement bollocks. Training that nobody does properly. Then half the staff hate it because it makes their jobs harder. I want it to work. I really do. But I'm not holding my breath.

And Investopedia — "AI in Business: Applications, Future Trends, and Ongoing Concerns." Ongoing concerns? What does that even mean? It means they haven't got a clue. It means they're hedging. Covering all bases so nobody can say they were wrong. That's not insight. That's cowardice.

Here's what I think. Here's what's actually real:

Most of this AI stuff is noise. Corporations pretending they've got a strategy. Governments pretending they've got a plan. Consultants pretending they've got answers. Everyone performing progress while nothing actually changes for the people on the ground.

But — and this is important — there are people building proper things. Quietly. Without the press releases. Without the conferences. They're in garages and small offices and university labs and they're actually solving problems. Not "disrupting industries." Not "maximising shareholder value." Just making something that works. That helps. That matters.

Those are the ones I respect. The ones who don't need to announce what they're doing every five minutes. The ones who just do it.

The rest? They can do one.

Five-year plans. Global summits. Ongoing concerns. It's all bollocks. You want to know if AI is actually good? Don't read the reports. Don't watch the keynotes. Look at what people are actually using. What's actually making their lives better. What's actually working.

Everything else is just noise.

Simple as.

OX x