Something Is Happening

A viral piece this week compared the current moment in AI to February 2020 — the "this seems overblown" phase before everything changed in three weeks. It's been shared widely, often with the instruction to send it to everyone you care about.

Here's where we stand on all of this.

AI capabilities are improving faster than most professionals realise. That's true. If you last tried ChatGPT in 2023 and wrote it off, your mental model is out of date. The current generation of models can do things that would have been unthinkable eighteen months ago, and the pace isn't slowing down. Anyone telling you this is a fad is wrong.

But there is an enormous difference between a technology being capable and a technology reshaping the economy, and nearly all of the viral AI discourse right now collapses that difference to zero.

An AI model that can draft a contract is not the same thing as law firms laying off associates. An AI that can build a working app from a description is not the same thing as software engineers being unemployed. Between capability and consequence sit years of procurement cycles, regulatory adaptation, liability frameworks, integration costs, organisational inertia, and the simple fact that changing how millions of people work is slow, messy, and uneven — even when the underlying technology is ready.

This isn't cope. It's how every major technology has ever diffused through an economy. The dynamo was demonstrably superior to the steam engine by the 1880s. It took until the 1920s for factories to be fully redesigned around electric power. Not because factory owners were stupid, but because rewiring an economy is a structural problem, not a perception problem. Knowing the technology works doesn't make the transition instant.

The COVID analogy gets this exactly backwards. A virus spreading through a population faces almost no institutional friction. Technology adoption faces almost nothing but. One reshapes the world in weeks. The other takes years, sometimes decades — and the pace is determined less by the technology's capability than by the readiness of every system around it.

So no, we don't think you need to treat this like a fire alarm. We do think you need to treat it like a serious shift that rewards understanding over panic.

The most aggressive timelines circulating right now — 50% of white-collar jobs displaced in one to five years — come almost exclusively from people who run AI companies. That doesn't make them wrong. It does mean you should notice the incentive before you absorb the timeline. The people most certain about the speed of disruption are the people whose valuations depend on everyone else believing it's imminent. Again: they might be right. But certainty from interested parties is not the same thing as evidence.

What we'd actually recommend is less exciting than what the viral posts suggest, and more useful.

Engage with the current tools seriously. Not as a novelty — as part of your real work. See where they help and where they don't. Build firsthand experience, because secondhand urgency is a poor basis for career decisions.

Learn the fundamentals, not just the tools. The specific models that exist today will be obsolete within a year. The ability to understand what AI is, what it's good at, where it fails, and how to evaluate new capabilities as they arrive — that compounds. Chasing the tool of the week doesn't.

Think about your specific role, not "the economy." The impact of AI varies enormously by industry, function, and task. "AI will change everything" is too vague to act on. "AI is already changing how my team does financial reporting, and here's what I should understand about that" — that's actionable.

Give yourself permission to be thoughtful about this rather than reactive. The people who navigate this well won't be the ones who panicked earliest. They'll be the ones who understood most clearly.

Something is happening. It deserves your attention. It does not require your fear.