The AI Diary: What Everyone Is Getting Wrong About AI And Jobs

Nov 5, 2025

For years, the AI debate has swung between “robots will take all our jobs” and “this is all hype.” The truth, as usual, sits in the middle — and it’s more practical and more optimistic than either extreme.

A Quick History Check

Back in 2016, Geoffrey Hinton suggested we might not need to train new radiologists because AI would soon read scans better than people. Fast-forward nearly a decade: radiologists are busier than ever. Yes, they use AI tools that read images quickly and accurately. But the tools didn’t erase the job — they changed it.

Two reasons. First, regulation and liability still require a human in the loop. More importantly, when something gets cheaper — like medical imaging — people use more of it. Cheaper MRIs mean more scans, and more scans mean more complex cases and treatment decisions that need human judgment. Economists have a name for this:

Efficiency Usually Creates More Work, Not Less

We’ve seen this before:

• Container shipping got cheap. Some dock jobs disappeared, but global trade boomed, and new logistics industries sprang up.
• Cloud computing cut IT costs. Traditional server work turned into DevOps and cloud architecture, and companies built systems they couldn’t have dreamed of before.
• AI gets cheaper to run. Demand for the tech and the people who can deploy it jumps. Just look at the scramble for GPUs and machine learning talent.

So What Actually Changes?

Jobs don’t vanish wholesale; they shift. Tasks that were once tedious or repetitive move to software. People move up the stack: from doing the task to checking, guiding, and improving the systems that do the task.

Some roles will shrink — especially those that are routine, low-context, and tolerant of mistakes (basic data entry, simple customer support). But even there, the work often morphs into managing AI-driven workflows, handling edge cases, and focusing on tricky, high-stakes interactions.

You can already see it happening:

• In sales, AI agents handle first-pass outreach so teams can focus on big accounts and tricky negotiations.
• In healthcare, paperwork automation frees staff to coordinate care and manage complex cases instead of chasing forms.

The Bottom Line

AI won’t wipe out work. It will change the work — and usually expand the market for it. Expectations will rise. Quality bars will rise. And the people who learn to use these tools will pull ahead.

This isn’t a call to wait around for a fully automated future or pin everything on universal basic income. It’s a nudge to get practical: figure out where AI can take the grunt work off your plate, learn how to supervise and improve it, and build from there. That’s where the real opportunity is.