The Cheaper AI Gets, The More We'll Need You

In 1865, a British economist named William Stanley Jevons noticed something strange. As steam engines became more efficient, Britain didn't use less coal. It used dramatically more. Better technology lowered the cost of energy, which created so many new uses for it that total consumption exploded. It became known as the Jevons Paradox: efficiency doesn't reduce demand. It multiplies it.

I've been thinking about this a lot lately, because I think it's the most important idea nobody is talking about when it comes to AI and our generation's job predictions.

Everyone I know at Bocconi has, at some point in the last year, had the conversation. You know the one. What's even the point of becoming an analyst if AI can build the model faster than you? I've had it too, usually late at night before an exam, spiraling through LinkedIn posts about automation and trying to convince myself it won't be that bad.

But here's the thing. According to Epoch AI, the cost of running a large language model at a given performance level has been dropping at a median rate of 50x per year. And yet OpenAI's annualized revenue went from $2 billion in 2023 to more than $20 billion in 2025, while its computing capacity tripled in a single year. AI is getting cheaper and faster, and instead of shrinking the market, it's exploding it. The Jevons Paradox, playing out in real time.

Think about what this means. When analysis becomes cheaper, you don't do less analysis. You do more. Companies that could only afford one consultant's report now commission 10. Investment funds that modeled twenty scenarios now model two hundred. The demand for insight doesn't go away, it scales with the supply of tools to generate it. And someone still must commission it, interpret it, challenge it, and make the actual call. That someone is us.

The World Economic Forum's Future of Jobs 2025 report notes that roles like fintech engineers are among the fastest-growing jobs because even in a sector under pressure, the work that sits at the intersection of finance and technology is expanding, not contracting. The floor is moving, yes. But it's moving upward for the people who position themselves correctly.

I'm not saying this to be naively optimistic. Moody's chief economist Mark Zandi recently described companies as reaching a "Cortés moment" on AI, a point of irreversible commitment, like the conquistador who burned his boats on arrival in Mexico, leaving his troops no choice but to advance. There's no going back. The transition is real, and it will be uncomfortable for a lot of people who aren't paying attention.

But I think the Bocconi panic about AI is slightly misdirected. The threat isn't that AI will replace us. The threat is that AI will raise the baseline, and the people who don't adapt will find themselves doing the same tasks as before while everyone around them is operating at a completely different level.

The Jevons Paradox is optimistic, but it comes with a condition. Coal consumption exploded because people found new things to do with cheap energy. The same logic applies here. AI creates more demand for human judgment, creativity, and decision-making, but only if we're developing those things, not just coasting on the skills that are getting automated.

So yes, I think we'll be okay. More than okay. But not automatically. Not just by showing up with a Bocconi degree and decent Excel skills.

We have to be the people who know what to do with the cheap energy.

Article by Daria Pricop - BSMS Partnerships