In 2030, Why I Would Regret Not Asking Better Questions

I wish I'd asked the right question.

It's 2030, and I'm unemployed. Possibly unemployable.

This is not how I expected life to turn out.

Five years ago, I was a partner at a respected law firm. I was billing hours, growing clients, making money. I believed I was doing everything right.

What frustrates me now is that my situation today was predictable then.

I just didn't ask the right question.

How an early AI bet made one law firm unbeatable, briefly

So, how did I get here? Well, in 2023, when ChatGPT first appeared, I saw the opportunity immediately. This wasn't incremental technology. It felt transformative. A tool that could increase productivity tenfold. Maybe more.

So, I pushed hard. My company adapted quickly. Faster than most.

The transformation wasn't always smooth, though.

In the early days, I made some expensive mistakes. I perhaps trusted the technology a little too much, a little too early.

Some of the large language models I used leaked confidential client information.

I also drafted legal arguments built on what the models had invented, complete with citations that looked real but were not.

But issues like this faded quickly. The models got better. More accurate. They started to understand context and nuance at levels junior lawyers couldn't ever compete with.

To be honest, they quickly surpassed my abilities too.

And that allowed us to scale. The benefits far outweighed the risks. So, we took on more clients, expanded our services, and increased output, without having to increase headcount. And when staff retired or moved on, we decided we didn't need to replace them.

For a while, it felt like we had cracked the code.

Margins doubled. Profits soared. We were making millions. I made millions.

But during that period of adoption and growth, I was busy. Perhaps busier than I'd ever been. I burned the candle at both ends.

And that meant I was too busy to think too far forward.

The inevitable slide: margin compression, then no clients at all

Looking back, I realize that what then happened was inevitable.

Slowly at first, margins began to compress. You see, competitors had adopted the same tools. So, everyone could do more, for less. To stay competitive, we began letting go of some of our best people.

By the end of 2028, we were down to just a handful of staff.

And in March 2030, we closed our doors for the last time.

Not because of margin compression, though, but because clients no longer needed law firms at all.

You see, they too had started to trust the algorithms.

Our expertise, years of training and decades of experience, had been embedded in software anyone could access.

But it wasn't democratization. It was consolidation; the collective knowledge of thousands of firms absorbed into algorithms owned by a handful of global tech companies.

AI was never just another productivity tool

Only in hindsight did our real mistake become obvious.

I thought adopting AI early made me progressive. In reality, I was accelerating my own irrelevance.

Because AI was never just another productivity tool.

Comparing it to the wheel, the computer, or the internet was a category error. Those were dumb technologies. They amplified human effort. But they still needed people to create value.

AI, though, was different.

Yes, it increased production. Just not human productivity. It shifted value creation from people to silicon chips. Chips that didn't sleep. Didn't complain. Didn't require holidays. Didn't need careers. Across knowledge industries, we stopped paying people.

We started paying the machines.

And the handful of companies that owned the algorithms benefited handsomely as more and more of society became dependent on their technology.

What Should We Have Done Differently About AI?

Looking back, I regret not taking the time to think forward. To ask better questions.

It was obvious that artificial intelligence was going to be transformational. Disruptive. And that this would happen at unprecedented speed and scale.

I should have lobbied policymakers to treat AI more like a utility, governed in the public interest, much like water, power, roads, and other foundational infrastructure.

Instead, I watched as a handful of technologists were allowed to decide who would get access. At what price. For what purpose.

And all without accountability for the inevitable unintended consequences.

Who Will Pay for the People AI Fails?

That's the question I wish I'd asked. I wish we'd all asked.

Who will pick up the tab for those people that AI fails? The displaced. The devalued. The discarded.

Our custodians might not have listened. But that doesn't mean we shouldn't have tried.

Author's note: This is a fictional future, written to explore a real question: who will pay when artificial intelligence reshapes work faster than society can adapt?