The Cost

I work at a large company that's bullish on AI. For a couple of years, I wasn't. I tried integrating AI into my workflow — pair programming, code review, drafting — and the results were mixed at best. It felt clunky. I fought it more than I used it. And I wasn't alone — studies were showing that even senior-level developers were being slowed down by AI integration. That was the prevailing sentiment in my community and in the news for quite a while. The data backed up what I was experiencing.

Then I started paying attention to what other people were actually doing. Not the LinkedIn thought-leaders regurgitating press releases — the engineers and builders I respected. Teammates. Friends. People whose judgment I trusted. They weren't just enthusiastic. They were transformed. Their output had changed. Their speed had changed. The nature of their work had changed.

So I dug in. Studied the latest models. Learned how these people were actually using the tools. And I found the gap.

There is a real chasm between what most people do with AI and what people who understand AI are doing with it. The results are not remotely comparable.

Most people use AI like a slightly smarter Google. Ask a question, get an answer, move on. Or they use it as a pair programmer — writing code alongside them, catching errors, suggesting completions. That's useful. It's also leaving 90% of the capability on the table.

Using AI effectively doesn't mean pair programming. It doesn't mean delegating basic menial tasks you think it won't mess up. I spent months fighting AI trying to help me code, because fundamentally, our minds — if you want to call it that — our styles, our training, they're very different. I was trying to keep my process and have AI improve it.

Thought patterns. Processes. Habits. All of it needed to change. I needed a tabula rasa mindset — look at the work with fresh eyes and ask: "If I were starting from scratch, with these tools, how would I approach this?"

Once I started doing that, everything shifted. That's where the Beyond the Screen concepts came from. The realization that I don't need to write code. That I don't need monitors. That typing is actually the bottleneck. That I can direct AI agents with my voice while walking around the house.

And then I noticed: the people on the cutting edge are all arriving at the same conclusions. None of the best engineers I know are still writing code. They're directing systems that write code. The skill set that matters has fundamentally shifted, and it's happening right now — not in five years, not gradually. Now.

To keep doing things the way we were is a sure-fire way to get left behind. That's not pessimism — it's pattern recognition. This is classic Darwinism happening in real time, in what is easily the biggest revolution in human history.

That sounds hyperbolic. It isn't. The industrial revolution mechanized physical labor over the course of a century. The internet connected information over the course of two decades. AI is transforming cognitive work itself — the thing that was supposed to be uniquely, irreducibly human — in years. Not decades. Years.

The .com boom and the industrial revolution will look like sidenotes. Not by a small margin.

I see this clearly now. And I'm trying my best to ride the crest of that wave so as not to get drowned by it. That is why I use AI every day. That is why I'm building and writing and thinking so much about it. Ultimately, I have a family I need to help provide for, and it would be irresponsible for me not to do everything in my power to be successful on that front.

AI is here. It's going to saturate everything. There is no putting the genie back in the bottle.

And of course we're going to have problems. Societal breakdowns. Reorganizations. Job displacement at a scale we've never seen. Power concentrating in the hands of whoever controls the models. Privacy eroding in ways that make social media look quaint. Real costs, borne by real people.

But consider the oldest version of this story we have. Prometheus steals fire from the gods and gives it to humanity. Zeus punishes him — chains him to a rock for eternity. The message: powerful tools come with consequences. Always have.

Fire can burn and destroy. Or it can warm and protect. That is always the delicate balance we have the responsibility, as wielders of these tools, to get right. AI is a tool. It is fundamentally no different than any other tool or concept before it — just faster, broader, and more consequential.

The conversation around AI tends to focus on what we're about to lose. Jobs. Skills. Privacy. Agency. These are legitimate concerns and they deserve serious attention.

But there's a question that gets asked far less often: what have we already given up? Long before AI, long before the internet, long before electricity — how many trades did humanity make, knowingly or not, that reshaped what it means to be human?

Yuval Noah Harari makes a provocative argument in Sapiens: we didn't domesticate wheat. Wheat domesticated us. Hunter-gatherers had diverse diets, flexible schedules, and varied physical demands. The agricultural revolution chained us to fields, narrowed our diets, destroyed our joints, and created social hierarchies that didn't exist before. But it went further than that. Stored grain created surplus — and surplus created something worth stealing. Permanent settlements created territory worth defending. Dense populations living alongside livestock created the perfect breeding ground for zoonotic diseases: smallpox, influenza, measles, plague. Hunter-gatherers didn't have organized warfare. They didn't have pandemics. We traded those into existence — for grain. It was sold as progress, and in many ways it was. But the individual human arguably got a worse deal. The species thrived. The person suffered.

Graham Hancock — a polarizing figure whose claims about ancient civilizations draw as much skepticism as fascination — nonetheless lands on something that resonates when he calls us a "species with amnesia." Whatever you think of his broader work, there's a fundamental truth in that refrain that stands on its own. We don't remember the sacrifices our ancestors made for convenience, safety, or survival. Even if we were all great students of history, recorded history represents a blip in the timeline of humanity. We can't know all the paths we've walked or the tradeoffs we've made.

Every one of these transitions was irreversible. Every one created winners and losers. Every one had advocates who saw only the upside and critics who saw only the cost. And every one reshaped humanity in ways that were impossible to fully predict at the time.

AI is the next entry in this table. We will gain things and lose things, and we won't fully understand the tradeoffs for generations.

Here's where my perspective diverges from the standard tech-ethics essay. Because AI isn't only taking things away. In some cases, it's giving things back.

The agricultural revolution gave us food security but took our physical freedom. The industrial revolution gave us production but took our craft. The digital revolution gave us information but took our attention.

Is it possible that the AI revolution gives us back some of what the previous ones took? Physical movement. Creative agency. Time. Craft.

Nobody is going to stop using AI. That debate is already over. So if there's a chance these tools give back some of what previous revolutions took — we should be paying attention, and making sure we actually capture it.

I don't have a policy agenda. I'm not a researcher or a politician or a pundit on the sidelines. I'm an engineer and a trainer with skin in the game — in the arena, getting my hands dirty with these tools every day, trying to build a good life for my family while riding the biggest technological wave in human history.

Roosevelt had it right: "It is not the critic who counts; not the man who points out how the strong man stumbles. The credit belongs to the man who is actually in the arena, whose face is marred by dust and sweat and blood." The people writing think-pieces about AI from a distance aren't wrong, but they're not the ones navigating it daily. The costs I'm describing aren't theoretical to me. Neither are the gains.

What I have is a conviction: the people who use these tools have a responsibility to be honest about them. Not breathlessly optimistic. Not reflexively fearful. Honest.

AI is fire. It will burn things down and it will light the way forward. The question — the only question that's ever mattered with powerful tools — is whether we're paying enough attention to wield it well.

I'm paying attention. I hope you are too.