← All writing

AI & Society

What 30 Years of Building Teams and Technology Taught Me About AI

Every big shift in technology looked like the end of something. What actually ended was usually a habit, not a job.

I have watched the ground move under this industry many times. Each time it felt like the end of something. Each time, what actually ended was smaller and more personal than the headlines said.

My working life started in 1996 at Newark airport, running operations and staffing for an aviation services company. Three years later I co-founded a digital media company in New York called Bla-Bla Networks, back when the web was new enough that getting a page to load felt like an accomplishment. Since then I have built products through almost every wave that followed. Online entertainment at UGO, which Hearst acquired in 2007. Mobile advertising at Receptiv, a company I co-founded that Verve acquired in 2018. Interactive and connected TV at BrightLine, where I was VP of Product and acting CTO. Cloud transformation at LivePerson. Ad tech at NBCUniversal. AI-driven workforce tools at Talisai. And now Kbean, where I am CTO.

I am not writing this to tell you whether AI is overhyped or underhyped. Both are true depending on the week. I am writing it because I have been through enough of these transitions to recognize the pattern, and the pattern is more useful than any prediction.

What changes

The first thing that changes is the cost of the thing everybody used to be proud of.

When the web arrived, the hard part was getting a page to load at all. People built careers on knowing the tricks. Then the tricks became tools, and the tools became defaults, and nobody was proud of loading a page anymore.

AI is doing that to a lot of things at once. Writing a first draft of a function. Summarizing a document. Turning a rough idea into a working prototype in an afternoon. These used to be where the effort went, and the effort was the proof that you were good.

That is the uncomfortable part. When the effort gets cheap, the proof goes with it, and you have to find another way to know you are worth something at work.

What endures

Here is what has not changed once in my whole career.

Somebody still has to know what the problem is. Not the stated problem. The real one, the one underneath the ticket, the one the customer could not put into words. Most of us have sat in a room where everybody agreed on the solution and nobody had checked whether it solved anything.

Somebody still has to decide what good looks like. A model can produce ten versions of something. It cannot tell you which one your users will trust, or which one your team can maintain in two years, or which one quietly breaks a promise you made to someone.

And somebody still has to own it when it goes wrong. That has never been automated and I do not expect it to be.

The habit that has to go

I have a pattern in me that I have never fully gotten rid of. Something is broken, I build a thing. It is a good instinct most of the time. It has also cost me, because building is comfortable and asking the harder question is not.

AI makes building almost free. That means the instinct that served engineers for decades, the urge to start making something, is now the cheapest part of the job. The expensive part is stopping before you build and asking whether you should.

The engineers I see struggling right now are not the ones who lack AI skills. They are the ones whose sense of themselves is tied to the part that just got cheap.

The ones doing well have quietly moved their pride somewhere else. Into judgment. Into understanding people. Into knowing which of the ten versions is the right one, and being able to say why.

What I would tell myself at the start

Learn the tools. Use them every day. Do not be the person standing outside the room insisting the old way was better. The room does not wait for him.

But do not confuse the tool with the work. The work was always about understanding something well enough to change it on purpose. The tools keep getting better at the changing part. The understanding is still yours.

If you lead a team or a company through this, I would love to compare notes. I speak and advise on exactly this transition, and I would rather have the conversation than give the lecture.

AIsoftware engineeringcareertechnology transitions
Jaime Boyaca

Technology executive, speaker and explorer of consciousness. This idea is also the heart of his keynote From Reaction to Response.