Years of leading developers, analysts and testers — then the tools changed completely and the job of directing them barely changed at all. Here is what transferred.
Whatever shipped, a person typed it. The craft was in the hands.
You would watch the demo, laugh, and go back to writing it yourself.
They lead the agents. Briefing, directing, reviewing, deciding.
I assumed the new era would ask me for a new skill set. Instead it handed me a bill for the one I already had — the same moves I learned leading 20 people: developers, business analysts, testers, and a project manager.
Different words, different research literatures — but from the lead's chair it is the same failure, with the same root cause and the same fix. When the brief is bad, the failure belongs to whoever wrote the brief.
Covers enormous ground, a little arrogant, leaves a trail to clean up — and clocks out after six hours, no negotiation.
Top of the benchmarks. Quick, sharp, does not always check thoroughly — and those tokens run out very fast.
Never the star, never the bottleneck. Quietly carries a third of the sprint without drama.
The most balanced of the roster. Sufficiently intelligent, web crawling good enough for real work, price that does not sting.
Unmatched at one thing. You do not hire a specialist and then complain that they are a specialist.
The best web crawling I have used — Google origins showing. Other abilities can fall short, and that is not what I bring it in for.
Writes the spec, shapes the design. Where a little arrogance costs nothing.
The long, patient, unglamorous stretch of actually building it.
Last pass — testing, and tightening whatever got left behind.
He was almost fired — not for his work, for his eccentric behaviour. He kept hours nobody had approved of. So I scouted him, took him onto my team, and did the only genuinely clever thing in the story: I left him where he deserved to be. No sanding him into a normal employee. Within a short stretch he became one of the best employees the company had. The same behaviour that nearly cost him his job was, in the right seat, the reason he outperformed everyone.
Very good results from people I trust, at a much cheaper token price — exactly the profile I have learned to chase. Currently not taking new subscribers, so the door is closed for now. The moment it opens, I am finding out what it is actually good at.
Planning too far ahead can bite you — these tools move fast, and a plan written against last month's capabilities can have you routing carefully around a limitation that quietly stopped existing. That risk is real. Going in with no plan at all is still, by a wide margin, worse.
.md
The most common phrase I hear when I ask what went wrong — and the stunning part is that it is not untrue.
The model really did write the broken thing and invent the wrong assumption.
If you take no responsibility for the result, you have described yourself as a pass-through.
Nobody needs to employ a pass-through. The token price is cheaper — and then the mistake is at least mine to own.
Not a nice-to-have on top of the technical skill — the specific, concrete thing a human is still paid for. Someone who checks. Someone who catches it before the client does. You will still mess up with the AI; that was never the failure. The failure is handing the mistake to someone else with the AI's name stapled to it.
The operating parts — the ones that fill your calendar.
Brief clearly Assign to strengths Plan before you start Own the result Keep scouting the rosterThe relationship parts — real differences, honestly stated.
No career to grow No Monday-morning mood Unmoved by praise or blunt feedback No memory of yesterday unless you arrange it Never says "we are building the wrong thing"Every difference is about the relationship. Every similarity is about the work. And it is the work that fills your calendar.
We spent years being told human skills were about to be devalued. Close to the opposite happened. Context, delegation, planning, ownership — all still load-bearing, all still the difference between a good week and a wasted one. Human skill is not wasted. It is more important now than it has ever been.