THE THREE SKILLS OF THE AGENTIC ERA

Never liked geometry? Neither did geometry, particularly. But there's a new shape to get your head around. And it looks like a trident.

Those of us who have marched off and started to put AI agents into our everyday workflows are increasingly reporting that we seem to be developing three key skills:

1. Curating the right work.

When you can achieve so much, you have to ask: 'What's worth doing?' And the corollary: 'Which thousand ideas need to be given a firm no?'

2. Running the chain gang.

Orchestrating multiple agents isn't entirely straightforward. Not least because of the constant task switching. But you also have to translate what you want into agent-friendly work packages - and then decide how to evaluate what they do.

3. Grounding the output.

Sure. An army of agents is super-productive. But humans still need to provide the domain expertise that models lack - the practical constraints, edge cases and reasons why things, in the real world, are the way they are. Usually meaning messy, complicated and likely to overheat.

And so the mental geometry you find yourself bent into is indeed three-pronged: An Enterprise Director. An Agent Manager. And a Subject Matter Expert.

Once upon a time it was enough to be T-shaped - holding on to one deep subject matter expertise and some convincingly general competence. That was sufficient when collaboration meant handoffs between specialists.

But it does seem the agentic era will demand 50% more geometry.

Of course, the whole experience is only a few months old. And frontier models will keep making agentic orchestration easier. They'll also better approximate domain expertise. So perhaps the first prong of the trident, that of enterprise director, will be the longest and sharpest?

Maybe. But it's also quite possible we'll simply juggle more agents and push even harder on our own expertise when the models improve.

As it stands today, all of us have work to do to develop our nascent agentic orchestration skills. Which is why you'll find us swapping notes after a (cautious) session with OpenClaw.

Let us know if you'd like to be counted in on that too. 🦞