When Jack Dorsey announced that Block was laying off nearly half its workforce — from over 10,000 people to just under 6,000 — he said the company was no longer focused on accelerating the work of developers. They were building a different kind of company, one with intelligence at the core of everything they do.

Later that summer, Mark Zuckerberg published a 6,000-word essay called "The Future is for Everyone," in which he used the word "superintelligence" 60 times and promised that everyone would have "invention superpowers." He barely mentioned jobs.

Between Dorsey and Zuckerberg, Palantir CEO Alex Karp went on CNBC and ranted about how enterprises are paying for "tokens that create no value" and that something has gone completely wrong with how AI is sold.

Here is the thing about all three of these leaders. They are all correct — and none of them are people you would naturally trust.

Dorsey is right that the structure of a technology company is changing. He is also the person who presided over years of chaos at Twitter and is now cutting 4,000 people without explaining what the remaining 6,000 are supposed to do differently. Zuckerberg is right that AI will give individuals capabilities they did not have before. He is also the person who spent the last decade selling attention to advertisers and now wants you to believe his vision of the future is benevolent. Karp is right that most enterprises are wasting money on AI that produces nothing — 80% of what he said in that CNBC rant was accurate. He is also the CEO of a company that sells AI to the government and the military, and is not necessarily someone you can trust when he tells you the problem is that everyone else is doing it wrong.

The world we are in

This is the current state of AI leadership: a series of people you would feel wary of trusting, all having valid opinions and concerns, and none of them providing a model you can actually look to. Dorsey fired 4,000 people and did not explain the new job. Zuckerberg wrote 6,000 words about superintelligence without describing a single role that will exist in this future. Karp is right about the waste but has his own angle — he is selling the alternative.

The issue is that people do not understand who to listen to anymore. It is a crisis of confidence, a crisis of trust, and it is happening at exactly the moment when the stakes are highest. There are tech leaders who are simultaneously right about the trajectory and wrong about the details, accurate about the destination and vague about the route, and the people who are supposed to make decisions based on their guidance are left trying to figure out which version of the future is real.

What none of them will say

All three are describing the same transition. None of them are giving you the model for what the work looks like and who is responsible for it. That model is the orchestrator — the person who directs delegated intelligence to create systems, who owns the outcome, who can open the hood when something breaks.

None of them will say that, because saying it would require admitting that the job they are describing does not have a name yet, and naming it would mean the rest of us could hold them accountable for whether they are actually building it.

It is easier to talk about superintelligence than it is to describe what the person who remains is supposed to do on Monday morning.

---

This is part of a larger argument about the transition from programmer to orchestrator. The Orchestrator, forthcoming.