The Technical Context
I was hooked not by the fact of layoffs at Block, but by what exactly Dorsey is trying to build in place of traditional management. Based on materials from late March 2026, a three-layer scheme is emerging: deep specialists as ICs, hybrid managers as player-coaches, and task owners as DRIs. Between them isn't just Slack, Jira, and calls, but a central AI system designed to maintain an up-to-date model of the business.
So, the idea isn't just to “give employees Copilot and see what happens.” Here, they are trying to make AI the infrastructure for coordination. An IC gets context not through a chain of command, but directly from the model. A DRI doesn't chase departments for status updates but relies on the overall picture. A player-coach isn't buried in reporting but works hands-on and pulls the team along.
Translated into the language of AI solution architecture, Block isn't just building an internal chatbot. They want a layer that aggregates updates, decisions, dependencies, and task statuses, turning it all into a working context for people. In essence, it's the company's operational bus with an AI layer on top of the data.
And here’s a crucial point: this story is fresh, not a retrospective. But it already comes with a disclaimer from its authors that a lot will break before it works. In my opinion, this is an honest part of the story, because such things look great in a memo but are very painful to unfold in real processes.
Impact on Business and Automation
The strongest signal here isn't about Block, but about a new type of organizational design. I've long been telling clients: implementing artificial intelligence is almost never just about the model. Usually, everything breaks down on how context flows, who makes decisions, and who is responsible for the last mile of execution.
Dorsey's model cuts right through these bottlenecks. Fewer people who just relay information. More people who either do the work or own the outcome. If the central AI system can truly synchronize knowledge without significant degradation, the company gets a shorter path from signal to action.
The winners are teams where expertise is already close to execution—strong developers, data people, product leads accustomed to making decisions with incomplete data sets. The losers are structures where the value of middle management was mainly in forwarding statuses and manually syncing between departments.
But there's a downside. If the AI context is flawed, outdated, or politically “enhanced,” the whole system quickly starts to replicate garbage. Then, instead of acceleration, we get a very expensive, company-wide hallucination. That's why AI automation without proper data discipline and clear lines of responsibility almost always turns into a showpiece.
I see this in projects regularly. When we at Nahornyi AI Lab create AI solutions for business, the most difficult question usually isn't “which model to use,” but “where does the model get its truth and who confirms the action.” If there's no answer, AI doesn't replace coordination; it just creates new layers of chaos.
That's why the Block case interests me not as news about another layoff. It's a field experiment: can you really replace a significant chunk of the management layer with an AI system and a more rigid role structure? If they succeed, many companies will start copying the form. But it will only work for those who can handle the AI architecture, data integration, and decision-making discipline.
I'm Vadym Nahornyi, Nahornyi AI Lab. I don't just retell these things from someone else's slides; I break them down through the prism of real-world AI implementation in processes, teams, and products. If you want to see how such a model could fit your business without the trendy fluff, contact me, and we'll analyze your case together.