Block cut about 4,000 jobs in February, roughly 40% of its workforce. Now the payments company says continuous AI models will coordinate work and eliminate a permanent middle-management layer. Jack Dorsey and Roelof Botha laid out the plan in a blog post titled "From Hierarchy to Intelligence," calling the change experimental and incremental. The rewrite reorganizes staff into three roles, individual contributors, directly responsible individuals, and player-coaches, and says humans will still handle major business and ethical decisions as pilots run 90-day outcome cycles.

Block is recasting itself as an "intelligence" powered by continuously updated AI models and is explicitly targeting middle management as an avoidable layer, Jack Dorsey and Roelof Botha wrote in a company blog post. The pair argued that two so-called world models can replace much of the coordination work managers have historically performed.

Those models do two jobs, according to the essay "From Hierarchy to Intelligence." One aggregates internal operational data. The other maps customer and transaction behavior. Both feed an "intelligence layer" that the company says can monitor work, assign tasks and even compose financial products dynamically.

So what changes for Block employees on the ground? The company plans to reorganize staff into three distinct roles. Individual contributors will be deep specialists who receive context and direction from the models rather than from managers. Directly responsible individuals will own cross-cutting problems and have the authority to draw on system resources to solve them.

Player-coaches will combine hands-on work with mentoring and talent development, taking on people-related duties that middle managers used to handle.

Dorsey and Botha were blunt in the post. "There is no need for a permanent middle management layer," they wrote, arguing that the system-generated coordination will remove many of the routing, prioritization and monitoring tasks that create hierarchical layers. They also warned the shift will be difficult and uneven, saying parts of the redesign "will likely break before they work," and describing the effort as "in the early stages."

Layoffs, timing and the AI rationale

The post and company messaging link the new operating model to the company's February workforce reduction. Block cut about 4,000 employees in that round, a reduction company leaders have described as approximately 40% of the workforce. Dorsey has said advances in AI were a primary factor in that decision.

Mastheads

Company spokespeople and the essay stress that the changes aren't purely cost cutting. They frame the moves as organizational restructuring enabled by faster AI capabilities.

But reporting that quotes current and former employees pushes back on the timeline for full automation of managerial tasks. Those employees say most AI-generated code changes still require human modification and that AI can't yet lead in regulated functions such as banking and money transfers.

That tension shows in internal reactions. One set of internal Slack messages shows employees confused about whether their managers are truly becoming "player-coaches" or simply "coaches," suggesting ambiguity in how the new roles will be implemented at team level. The company says implementation is incremental rather than a plug-and-play replacement of managers overnight.

Capabilities claimed and the evidence cited

Dorsey told interviewers he saw a capability shift last December in models from other vendors and pointed to specific model versions as evidence that tools can now operate in large codebases. He used that observation to justify the timing of Block's organizational changes. The essay also proposes that when the intelligence layer lacks a capability, that gap effectively defines what the system should build next, replacing fixed roadmaps with a system-generated backlog.

Not all the changes are theoretical. The blog post and company statements say the model will monitor operational data and customer behavior continuously, assign tasks and compose financial products dynamically when conditions call for them. Still, current and former employees quoted in reporting say AI still falls short in practical ways. They note that AI-assisted code often needs human modification and that regulated financial work brings compliance and risk challenges that are hard to hand off to a model.

There are early signs of messy transition. One account indicates some staff who were let go in February were quietly rehired in March. That suggests teams are still calibrating headcount and capabilities as the company experiments with the new operating model.

Block says the redesign doesn't remove humans from major decisions. Company spokespeople emphasize that people will remain central for business and ethical choices even as the intelligence layer handles routine coordination and task assignment.

The company is also experimenting with operating rhythms for the new roles. It has started testing 90-day outcome cycles for directly responsible individuals, a cadence the essay highlights as a way to focus ownership and measurable results without reestablishing permanent middle management.

Related Articles

Block says it will scale the experiment gradually and is already piloting 90-day outcome cycles for directly responsible individuals as part of the transition.

This article was created with AI assistance.