Jack Dorsey cut nearly half of Block’s workforce in February, then co-wrote an essay with Sequoia’s Roelof Botha arguing that the org chart itself is obsolete. Brian Halligan, a HubSpot co-founder and investor, amplified the thesis on social media, mixing Dorsey’s framework with his own observations about the CEOs he works with daily. Redpoint Ventures published a companion slide deck spelling out what makes an AI-native company structurally different from a traditional software firm. Together, the three pieces mark a coherent, and VC-backed, blueprint for how companies should be built in 2026.
Block’s Restructuring as a Proof Point
On February 26, 2026, Dorsey announced that Block would reduce its headcount from over 10,000 to just under 6,000 — a 40% cut made from a position of financial strength. The company posted gross profit of $10.36 billion in fiscal year 2025, up 17% year over year, and raised its 2026 gross profit guidance to $12.20 billion. Block’s stock surged roughly 22% on the announcement, with Morgan Stanley upgrading the stock to overweight and Goldman Sachs raising its price target, citing AI-driven efficiency gains.
Dorsey’s internal AI coding agent, Goose, was already saving engineers eight to ten hours a week, he told investors. 90% of Block’s recent code submissions were AI-authored by the time the cuts were announced. “Intelligence tools have changed what it means to build and run a company,” Dorsey wrote. “A significantly smaller team, using the tools we’re building, can do more and do it better.”
Skeptics were quick to note that Block had grown from roughly 4,000 employees in 2019 to over 10,000 during the pandemic. Goldman Sachs pointed out the reduction effectively returned Block’s headcount to 2020 levels. Bloomberg reported that the cuts aroused suspicions of “AI-washing”; dressing up old-fashioned cost reduction as technological futurism. Om Malik, writing on his personal site, called it a “narrative substitution” template: cut half the company, blame the machines, watch the stock pop. Even Sam Altman had publicly noted that many AI layoff stories were getting ahead of reality.
Sequoia’s Organizational Argument
Three weeks after the Block announcement, Dorsey and Botha co-published “From Hierarchy to Intelligence” on Sequoia’s website. The essay frames corporate hierarchy not as a management philosophy but as a two-thousand-year-old information routing protocol, invented by the Roman Army, formalized by Prussian military reformers, and imported into American business via the railroads in the 1840s. The constraint it was solving, how many people one leader can effectively supervise, has never changed. What changed, Dorsey and Botha argue, is that AI can now perform the coordination work that middle management exists to do.
At Block, the essay describes a shift toward three roles: individual contributors who own deep technical layers, Directly Responsible Individuals (DRIs) who own specific outcomes for defined periods, and player-coaches who combine building with people development. There is no permanent middle management layer. A continuously updated “company world model” built from remote-first, machine-readable work artifacts replaces the information that used to travel up and down the chain of command. Sequoia frames speed as the key metric: “Speed is the best predictor of start-up success,” the essay opens, arguing that AI-enabled organizational redesign compounds speed as a competitive advantage.
Halligan’s “Dorsey Mode” Framework
Brian Halligan, who co-founded HubSpot and now works closely with hypergrowth AI-native CEOs, posted on X that he had interviewed Dorsey for his Long Strange Trip podcast and synthesized the ideas into what he called a first draft framework. Sequoia reposted the thread. Halligan’s contribution was coining a third management archetype ;”Dorsey mode”, sitting beyond the familiar Manager mode (pyramid org, annual planning, hire for experience) and Founder mode (flat structure, quarterly cycles, hire for slope), a distinction Paul Graham had popularized in a 2024 essay.
In Halligan’s taxonomy, Dorsey mode is characterized by a circular org shape, AI as the primary decision-maker (“AI+”), planning cycles that are continuous rather than quarterly, and hiring for taste rather than experience or trajectory. The analogy he chooses is Suno, the AI music generation platform, rather than an orchestra or a jazz band. The organization is described as “legible” rather than organized or chaotic, speed as “warp” rather than medium or fast, and the moat as deliberately “shallow”; replaceable and adaptable rather than wide or deep. The framing is deliberately provocative: in Dorsey mode, focus is explicitly “No.”
Redpoint’s Structural Economics
Redpoint Ventures published a slide deck arguing that building an AI-native company is “fundamentally different” from building traditional software. The differences are not cosmetic. On gross margins, traditional SaaS targets 75-85% with near-zero marginal cost per user; AI-native companies run at 50-70% blended at scale due to inference costs. On product development, traditional software is customer-led — listen, build to spec; AI-native is possibility-led, requiring teams to understand what models can do and build ahead of capability curves. Engineering and QA shifts from deterministic to probabilistic. Sales moves from packaged products to the FDE (Field Deployment Engineer) model. Customer success, often low-effort in SaaS, becomes critical in AI-native because everything starts as a proof of concept requiring education. Pricing, historically seat-based ARR, becomes consumption or outcome-based and remains experimental.
Redpoint’s AI64 report had previously noted that AI companies reach $5 million in annualized revenue 13 months earlier than the prior SaaS cohort and raise at valuations carrying a 24-40% premium between Seed and Series C. The firm estimates that if the cloud software market is valued at roughly $600 billion, AI-native applications could expand that addressable opportunity at least threefold.
What VCs Are Funding
The investment implications are concrete. Redpoint’s 2026 outlook describes 2026 as the year pilots convert or quietly disappear, with AI-native apps positioned to leapfrog both legacy incumbents and the first generation of AI startups. The firm sees the attack vector for new entrants as speed: large incumbents, in their framing, “have more lawyers than engineers.” Sequoia’s decision to co-author and publish Dorsey’s organizational essay signals that the firm views the management model itself , not just the product , as an investable differentiation.
For investors, the Dorsey-Halligan-Redpoint alignment creates a coherent due diligence lens. Companies that retain conventional org design while deploying AI as a productivity layer are running what the Sequoia essay calls a “copilot” strategy — making the existing structure slightly better without changing it. Companies that reorganize around AI as the coordination mechanism are the ones the current VC consensus appears to be pricing at a premium.
Friction and Open Questions
Oxford Economics published a January 2026 report finding that many layoffs attributed to AI were corrections for pandemic-era overhiring. Analyst firm CCS Insight cautioned that universal adoption within a year underestimates institutional friction in regulated sectors; financial services, healthcare, and public infrastructure face regulatory intensity, labor frameworks, and legacy integration complexity that will slow structural change significantly.
Piper Sandler analysts maintained an underweight rating on Block after the announcement, noting that transaction losses increased to 18% of gross profit from 11% a year earlier. The organizational thesis is compelling; the financial execution remains unproven at scale. Dorsey’s 2026 guidance implies productivity per remaining employee more than doubling in a single year , a projection that multiple analysts have flagged as difficult to swallow.
The more durable question is whether the “world model” concept; a continuously updated internal representation of company operations that replaces managerial information routing, is engineering-feasible outside of a company like Block, which is remote-first, transaction-rich, and already machine-readable by design. For most enterprises, the raw material for such a model does not yet exist in structured form.
Whether the Dorsey-mode framework ages as organizational theory or as a justification for mass layoffs dressed in futurist language will depend on what Block’s financials look like in 18 months. In the interim, Sequoia has made its position clear: the firms that survive the current transition will be those that treat AI as an architecture decision, not a feature.
