A repost by Yoav Rechtman, a product and growth operator, has been making the rounds in founder and investor circles on X. The thesis is blunt: as AI absorbs more of the work done inside technology companies, the roles that survive will collapse into four archetypes. Brian Halligan — co-founder of HubSpot, senior advisor at Sequoia Capital, and a lecturer at MIT Sloan — flagged it publicly, writing: “Not sure if I agree, but it’s a really interesting take.” That caveat did not slow the spread.
The four categories mentioned are: product engineers and generalist builders who ship at high velocity using AI tools; security, SRE, and infrastructure specialists who hold together the growing volume of AI-generated output; socially adept, relationship-forward people in sales, customer experience, and people operations; and institutional adults in legal, finance, and governance roles who act as governors on an otherwise accelerating organization. The framing strips away job titles entirely and bets that underlying traits; velocity, systems thinking, social intelligence, risk management are what actually determine who stays.
What the Data Says
The thesis lands in a labor market where early displacement signals are real but concentrated. Goldman Sachs research found that unemployment among 20- to 30-year-olds in AI-exposed occupations rose by nearly 3 percentage points in the first half of 2025, outpacing the same cohort in other sectors. Entry-level hiring for software development, customer service, and clerical roles has slowed sharply as companies deploy AI. If AI were applied across the economy to every task it can currently handle, the same research estimates roughly 2.5% of U.S. employment would be at direct risk of displacement, a figure that rises substantially as adoption accelerates.
Employers announced more than 1.2 million job cuts in 2025, the highest count since 2020, and AI was cited in roughly 55,000 of them; about 4.5% of total layoffs, according to Challenger, Gray & Christmas. Amazon eliminated 14,000 corporate roles citing leaner AI-enabled structures. Salesforce reduced its customer support workforce by 4,000 after its CEO stated AI now handles up to half of internal workload. Goldman Sachs and Hewlett-Packard both said AI initiatives would reduce headcount, with HP targeting up to 6,000 cuts by 2028.
Employment growth has already slowed in sectors that fit squarely into the roles the viral framework would eliminate: marketing consulting, graphic design, office administration, and telephone call centers have all dropped below trend, Goldman found, amid reduced labor demand tied to AI efficiency gains.
How Investors Are Reading It
Halligan is a senior advisor at Sequoia Capital, which manages $85 billion and has backed more AI unicorns than any other firm. Sequoia has said it views AI as a “full-stack revolution” and committed $950 million in new funds for early-stage AI startups in late 2025 alone.
This framework aligns with how investors are currently structuring bets. The first category, the high-velocity generalist builder, maps onto the explosion of AI coding assistant investment. Anysphere, parent of Cursor, was reported to be in talks to raise at a $27 billion valuation. Andreessen Horowitz, which now directs more than 40% of its portfolio toward AI companies, has backed Harvey (legal AI, valued at $8 billion), Vanta (automated security compliance, $4.2 billion), and Rippling (workforce management, $11 billion): companies that sit at the intersection of the categories rather than being built around any single traditional function.
The security and infrastructure category is drawing the most capital in absolute terms. Andreessen Horowitz raised $15 billion across five funds in January 2026, with $1.7 billion directed specifically at AI infrastructure. The argument is straightforward: as AI-generated code, content, and decisions proliferate across organizations, the humans responsible for making it stable, auditable, and secure become structurally irreplaceable.
Sequoia partner Sonya Huang put a specific frame on this at the firm’s AI Ascent conference in May 2025: “We’re entering the Age of Abundance, where AI makes once-scarce labor available everywhere at near-zero cost.” That claim presupposes that the scarce thing is no longer execution volume, but the judgment and relational intelligence that sit above it, which is precisely what frameworks “hot people” and “grown ups” categories represent.
Halligan’s Own Thesis on What This Means for Founders
On X, Halligan has separately described the fastest-growing startups he observes as running with “very very lean headcount,” mostly homegrown exec teams, and 80-plus hour weeks. The framework gives that lean structure a taxonomy. The implication for founders is that over-hiring into middle-layer roles — project management, content production, data entry, internal coordination — is now both financially inefficient and strategically exposed, because those are the first roles AI replaces.
The counterargument, which Halligan himself gestured at with his initial hedge, is that the framework is too clean. Real organizations contain hybrids: a sales engineer who also ships demos, a legal counsel who writes product specs. And some of the most structurally vulnerable roles in the schema; the junior analyst, the entry-level support agent, are also training grounds for the judgment that makes “grown ups” valuable over time. Removing them shortens the pipeline.
PwC’s 2025 Global AI Jobs Barometer found that AI-exposed roles grew 38% in job postings, but the premium accrued almost entirely to workers who had demonstrated AI skills. The four-job framework is less a prediction about titles and more a bet on which underlying capabilities AI cannot yet replicate at the speed and cost required by a technology company. So far, the data supports the bet.