The "Lean AI Team" Thesis Hits Reality: Chatbase Founder Reverses Course, Plans to Double Headcount

A startup founder who championed skeleton crews now calls that bet "quite wrong" — arguing AI's speed gains should raise ambition, not cut jobs. The counter-narrative to Block and Coinbase's mass layoffs.

Yasser Elsaid, founder of customer-service AI startup Chatbase, is doubling his 40-person team over the next six months — a direct reversal of his earlier conviction that AI-era companies should run lean. His October 2026 post on X admitted his previous stance was "quite wrong," and his reasoning reveals a fault line in how the tech industry thinks about AI-driven productivity.

The prevailing Silicon Valley consensus is clear: AI makes humans redundant, so cut them. Block shed nearly half its workforce in February 2026, with Jack Dorsey citing AI and "smaller, flatter teams." Coinbase cut 14% of staff in May, arguing engineers could now do in days what previously took teams weeks. Sam Altman has predicted one-person billion-dollar companies. Cursor reportedly scaled from $1 million to $100 million ARR in under a year with fewer than 50 employees. Elsaid's counter-argument is structurally different from a jobs-protection narrative. He contends that if AI accelerates development speed, the rational response is to raise the difficulty ceiling — build harder things — not to shrink the team. "If development got easier, build harder things. Much of a software company's competitive advantage comes from someone working on something hard for a long time and being right about it." The competitive moat, in other words, is still human judgment applied over time. The practical evidence from Chatbase supports this. Elsaid reports that roughly 90% of his engineers' work is now product development — conducting research, talking to customers, forming opinions, and translating those into specifications that AI agents then code. The remaining work requires engineers who can read code and ensure multiple AI agents don't produce contradictory outputs. The bottleneck moved from writing code to deciding what code should exist. Chatbase is also hiring in sales and customer success — roles that require human persuasion and relationship management with enterprise clients. This is the unglamorous reality that the "AI replaces everything" thesis consistently underweights: enterprise customers still need humans to trust, and trust still requires humans. Elsaid's framing exposes a genuine strategic fork. One path treats AI productivity gains as a cost-reduction tool — same output, fewer people, higher margins. The other treats them as a capability multiplier — more output, same or more people, higher ambition. The first path is immediately legible to investors. The second requires patience and conviction that harder problems yield durable competitive advantages. The honest caveat is that Elsaid himself acknowledges this may be temporary. "The day may come when humans are not needed in the development process at all. But that day is not today." Chatbase is a 40-person startup with a specific product thesis, not a universal template. But his reversal is a useful data point against the reflexive assumption that AI + fewer humans is always the correct equation.