NEW YORK - s4story -- Roy Lee, the 21-year-old founder of Cluely, spent most of 2025 doing something that sounded like genius in a pitch deck and looked like arson in retrospect. He bragged on stage at TechCrunch Disrupt in October that his annual recurring revenue had jumped from $3 million to $7 million in a single week, told an audience of founders that "reputation is sort of a thing of the past," and explained that rage-bait marketing was the only growth strategy that mattered in a saturated AI market. Five months later, in March 2026, he posted on X that the $7 million figure had been "the only blatantly dishonest thing I've said publicly online" — a confession that arrived after Andreessen Horowitz had already wired him $15 million in Series A money. Founders studying the wreckage tend to focus on the lying, but the more interesting failure is upstream: Cluely shipped narrative faster than it shipped substance, and the gap eventually swallowed the company. Operators who treat communications as an engineering discipline — the kind of structured external positioning work mapped out in this resource (https://techwavespr.com/services/public-relations/) on technology PR — generally avoid this particular failure mode, because the discipline itself forces the gap to close before a journalist or a competitor finds it. The lesson of 2025 and the first half of 2026 is not that attention doesn't matter. It is that attention without a defensible technical substrate is now a leading indicator of collapse.

More on S For Story

• Heracles: Ancient Greece, Cult Tradition, and Regional Myth
• Research reveals "The Borderless Pay Standard," a 48-point gap between multinational employers and workers on transparent pay expectations
• Global.ai Appoints Freedomtech Solutions as Specialist Partner for Agentic AI
• Liberty Book Writers Officially Launches to Help Aspiring Authors Publish Professionally
• Audiobooks Are Transforming Publishing California Book Publishers Helps Authors Reach Modern Readers

The Builder.ai Object Lesson

If Cluely is the small version of the problem, Builder.ai is the industrial-scale one. The London-based startup raised more than $445 million from Microsoft, SoftBank, the Qatar Investment Authority, Insight Partners, and Iconiq Capital on the premise that its AI assistant "Natasha" could generate working software the way DoorDash generates dinner. The pitch was clean. The product was theater. When the company collapsed into Chapter 7 in May 2025, reporting from the Financial Times and the Times of India revealed that the actual code was being written by approximately 700 engineers in India working manually behind a marketing curtain. An audit found that 2024 revenue had been inflated by roughly 300 percent. The U.S. Attorney for the Southern District of New York opened a federal investigation, subpoenaed the former CFO, and began assembling a grand jury in Manhattan.

What is striking about Builder.ai is not that it failed but that the warning signs were public for six years. The Wall Street Journal had questioned the company's AI claims in 2019. A former executive had sued in the same year alleging that investors and customers were being misled.