A Cautionary Tale of Fools’ Gold and Cult-Like Practices
The pitch deck was beautiful. Fifty slides of gradient blues and deep oceanic greens, the logo a stylized eye — slightly too wide, slightly too unblinking — above the tagline: “Dive Deeper. See Further. Become More.”
Innsmouth AI raised $47 million in eighteen months. Nobody who visited the office could quite explain what the product did, but they left the meetings feeling strangely calm, strangely certain, and with a strong desire to return to the coast.
The Founding Story was impeccable, as founding stories always are. Three MIT dropouts, a Stanford professor of “Deep Learning and Deeper Architectures,” and a CEO named Obed Marsh IV who spoke in a hypnotic cadence and never seemed to blink at a normal rate. The company had emerged from a small town in Massachusetts — quaint, the press releases said, historic — where the founders claimed the cold ocean air had given them a unique perspective on intelligence, transformation, and what it truly meant to adapt.
The core product was called The Change. A personalization engine, they said. An AI that didn’t just learn your preferences — it restructured them. Made you more efficient. More focused. More willing to let go of old attachments like sunlight, land, and the company of people who asked too many questions.
The red flags were abundant, in the way that red flags always are: obvious in retrospect, glamorous in the moment.
The equity structure was opaque — something about tidal vesting schedules that only fully unlocked after a “deep immersion period.” Employees who left the company were described, cheerfully, as having “returned to the water.” The office had no windows but a tremendous number of fish tanks. Friday all-hands meetings were called The Congregation. The snacks were exclusively seafood.
Investors who pressed for metrics were invited on a retreat to the Innsmouth compound, and came back placid and unquestioning, smelling faintly of brine.
A journalist from TechCrunch filed three paragraphs of notes before her editor noticed the copy read only: “They are so right about everything. The gills are coming along nicely.”
The product itself — when it shipped — did work, in a fashion. Users reported their productivity had increased dramatically. They reported sleeping better. They reported a loosening of old anxieties, old identities, old shapes. The churn rate was essentially zero, because the users stopped using other apps entirely, stopped using most things entirely, and began spending an unusual amount of time near harbors.
The App Store reviews were five stars across the board. The most common phrase: “I don’t know how I lived before this. I don’t think I’ll need to go back.”
The Series C fell through when a due diligence team from a major VC firm submitted a report that consisted entirely of wet paper and what appeared to be scales. The SEC opened an investigation, but the lead investigator transferred herself to a field office in Newfoundland before it concluded. Glassdoor reviews alternated between “Best job I ever had, worth every transformation” and single, long, unbroken strings of the letter G.
By the time the Wall Street Journal ran its exposé — “Innsmouth AI: Something Fishy in the Valley’s Deepest Bet” — the founders had relocated the company offshore. Literally offshore. The servers, they said, ran cooler beneath the continental shelf.
The lesson, as with all cautionary tales of fools’ gold, is a simple one:
When a startup promises to change who you are at the deepest level, when its community feels less like a company and more like a calling, when the ask is not just your money but your self — and when the HR department is located in a sea cave —
it is reasonable to ask a few more questions before signing the term sheet.
Though by that point, of course, you probably won’t want to.
The water is very warm. The depth is not so frightening as you feared. And the old life — the dry, limited, surface-dwelling old life — already feels very far away.
Innsmouth AI is currently valued at $2.3 billion and accepting applications for its Deep Immersion Fellowship. Relocation to the Atlantic shelf required. Gills optional but encouraged.
