I know how the AI bubble ends now
TL;DR
The tell is in the verb — Mo Bitar fixates on Allbirds saying it would "seek to acquire" $50 million of AI compute, arguing that the hedge word "seek" reveals a company signaling AI ambition without actually committing to become an AI business.
Allbirds' $50 million AI plan is framed as theatrics, not substance — he says $50 million of compute in 2026 is trivial given Nvidia scarcity through 2028, and the fact that the funding comes from an undisclosed investor makes the whole announcement feel even thinner.
This is the 2017 Long Blockchain playbook in new clothes — Bitar retells how Eric Watson turned a Long Island iced tea company into Long Blockchain Corp, without engineers or product changes, and watched the stock jump 400% in a day.
The AI bubble lowers the bar for scammy signaling even further — his punchline is that dot-com scams needed a website, crypto scams needed a white paper, and now AI hype can move markets with "just a damn verb."
The real damage lands on ordinary buyers, not the promoters — he imagines a plumber in Ohio chasing a 600% pop at $17 and getting crushed a day later, making the point that these pivots are funny until retail investors eat the loss.
If Allbirds wanted a credible AI pivot, it had one — instead of chasing GPUs, Bitar says its decade of supply-chain, warehousing, inventory, and delivery experience could have supported an actual AI logistics business, but that would require building real software.
The Breakdown
Russell Crowe, a hotel bathroom brawl, and the setup for a finance parable
Bitar opens with a completely unhinged but memorable story: in 2002, Russell Crowe allegedly gets into a bathroom fight at a London hotel with Eric Watson, a New Zealander whose day job involved selling women's underwear through Ellie McFersonson Intimates. It sounds like tabloid trivia, but he asks you to remember Watson because, 15 years later, he becomes central to a story about how bubbles really work.
The one-word scam detector: "seek"
His big thesis is that every scam has a linguistic tell, and this week's tell is the word "seek." He points to Allbirds' press release saying it will "initially seek to acquire" high-performance AI compute hardware, and mocks the phrase as the "metaphysical precedent of acquiring" — not doing the thing, just gesturing toward maybe doing the thing.
Why the Allbirds AI announcement smells off
Bitar argues the numbers don't hold up: $50 million of AI compute in 2026 is, in his telling, laughably small in a world where Nvidia is sold out until 2028. He also zeroes in on the money coming from an undisclosed investor, treating that as another red flag that this is more about surviving public scrutiny than building an actual AI business.
The Long Blockchain iced-tea farce
Then he rewinds to 2017: Eric Watson buys a controlling stake in an iced tea company in Farmingdale, Long Island, and during the crypto boom rebrands it as Long Blockchain Corp. The joke lands because nothing operational changed — the factory still made iced tea, "Gary packed a pallet of diet mango tea on Monday, and by Tuesday Gary was legally a blockchain engineer," with Bitar leaning hard into the absurdity of 14 factory workers all seemingly named Gary.
The stock pop, the broker calls, and the two-hour bag dump
After the name change press release, the stock jumps 400% in a day. Bitar says Watson had called broker Oliver Barrett Lindsay, who called Gannon Giguiere, who bought 35,000 shares the night before the announcement and sold them within two hours the next morning; the SEC eventually caught them, but Bitar shrugs that little really came of it.
From dot-com to crypto to AI: scam gravity keeps dropping
This is where he makes the broader point: each bubble needs less actual work to fake legitimacy. Dot-com required a website, crypto required a white paper, and AI now requires "not even the PDF, just a damn verb," which he turns into a great line about "Moore's law for scams" and a coming "pure photon scam" made entirely of intent.
The part that's actually bleak
Bitar pivots from jokes to consequences, imagining a plumber in Ohio seeing a bird stock rip 600%, buying in at $17 on Wednesday, and being down a third by Thursday. He pushes back on the "great companies pivot too" defense by contrasting Allbirds with Nintendo's century-long progression across adjacent entertainment categories.
The honest pivot Allbirds didn't choose
His closing argument is that Allbirds actually had a plausible lane: 10 years of consumer supply-chain experience in warehouses, inventory, and delivery. An AI logistics spinoff would at least connect to what the company knows, but buying GPUs is, in his words, the laziest legal way to say you did AI — and since regulators and media won't stop it, he jokingly calls on the only proven enforcer left: Russell Crowe.