We're all short intelligence
1. Key Themes
The AI IPO Supercycle Is Concentrating Financial Markets Around a Single Narrative
The scale of upcoming AI-related public market events is historically unprecedented, and Rechtman frames it as a reflexive feedback loop where financial enthusiasm is overtaking technological enthusiasm.
"Anthropic filed its S-1 and SpaceX goes public next week. There's ≈$3T of market cap hitting in the next few months with at least another trilly on the way (OAI). Google is selling ≈$80B of stock and it's not much of a story. We are witnessing the total takeover of already-concentrated financial markets by a single story."
"The exuberance about AI the technology will have become fully overtaken by exuberance for AI the financial asset."
Foundation Models May Be Great Products But Structurally Terrible Businesses
The bear case is laid out with precision: training costs are ongoing, models depreciate fast, and competitive pressure from well-capitalized incumbents and open-source distillation prevents pricing power.
"Training is a cost of revenue, not an R&D cost. Stop training, everyone churns and moves on to your competitor or a distilled Chinese version. The labs are perpetually investing unbelievable sums in building these fast depreciating assets (capex and training)."
"Every model is bigger and more expensive to run and also gets distilled faster, so these companies have to make huge up-front investments that depreciate extremely quickly. They will just run out of money and places to get it."
Everyone With a White-Collar Job Is Structurally Short Intelligence — And Must Get Long
Rechtman reframes AI exposure not as a speculative bet but as a structural hedge every knowledge worker must take, collapsing the distinction between investment thesis and personal risk management.
"Everyone who works a white-collar job is structurally short. We're all waking up every day betting that our intellect and creativity will matter. Nobody has any incentive not to be in it and every incentive to be in it."
The Bubble May Be Worth It for the World, Even If Not for Equity Holders
Drawing the railroad/crypto analogy, Rechtman argues the infrastructure and capability overhang created by the bubble has positive externalities regardless of whether the labs themselves survive.
"The intelligence and capabilities are a positive externality of the bubble/exuberant build outs. And the more distillation happens, the more the frontier gets eaten by open source, the better for everyone."
"It does seem clear by now that as long as we don't completely bungle the second order impacts...that we are all going to be the winners. Productivity is the main thing."
2. Contrarian Perspectives
The Stock Price Is Not a Verdict on the Business
Against the consensus view that market cap validates business quality, Rechtman explicitly decouples the three layers — product, business, and stock — warning that the coming liquidity bonanza will not answer the real question.
"None of this answers the actual question of the business vs the stock and the bonanza won't settle it one way or the other. A business is not a stock, and a stock is not a product; the price won't be a verdict."
Evidence: He notes that even at potentially bubble-level prices, the incentive structure forces participation, which means price signals will be reflexively distorted rather than fundamentally grounded.
Distillation and Open Source Are a Feature, Not a Threat — For Society, Not Labs
While the industry treats Chinese distillation and open-source model proliferation as competitive threats to be defended against, Rechtman inverts this: the faster distillation happens, the better the outcome for humanity, even as it destroys the economics of frontier labs.
"The more distillation happens, the more the frontier gets eaten by open source, the better for everyone."
Evidence: This is consistent with the bear case on lab economics — distillation compresses the window in which labs can monetize expensive frontier models before cheaper alternatives emerge.
Betting the Bull Case Is Rational Even If You Believe It's a Bubble
The conventional framing is: if it's a bubble, avoid it. Rechtman argues the opposite — even bubble-believers with low enough cost of capital or structural short exposure should get long.
"On balance, it's almost certainly better to bet on the bull case, despite how much upside has already been sucked out / pulled forward at these prices. If there's any chance the bull case is right, and you have either a low enough cost of capital or are otherwise short the proliferation of intelligence, you have to get exposure here."
3. Companies Identified
- Description: Leading AI foundation model lab
- Why mentioned: Filed S-1, part of the ~$3T market cap wave hitting public markets
- Quote: "Anthropic filed its S-1...There's ≈$3T of market cap hitting in the next few months"
- Description: Elon Musk's aerospace and satellite company
- Why mentioned: Cited as going public imminently as part of the same concentrated liquidity event
- Quote: "SpaceX goes public next week"
- Description: Creator of ChatGPT and GPT model series
- Why mentioned: Identified as the next trillion-dollar liquidity event after Anthropic and SpaceX
- Quote: "At least another trilly on the way (OAI)"
- Description: Alphabet's core search and AI business
- Why mentioned: Selling ~$80B of stock, yet framed as almost unremarkable against the scale of AI lab listings
- Quote: "Google is selling ≈$80B of stock and it's not much of a story"
4. People Identified
- Description: NYC tech veteran and entrepreneur (co-founder of Meetup)
- Why mentioned: Built a rapid-response mini app to find Knicks watch parties around NYC, cited as an example of a "real NYC tech OG"
- Quote: "Scott Heiferman (a real NYC tech OG) rolled a mini app to find Knicks watch parties around the city"
- Description: Partner at Slow Ventures, leading pre/seed rounds from a ~$325M fund
- Why mentioned: Author; generalist investor focused on AI second-order effects, real-world businesses, hybrid software, healthcare, network effects, and fintech
- Quote: "I'm a generalist investor looking for weird takes on important stories: N-of-1 companies taking non-obvious approaches to markets that matter."
5. Operating Insights
For Founders: Training Cost Classification Is a Business Model Question
Rechtman's bear case hinges on a specific accounting and operational insight — training is a recurring cost of revenue, not a one-time R&D investment. Founders building on or adjacent to foundation models should pressure-test whether their core capability has a similar cost structure, because it determines whether the business can ever reach sustainable margins.
"Training is a cost of revenue, not an R&D cost. Stop training, everyone churns and moves on to your competitor or a distilled Chinese version."
For Investors: Cost of Capital Determines Whether the Bull Case Is Actionable
Rechtman frames AI exposure as rational for those with low cost of capital or structural short exposure — implying that the type of capital matters as much as the thesis. Patient capital (long-duration funds, personal balance sheets) can absorb reflexive bubble dynamics; short-duration capital cannot.
"If there's any chance the bull case is right, and you have either a low enough cost of capital or are otherwise short the proliferation of intelligence, you have to get exposure here."
6. Overlooked Insights
The Railroad vs. Crypto Framework for Evaluating Bubble Utility
Rechtman briefly introduces a two-scenario model for assessing whether a technology bubble is worth it: railroads (infrastructure surplus that created lasting real-world value) versus crypto (largely a wash, though it did accelerate chip development that enabled AI). This framework is underexplored but offers investors a practical mental model for distinguishing productive bubbles from extractive ones — and Rechtman leans toward the railroad outcome for AI.
"Will it be like railroads, which ultimately produced a ton of valuable infrastructure (I think yes), or like crypto, which was basically a wash (though of course we wouldn't have AI without all the chips and stuff from crypto)?"
"Every Task Is Actually a Coding Problem" as a Hidden Scaling Thesis
Buried in the bull case is a specific and non-obvious claim: that LLMs may not need to demonstrate new scaling laws across all domains if every human cognitive task can be reframed as a coding or structured reasoning problem. This would mean scaling observed in code generation generalizes universally — a much stronger claim than mainstream AI discourse typically asserts.
"LLMs either scale in other domains like they've done in coding or it turns out that every task is actually a coding problem after all."