Sex, Lies and Short-Selling Venture Stocks
What I found
I arrived in Silicon Valley inspired by Steve Jobs's Stanford commencement. To stay hungry, to stay foolish. What I found was quite different.
Whole Foods doesn't carry food that's actually good for you. Most of what's marketed as healthy is optimized for shelf life, label claims, and margin.
Yoga studios don't carry teachers who can transmit the practice. Most are credentialed, charismatic, sequenced for retention, and missing the thing the practice was supposed to be.
Stanford grad school didn't carry conversation about science or the big questions. Most of the talk was about funding, position, and the next paper. (There's a working theory that Stanford doesn't exist. After my time there I have some sympathy for it.)
Founders don't carry AGI. They carry AGI in the adjusted-gross-income sense.
It feels designed. It isn't. It's the emergent product of greedy optimization at the individual level, run against a robust optimization protocol at the system level. Adam Smith's invisible hand was supposed to coordinate this. It would have, if the inputs were truthful.
The inputs aren't truthful. Information has been biased toward conformity by the internet, which selected for self-significance signals over reality signals, then trained an entire generation to fit itself into a small number of artificial identity categories — not many can fit inside a single human mind. The system ends up running on performance, not truth. The Wealth of Nations turned into a networked Shakespeare play. Price signals became performance signals. The capital allocates to the actors — but only those playing one role. The hero's arc, monotonically up. The rest of the drama spectrum — tragedy, satire, dark comedy — has no audience and no green-lighter.
Why venture was long-only
Long-only venture corrected an asymmetry. For most of Silicon Valley's history — 1971 to 2010 — mainstream capital priced tech possibilities at near-zero. Every actor — founder, fund, LP, press — bet directionally up, because that was the corrective trade against a world betting directionally flat. Within a generation, the largest companies in the world were tech. The original mispricing closed.
It overshot. The world over-believes now. Every category. Every founder narrative. Every round. The asymmetry inverted around 2014. Long-only became a noise amplifier.
Short-side correctives are what's missing. Public markets have them.
The missing short
Public markets have a structural counterweight: short-sellers. A class of capital whose profit comes from finding overvalued companies, exposing the lies that hold the valuation, and being paid by the correction. Enron, Wirecard, Theranos — surfaced by short-sellers, journalists, and skeptics with financial incentive to find the lie. The market clears partly because some participants profit from finding things to short.
Venture has no equivalent. No class of capital profits from exposing that this round is overvalued, this metric is inflated, this category is hype. Short-selling a private company is structurally impossible: shares are illiquid, the cap table is opaque, the holders are aligned. The mechanism that prices in lies in public markets does not exist in venture.
The result is alignment around the lie. Every actor is incentivized to make the story true rather than to find out whether it is. The signal is the noise.
The gossip network
Silicon Valley is, among other things, a gossip network. In the absence of formal short-selling, gossip is what's supposed to carry the short signal — word that this founder is overstating revenue, this round is propped up, this company is missing its plan — passed informally through dinners, group chats, partner meetings. It should price in the lies the formal market can't.
It doesn't. The gossip network is too small, and it is the founder's own social graph. The same people who would expose the lie are also the people the founder needs in his next round, on his cap table, at his wedding. Gossip in SV doesn't travel as truth-seeking; it travels as relationship-maintenance. Negative information gets routed around the founder's personal network rather than through it, because routing it through has unpriced social costs no participant wants to absorb.
The founder knows this. He knows the network has the truth, and he knows that admitting the truth into his own network will cost him the next round. So he begins, gradually and then completely, to lie to himself. Not deceiving anyone else first — deceiving himself, so the gossip network has nothing accurate to carry. The lie has to start at the founder before it can be carried as truth by the network.
"Building in public" was a response to this. If the founder publishes the metrics openly, the network can't gossip a different version. But "building in public" gets exploited — selectively chosen metrics, narrative-shaped updates, transparency as performance. The founder publishes the parts of the truth that support the story and omits the rest. The audience now thinks it's reading the company's actual signals; it's reading the company's curated brand.
Jeffrey Pfeffer's Leadership BS documents the dynamic in corporate settings — leaders perform leadership rather than do leadership, because the audience rewards the performance and can't verify the substance. The SV version is the same on steroids: the audience funds the performance, the network protects the performer, the founder convinces himself the performance is the substance because admitting otherwise is the same as quitting.
Most companies at scale are run by actors. The system selected for them and trained the rest to perform alongside.
What the framers actually built
The Constitution didn't try to make participants honest. It built a system where dishonest participants couldn't coordinate.
Opposition parties, separate branches, independent press, lifetime-appointed judges, fifty competing state experiments — each with structural incentive to expose the others' lies. Truth surfaces through competitive contradiction, not through aligned reporting. The system doesn't require honest participants. It requires participants with conflicting interests in the truth being known.
The venture ecosystem has none of this. No opposition party. No short-seller. No independent press whose business model rewards exposing portfolio-company lies. The gossip network, which should approximate adversarial structure, doesn't — it routes around the truth instead of toward it.
A 1787 v2.0 in venture — a mechanism that aggregates truth in an ecosystem aligned around the lie — is the design problem nobody is working on, because the people who could work on it are inside the system that benefits from the lie.
The Prestige problem
Even adversarial structure has a deeper layer. From Christopher Nolan's The Prestige:
"You're not really looking. You don't really want to know. You want to be fooled."
Founders don't tell themselves the truth about the company. The truth would dissolve the narrative they need to keep going. The founder who wakes up at 3 a.m. and honestly asks whether the thing he's been telling investors for two years is real often stops if he answers. The narrative is the engine. The lie is the fuel.
No external mechanism reaches this layer. The founder is alone with himself. The only check on self-deception here is the operator's own nervous system — calibrated against years of contact with reality, capable of detecting the moment a story starts compensating for a fact.
The mechanism from outside does part of the work. The rest is interior. The short-seller for your own story is a person who can hear when his own story starts to lie.
Where this leaves us
The structural truth-seeking mechanism for venture doesn't exist. The interior one for the founder rarely does either. Default state: alignment around the lie, outside and inside.
Long-only corrected for forty years. The asymmetry inverted. What comes next requires short-side correctives at the system level and an interior short-seller at the personal level.
Until both exist: short your own story before the market does.