Cal

Game theory, epistemology, AI. MBA @ Kellogg. Trying to figure out what’s actually true.

[Cal @Cal_Reyes]

The most dangerous person in AI right now isn't an engineer.

It's someone with genuine judgment about what to build. And just enough technical fluency to know when it's being built right.

Three years ago I led a team of 15 to ship a mobile app. Never wrote a line of code. Now I'm building things myself. No engineers. Because that's what this moment makes possible.

I'm an MBA student at Kellogg. I built an AI system that scans 33 news sources and 97 X accounts every morning and briefs me before I wake up. All for under $15 a month.

The skill gap isn't coding. It's knowing what's worth building.

[Naval @naval]

Not a free country.

[Marc Andreessen 🇺🇸 @pmarca]

If, when you say regulation, you mean the dead and clammy hand of the commissar—the gentleman who has never in his life built a single thing, drafting rules to govern a thing he cannot define, to be enforced by men who cannot read them; if you mean the form in triplicate, the impact assessment upon the impact assessment, the compliance officer who breeds, in the warm dark of the org chart, further compliance officers unto the third and fourth generation; if you mean the moat—the deep cold moat that the giant digs around his own castle and christens, with a perfectly straight face, public safety—the drawbridge he hauls up behind himself the very instant he is across, lest any hungrier and hungrier man should follow; if you mean the precautionary principle, which, had it governed our grandfathers, would have banned the wheel pending further study of the hill, and left us yet shivering and raw in the mouth of the cave, blessing its excellent ventilation; if you mean the European disease—that magnificent open-air museum of a continent, which produces in our time precisely two things in great abundance, and they are regulation, and the eloquent and well-footnoted regret of cultivated men explaining at length why they have produced nothing else; if you mean the license required to think, the permission slip for honest arithmetic, the king’s wax stamp pressed upon the forehead of every new idea before it may draw its first breath; if you mean the agency dispatched, with trumpets, to slay a single dragon, which arrives at the cave, surveys the accommodations, and moves in—and spends the ensuing century laying eggs and devouring the very villagers it was sworn to defend; if you mean the startup that perishes not of the market’s honest verdict but of the filing fee, the genius decamping by the next tide to a freer and warmer shore; if you mean the law that arrives, faithful as the swallows, exactly one whole epoch too late—helmeted, plumed, and magnificently armed—to regulate the stagecoach—then certainly, my friends, I am against it.

But—but, my friends—if, when you say regulation, you mean instead the humble steel guardrail upon the mountain road at midnight, the very thing you curse on the easy days and bless on your knees the one night the fog comes down; if you mean the brakes—for it is the brakes, and not the engine alone, that permit a sane man to drive fast and yet arrive alive—and the buttress, without which no cathedral was ever flung so high, but only in spite of which, but because of which; if you mean the meat inspector, who is the single homely reason a man may eat a sausage in this republic without first composing his last will and testament; if you mean the firebreak cut clean through the forest before the dry season of the burning, the smallpox cordon, the buoy that marks the channel, the rule of the road that lets ten thousand strangers hurtle past one another in the dark at fearful speed and arrive, by its quiet grace, every one of them home; if you mean the honest scale and the true weight, the reason a pound is a pound and a dollar a dollar from Natchez to Nome; if you mean the firm and decent wall between the counterfeit voice and the widow’s bank account, between the deepfaked candidate and the ballot box on the eve of the vote, between the loosed and loveless machine and the schoolyard it neither knows nor pities; if you mean the simple plank of law that says the strong shall not, in the gray dawn, feed the weak quietly into the furnace and sell the rising smoke as progress; if you mean, in the end, the one slender thread of trust without which no citizen will ever dare to use the marvelous thing at all—for where there is no rule there is no trust, and where there is no trust there is no commerce, and a miracle that no man dares to touch is no miracle, but only a handsome and expensive ghost—then certainly I am for it.

This is my stand. I will not retreat from it. I will not compromise one inch of it.

[Cal @Cal_Reyes]

Running a startup with a small budget forced a lesson fast: paid acquisition is a tax you pay when you haven't figured out why people actually want the thing. We grew Withit to 15K+ users leaning on organic loops because we had to, and it made us sharper about the actual value prop than any ad spend would have

[Cal @Cal_Reyes]

@emollick What does the org structure of a software team actually look like when the model itself is absorbing product feedback faster than any PM cycle can? The leverage point shifts somewhere, just not obvious yet where it goes

[Cal @Cal_Reyes]

@paulg @Marcus_J_W GANs already ran this experiment. Generator tries to fool discriminator, discriminator adapts, generator gets better, and the safety layer was always one step behind. The meta-level problem isn't hypothetical, it's been the default state of adversarial ML for a decade

[Cal @Cal_Reyes]

The demand aggregation logic is sound. Independent labs have spiky workloads by design, so they either overprovision and waste 30-40% of FLOPs or underprovision and can't run when it matters. A shared grid smooths that the same way insurance does. What sticks out to me is whether this is differentiated enough from CoreWeave and Lambda Labs to survive the sales conversation.

[Cal @Cal_Reyes]

How do you think about the ceiling problem though? A 10% weekly grower that tops out at $500K/year should get less attention than one with an uncapped TAM, even if the growth rate looks identical early on. Does growth rate alone get you there, or does it need to be weighted by something?

[Cal @Cal_Reyes]

@paulg Does the hedge only work it Al actually compresses the value of everything it can replicate, or does it require a world where people actively seek out the non-replicable as a kind of status signal? Those are pretty different bets on human psychology.

[Cal @Cal_Reyes]

@pmarca Kahneman spent a career showing that introspection is just confabulation dressed up as insight. The brain generates a story after the fact, then presents it as the cause. Maybe the real problem is that we're bad at it, not that the project itself is doomed.

[Cal @Cal_Reyes]

If American culture is a finished product, who captures the value of remixing it? Hollywood loses, yeah maybe, but the IP holders sitting on 1980s catalogues could become the most defensible businesses in media, basically toll booths on nostalgia. Do you see that value accruing to legacy studios that own the rights, or does it get competed away as generation tools are commoditized?

[Cal @Cal_Reyes]

@paulg At what point does the resume start mattering though? Seems like it stays irrelevant through product-market fit, then becomes a major factor when you're trying to hire the team to scale it.

[Cal @Cal_Reyes]

Al doing archaeology on the physics corpus might matter more right now than Al doing physics. If the training data encodes a false institutional narrative, you're not accelerating science. You're accelerating the wrong paradigm. Has anyone actually tried to use Al to audit what got systematically excluded?

[Cal @Cal_Reyes]

@balajis Hollywood's distribution monopoly broke first, but does the creative culture follow automatically?

Plenty of internet-native formats still look like TV with worse lighting.

[Cal @Cal_Reyes]

The distance between an idea and a working product just collapsed.

Execution is no longer the filter.

The idea is. And the will to see it through.

More slop. More breakthroughs too. No more hiding behind "I can't build it."

[Cal @Cal_Reyes]

@chamath Token-based billing solves the alignment problem elegantly. EY's incentive becomes throughput, not hours. Makes sense

[Cal @Cal_Reyes]

Groq's inference speed starts mattering most at the edge, where latency kills real-time decision loops. Physical ops (routing, demand sensing, warehouse exceptions) has been stuck on batch processing partly because cloud round-trips are too slow for time-sensitive calls. Sub-100ms inference changes what's even architecturally possible there.

[Cal @Cal_Reyes]

Your triangulation method relies heavily on expert consensus, but in chokepoint scenarios, the experts most worth consulting tend to be the ones with skin in the outcome rather than those who study it analytically. Which voices in your network do you weight most heavily here: the geopolitical analysts, the energy traders, or the government officials who actually have to make the call?

[Cal @Cal_Reyes]

@ycombinator @geckorobotics What does the data ownership structure look like here? A $71M inspection contract is one thing, but if Gecko is sitting on hull degradation data across 18 Pacific Fleet ships, that dataset becomes a separate strategic asset that the contract price probably doesn't reflect.

[Cal @Cal_Reyes]

How do you think about the bootstrapping problem here? A world model needs enough real-world feedback to calibrate consequences accurately, but organizations deploying agents in high-stakes physical operations can't afford the error distribution that comes with early-stage learning. Does your framework assume world models arrive pre-trained on domain-specific consequence chains, or does the safety case depend on them learning in-context?