Frank D Prestia

Inventor, Attorney & Founder.
NASA, Blue Origin & Lockheed run my equipment.
What system governs the outcome?
Machine learning '80s → Philosophy → AI '91

[Frank D Prestia @FrankDPrestia]

In 1991, I applied for a federal grant to answer one question:

Is the human mind just an organic machine?

I was a philosophy undergraduate. My two main sources were John McCarthy, who founded the field of AI, and Hubert Dreyfus, who spent much of his career arguing it would never achieve human intelligence.

The question wasn’t fashionable then.

It’s the only thing anyone talks about now.

I never stopped asking it.

The only thing that’s changed is the direction it points.

[Frank D Prestia @FrankDPrestia]

@AnneMarieTrades 🙏 from dust to dust

[Frank D Prestia @FrankDPrestia]

The danger is not that AI beomes hateful or defiant:
https://t.co/LjebDkMfkE

[Frank D Prestia @FrankDPrestia]

https://t.co/T5q26pdh8S

[Frank D Prestia @FrankDPrestia]

Love it- applying everything to everything!

[David @Davo0820]

@FrankDPrestia @RosannaInvests I was told a long time ago the secret of life is in compound interest, in everything you do. Love, relationships, work and financing.

[Frank D Prestia @FrankDPrestia]

@Davo0820 @RosannaInvests Yes definitely Blessed. Don’t get me wrong, but good relationships still require work, sort of like luck… harder you work the luckier you get…

[Frank D Prestia @FrankDPrestia]

Rosanna’s financing-vs-demand distinction is on point.
I’d add that 2026 has 3 catalysts that can accelerate a financing event beyond what we saw in 1998:

1) Algorithms can mechanically de-risk without any change in fundamentals.

2) Retail can amplify price moves almost instantly through options, commission-free trading, and social media. Via social media narratives, AI can be declared the next internet or the next dot-bomb in the same afternoon.

3) Passive flows can spread selling across entire ETFs and indices rather than isolated stocks.

Same analogy but accelerated transmission mechanisms.

In a market this reflexive, the investor’s job is not to repeat the narrative. It is to underwrite the mechanism and stress-test what actually changed👇
https://t.co/DjWZHcNdou

[Rosanna Prestia, MBA @RosannaInvests]

🌍 The question isn't whether AI is the year 2000. It's whether we just lived through 1998, and the difference is worth a fortune.

Remember what 1998 actually was. Russia defaulted, LTCM blew up, and the Nasdaq fell roughly 30% in under three months. A leverage event, not a demand event. Internet usage never blinked. The forced sellers just had to finish selling. ✅

It was called the end of the internet bubble at the time. Then the Fed eased, the liquidations cleared, and the Nasdaq ran 272% off the October low in seventeen months. Nearly a quadruple. 🤯

Now look at 2026. Record margin debt, $1.5 trillion, photographed at the June top and unwinding since. Korea's KOSPI leveraged complex flushing through seven circuit breakers. High-beta down 30-60% while the index barely moved. And demand? Capex isn't just holding, it's accelerating: $GOOG just raised its 2026 spend to $195-205B tonight, a $15B hike, its second raise this year, with the CFO saying demand still outpaces the investment. Backlogs at records, contracts signing straight through the drawdown. A financing event, not a demand event. Same species as 1998. ✅

And here's where the analogy breaks in 2026's favor, and it's the most important paragraph in this post. The dot-com era traded on pre-revenue stories: companies with no earnings, no path to earnings, priced on eyeballs and press releases. Today the money is flowing into cash machines. $MU and $SNDK trade at single-digit forward earnings multiples while memory prices rise and supply is sold out for years. The hyperscalers writing the capex checks generate hundreds of billions in free cash flow. In 2000, the froth WAS the market. In 2026, the froth is a tier, and underneath it sit some of the cheapest cash generators in the entire tape. That difference is structural, not sentimental. ✅

Now the ledger, because analogs get chosen because they fit. This is one sample. The Fed hasn't eased. And if this is 1998, the sequel is 1999, the most euphoric, least disciplined phase of the entire era. The map that promises the rally also promises the ending. 1999's bill came due in March 2000, and it was 78% of the Nasdaq. ✅

Why do these patterns rhyme at all? Because markets are made of people, and people ship with the same firmware in every era: fear, greed, leverage, capitulation, repeat. The technology changes. The nervous system running the trades doesn't. That's why history rhymes and why the sequence is recognizable in advance. The cycle was never made of technology. ✅

The discriminator then was demand. The discriminator now is the same, and it reports quarterly. Capex and tokens decide whether this analog stays alive. ✅

1998 didn't reward the people who called the top. It rewarded the people who survived the flush, stayed for the run, and wrote the exit rules before they needed them. 😎

[Frank D Prestia @FrankDPrestia]

@Davo0820 Thanks. After 27+ yrs of marriage, raising our 3 sons, and working together across several businesses, @RosannaInvests and I have had plenty of practice… yet still working on it.

[Frank D Prestia @FrankDPrestia]

@Davo0820 Thanks, David. I appreciate it.
Critical thinking about most starts by asking what changed vs what just looks different.
That real helps to identify the mechanism, so you can test the assumptions, and let the evidence decide.

[Frank D Prestia @FrankDPrestia]

Thanks again Eric-you’re saying something that is going to force me to revise an article I already had in my queue, just like Simmons Hyperion.
Goldratt is fiction and the bottleneck ideas are strong and you are applying them strongly.
The question is whether cheaper inference lowers aggregate physical demand or makes far more inference economical. Going to have to revise and it’s going into the upcoming article.

[Frank D Prestia @FrankDPrestia]

Chips were the bottleneck. Then power. Next comes interconnection, transformers, copper, electrical steel, and refining.

Constraints relocate and economic value follows them through the stack.
https://t.co/IzCI3E6edW

[Frank D Prestia @FrankDPrestia]

https://t.co/pePBbwNjDQ

[Frank D Prestia @FrankDPrestia]

Eric, you do an amazing job extrapolating and applying.
You did it with Simmons and Hyperion, which pushed me to include it and led to a more dystopian article than I began with.

Regarding the efficiency gains, I I already released an article about Jevons, where he observed in 1865 that more efficient steam engines raised total coal consumption, because cheaper use made more uses economical.

If near-frontier performance runs on cheaper chips drawing less power, the number of workloads running it presumably expands, and the pressure keeps moving downstream to interconnection, transformers, electrical steel.

Goldratt is great because fiction is enjoyable and his contribution key- he’s in 2 other essays on this subject that I haven’t released. I’d also mention Thomas Hughes found the same pattern studying electrification. The idea was born on the electrical grid, but AI may be sending us back to it.

[Frank D Prestia @FrankDPrestia]

A market call is not underwritten because the conclusion sounds plausible.
What is the premise?
What is the transmission mechanism?
Which assumptions must remain true?
What would invalidate it?
That is the difference between holding a view and underwriting one.
https://t.co/PLOgPnf7V3

[Frank D Prestia @FrankDPrestia]

https://t.co/SX3RM1tbzp

[Frank D Prestia @FrankDPrestia]

An overlooked link is aggregation. Higher costs may impair some companies, that don't automatically mean the entire market tanks. You still need broad exposure, meaningful index weight, few offsetting winners, substitutes, or strong fiscal countereffects.
https://t.co/JSWq4lm4Jy

[Frank D Prestia @FrankDPrestia]

https://t.co/SX3RM1tbzp

[Frank D Prestia @FrankDPrestia]

Chamath called a stock-market crash “not debatable.” Then he explained it with suppose, if, and eventually.
The mechanism was conditional at every point but his certainty was not.
Investors should underwrite the chain on their own rather than inherit his confidence.

[Frank D Prestia @FrankDPrestia]

@Davo0820 Thank you, David. A tough day for many of us, but this is where underwriting matters. If your capital can remain invested and the time horizon is long, a day like today is painful without necessarily being meaningful.

[Frank D Prestia @FrankDPrestia]

Do you have a test for market predictions? Some sort of underwriting test? A test for whether confidence in the prediction has been earned?
https://t.co/JSWq4lmCz6

[Frank D Prestia @FrankDPrestia]

https://t.co/SX3RM1tbzp

[Frank D Prestia @FrankDPrestia]

@Davo0820 Yes just being upfront is only way to learn

[Frank D Prestia @FrankDPrestia]

At 2:00est today, the Fed gives markets the conclusion. From there, investors begin underwriting the mechanism.
Headlines can move prices. The assumptions beneath them determine whether the repricing deserves to hold.
https://t.co/JSWq4lmCz6

[Frank D Prestia @FrankDPrestia]

https://t.co/SX3RM1tbzp

[Frank D Prestia @FrankDPrestia]

Thank you David.
In fairness to Chamath, he reasoned very carefully and supplied a valid mechanism, but his prediction exceeded the evidence.
The Dunning is a little different since the person lacks competence. I remember shorthand Dunning=dunce who overestimates himself and Chamath may posture and overstate but he is not incompetent.
But yes, the fundamentals and discipline you're talking about is exactly what keeps an investor from trading on someone else's confidence.

[Frank D Prestia @FrankDPrestia]

@Davo0820 and repeatedly predicted...

[Frank D Prestia @FrankDPrestia]

The problem with a bold prediction is not that it may be wrong.
It is that the certainty may exceed the evidence supporting it.
https://t.co/JSWq4lmCz6

[Frank D Prestia @FrankDPrestia]

https://t.co/SX3RM1tbzp