Item #1: Trump, Who Has Amassed Enormous Wealth in Crypto, Hosts Industry Leaders.
Bitcoin recovered from around $63,000 to $77,000 on renewed optimism about the Clarity Act. Or for no reason at all. It doesn’t matter. The gains will not hold. Nothing changed between now and 3 weeks ago. The media keep pushing this narrative that crypto donors “bought out” the presidency. Hardly. What donors want is direct investment by the government of crypto assets, similar to other sectors which got direct investment despite their CEOs having donated nothing to Trump’s campaign.
Meanwhile, crypto donors and industry leaders have nothing to show for their support but occasional meetings that amount to empty promises: no direct investment, no Bitcoin reserve, or even any shoutouts on Twitter endorsing crypto. Eric and Trump’s cost basis is zero, and used the inauguration hype to sell soon-to-be worthless tokens to gullible investors at inflated prices. That is the key difference: Trump and his son licensed their brads to launch various crypto projects, so the downside is shitted to investors, not them. This is a no-brainer. Iran and tariffs will not get better, further dragging Bitcoin lower. I added more to my short positions to profit from this.
I was the only one who predicted in 2024-2025 that Trump would let his crypto donors down. If there were a list of sectors or companies for consideration of direct investment: metals, semiconductors, quantum computing, Intel, Dell Computers, MicroStrategy, Bitcoin–in 2024/2025 everyone would have picked the last two, and never in the million years the others. Yet the opposite happened. I was the only one who saw this.
#2 Pangram is very accurate.
Ironically, the hottest AI start-up is a company that uses AI to try to thwart AI, that being Pangram. It’s almost supernatural how accurate it is. The above three paragraphs about Bitcoin are 100% human (try it), but rewriting it below with Chat GPT, although not much different in terms of content and style, comes out as “100% AI.” I even instructed it to write at a 7th grade level, to make the grammar less formal, and remove em-dashes. Not good enough.
If you had shown people a list in 2024 or 2025 of things the government might invest in, like metals, semiconductors, quantum computing, Intel, Dell, MicroStrategy, and Bitcoin, pretty much everyone would have picked the last two. Almost nobody would have guessed the government would invest in the others instead.
Bitcoin recovered from around $63,000 to $77,000 because of new hope about the CLARITY Act, or maybe for no real reason at all. It does not matter. These gains will not last. Nothing important has changed in the past three weeks.
The media keeps saying crypto donors “bought” the presidency. Hardly. What they really wanted was for the government to directly invest in crypto. Other industries got government support even though their CEOs gave nothing to Trump’s campaign. Meanwhile, crypto donors have almost nothing to show for their support besides empty promises. No direct investment, no Bitcoin reserve, and not even a post from Trump telling people to buy crypto.
Donald Trump and Eric Trump also had almost no cost basis in their tokens. They used the hype around the inauguration to sell tokens that will probably end up worthless to gullible investors.
Pangram is the answer to slop. Schools and media companies will adopt it, if they aren’t already, to detect AI-generated writing. I don’t see any easy way to thwart it.
Item #3: Iran and Polymarket.
Of course, the Iran situation isn’t getting better. I remember on August 2nd there were headlines of yet another “deal”, which proved as short-lived as the last 4 or so:
It’s the same pattern: Trump announces deal, Iran denies deal, stalemate, more strikes, strait closes, Trump announces new deal…
My prediction of the strait not fully reopening until 2028 at the soonest, went from a long-shot 5 months ago or easily dismissed as nonsense, to now even confirmed by Iran:
BREAKING: Iran declares the Strait of Hormuz will stay closed until 2029, the end of Trump's term, unless Trump accepts all of Iran's demands, per a senior Iranian source and Majid Shakeri, advisor to Parliament Speaker Ghalibaf.
Iran's demands in the US surrender document…
— The Hormuz Letter (@HormuzLetter) August 10, 2026
I had already figured out shortly after the war began how it would play out: Iran had nothing to gain by working with Trump. And, of course, Trump, being tied to Israel and Rubio wanting war, would never make any concessions either. The only possible outcome was a stalemate with the strait remaining impeded.
All Iran had to do was just wait as Trump’s poll numbers keep falling, and for Trump’s successor, presumably Vance or Rubio, to lose. Iran’s leadership isn’t that upset about losing their Ayatollah. He was a sacrificial lamb for a much longer game. Having oil prices surge and remain high helps Iran’s economy and makes Trump look worse. I was also right about NATO countries not helping. Why would they. Like Iran, they are perfectly fine watching Trump dig himself into a hole.
I continue to profit by betting “no” on the strait reopening. This is some of the easiest money ever. I have been running this strategy since the war began. The odds of the strait reopening by Dec 2026 have fallen from 85%, when I opened the “no” position, to just 29%:
Conversely, the “no” contract has increased from 15 cents to 71 cents in only 3 months. This is further evidence of being one of the greatest forecasters out there. It’s all backed by quantifiable evidence and results. Some people make lots of bets on Polymarket or Kalshi; I make only a few surgical bets, where I’m certain because my “mental model” of the world is the correct one.
Item #4. Trump’s approval ratings falls to 33%, the lowest of his presidency.
As discussed earlier, it’s clear the tariffs have not helped. Same for the Iran War. Even though the online-right are not representative of the entirety of his base, his falling approval rating suggests a broad loss of support among likely voters. He won in 2016 because of immigration. That is the issue everyone cares about first and foremost. They see the demographic change unfolding before their eyes. They see that society has lost trust. They don’t want more useless tariffs or wars.
At this stage, the Trump presidency is effectively over, as I don’t see much he can do to boost his approval ratings or policy-wise, as he’s already fully committed to Iran. It would take a repeat of something like 9/11 to give him meaningful polling bump and enough momentum for his successor (although it helps greatly that the dems choose the worst possible candidates). Otherwise, this looks like a repeat of 2008.
Item #5: He’s right:
I don't think elite higher ed is going to become irrelevant but it's probably tracking to become ~50% less relevant in the new steady state, and that's probably a good recalibration.
— Nate Silver (@NateSilver538) August 20, 2026
I predict elite colleges will remain as relevant and important as ever. Despite rapid gains in AI since 2022, elite colleges have not lost their luster. Harvard, MIT, Caltech, and Stanford are still extremely selective, and top employers still see those degrees are valuable signals of intelligence and competence. Employers need some way to filter for talent, and top-20 colleges serve that function well, still.
Colleges and employers will adapt to AI, rather than the institution of higher ed becoming obsolete, as many are hoping or expecting. The status quo will not be upended by AI, as it has exhibited a remarkable ability at adapting to technological disruption over the past half century, such as the rise of the personal computer and the World Wide Web. Solving AI cheating is easy: make 100% of the grade based only on in-class assignments/quizzes–and or–combined with AI-detection software such as Pangram. Take-home tests were always a dumb idea, even before the AI era. Of course, students are going to avail themselves of outside help if they can.