OpenAI's GPT-5 Flop, AI's Unlimited Market, China's Big Advantage, Rise in Socialism, Housing Crisis
2025-08-09 spoken.md · speaker-labeled ▶ watch ← E237 all episodes E239 →
Every number here is replayed from score_events — the same ledger the pool ranks on. Decay is what the 105 ideas nobody mentioned gave up this week; it applies only when an episode is processed.
Tier crossings
conviction thresholds crossed by this episode — 65 / 45 / 15 · ideas born here show where they landed. why 65 isn't always green
Kill dates that landed since E237
1 hit · 0 partial · 4 miss — windows that closed after 2025-08-01 and up to 2025-08-09, auto-scored against price data and never hand-set. verdict · R · α
| idea | verdict | R | α | closed |
|---|---|---|---|---|
| 🌍 BOJ is trapped — Japan can never raise rates and the yen keeps depreciating | MISS | +1.1% | -19.9 | 2025-08-09 |
| 🏛️ Google's search-monopoly remedy is a broad consent decree, not a breakup | MISS | -23.7% | -44.7 | 2025-08-09 |
| 📈 Recession and an equity rally at the same time, because the Fed cuts | HIT | +21.0% | +0.0 | 2025-08-09 |
| 📈 The real economy is already squeezed while software keeps its pricing power | MISS | -23.1% | -44.2 | 2025-08-09 |
| 📈 You cannot build durable public equity value on a bet-making business | MISS | -30.8% | -48.9 | 2025-08-02 |
Who moved the board
each voice's force on conviction this episode — supports and opposes, weighted exactly as the replay applied them · share = % of this episode's movement
What got argued (22 ideas)
ordered by how hard each idea moved · quotes are verbatim from the transcript, timestamps deep-link into the episode
if they're unwilling to move to places where the real estate is cheaper, then you are going to get more expensive real estate. And maybe you can build your way out of that. Or maybe you can't.
Yeah, it's not comparable to Colossus. So it's just, it's not going to be decisive in this, you know, AI, you know, race that certainly China believes we're in.
Friedberg's socialism thread re-aimed at the United States rather than Europe: stagnant wages against visibly higher grocery, rent and household-goods costs, plus his estimate that nearly half of Americans are employed directly or indirectly by government, produce a one-way ratchet in which every failed program is answered by a bigger program, so taxation and regulatory burden keep rising and the end state is seizing the means of production. Ben Shapiro supplies the trigger and the timing - a market or employment downturn converts a 'the rich got richer' grievance into full-scale redistributionism further left than Obama, a redux of 2007-08 - which compresses the multiple on US risk assets and on the financial and wealth-linked sectors most exposed to redistribution. Gavin Baker and Phil Deutch take the direct other side: the policies fail fast and visibly enough to end the experiment quickly, and the country has always swung back.
I don't see that, but if you want to see that, you can, but it will swing back. This country is incredibly resilient, and I have complete faith in that resilience.
So they won't work. And I think that'll be a quick, that'll be like a quick end to socialism.
If things take a downturn, then the backlash is going to be a lot stronger than anything that any of us can say about the failures of Zoran Mamdani, because what it's going to be is, look, the rich got richer and we got poorer, and it's going to be a redox of 2007, 2008, but with a much, even further to the left than Obama ever was, and that I'm deeply afraid of.
But I'm just shocked at their inaction. And aptitude and AI, how bad the products are from a user perspective. I mean, they're terrible.
it's been many years since they brought out anything that excites me as a consumer
and that's when you will see the Federal Reserve step in and lower those interest rates in order to sort of artificially juice the economy
But you know, we are waiting on Gemini. I would expect there to be a very, you know, an impressive new Gemini model.
So all of the dollars that are going into the ground are being used. We're not building ahead of demand. In fact, we're clearly still behind demand.
But this is the first time they have not released a model that was a decisively better. And the progress for Grok 4 was exceptional. Grok 5 is coming soon, you know, they have more.
Anthropic just released a paper this month saying, we want 50 gigawatts of power for the AI in America and over the next three years. That's about as much power as was added all this year.
It'll take a year, two years for those policies to actually hit and destroy all of New York.
And so the point of this chart is if we really need all we could eat on energy for artificial intelligence, and if artificial intelligence, the race is the, you know, race for national security and dominance in the world. This is saying we're taking a step backwards on this supply.
Yeah, so in the small modular reactor, Gen 4 stuff, we as well as others have investments there. A great company we like called X Energy.
the idea that the federal government ought to prevent the state of California from filling the field with some really bad AI legislation that essentially destroys competition among all the various players
So just to give credit where credit's due, I think all the stuff we're seeing on the small modular reactors at some level, owes its renaissance to AI and the energy demands of AI.
Otherwise, China is able to sort of climb out the window, which is, I think, what they've done on everything from Nvidia chips to TikTok to their actual tariff rate, which I believe is now lower than the tariff rate we have against Canada.
I would just say he showed his hand in April. He looked into the abyss and if it is clearly not working, he showed that he will change it.
The second thing I would say is I do think they are making an effort to throttle it up.
And if you didn't control the chop, the thing that made the stamp, you didn't actually control the factory. That is not a long-term recipe for success.
Is this stuff going to hit fast enough to actually justify the amount of investment, or are you going to see a crater here?
You can calculate something called return on invested capital. It's actually gone up since they started making these AI investments.
I am inclined to think that he is much more sensitive to the market than he lets on.
Episode digest
written during extraction and stored in data/extractions/ep238.json — the auditable source of truth, including everything market-adjacent that did not earn a capture
Emergency Friday show: Chamath and Sacks both bailed 90 minutes before record, so it is Jason and Friedberg with Gavin Baker (Atreides), Ben Shapiro and energy investor Phil Deutch (NGP). The GPT-5 launch is the news but Gavin refuses the diminishing-returns framing - he calls it the first time OpenAI shipped a model that was not decisively best, blames talent loss and benchmark saturation rather than scaling, and stays optimistic on Grok 5 off the largest Blackwell cluster (disclosing Atreides is an xAI shareholder); Friedberg reads GPT-5's real upgrade as the model router, i.e. UX not capability. On AI's market size Ben asks the bear question directly - do the productivity gains arrive fast enough to justify the spend, or do you get a crater - and Gavin answers it with receipts: ROIC has risen since the AI capex started, the dark-fiber analogy fails because GPUs get used so hard they melt, and we are still behind demand. China splits into two claims: Friedberg concedes their manufacturing/materials/supply-chain edge but says the CCP's throttling of entrepreneurship rate-limits them, Gavin counters that Xi is throttling back up, Phil says none of it lasts without rule of law (the wind-blade factory and 'controlling the chop'). Energy is the deepest segment - Phil vs Friedberg on whether solar/wind subsidies starved nuclear, Phil's OBBB chart showing generation stepping backwards, and Anthropic's 50GW-in-three-years ask - and the socialism debate is the real centerpiece: Friedberg re-ups his own prediction now aimed at America ('socialism will sweep over this nation'), Ben supplies the trigger (a downturn turns grievance into 2008-style redistributionism further left than Obama) and the NYC capital-flight receipt, while Gavin and Phil take the other side on fast visible failure and national resilience. Not captured as ideas: Gavin's 'best AR glasses in three or four years' Apple call (no clean price-resolvable expression), Phil's 2030 EV-battery-aggregation thesis (beyond the 36-month cap), and Friedberg's 'nobody wins the internet' anti-winner-take-all AI framing (no matching open idea).