E133: Market melt-up, IPO update, AI startups overheat, Reddit revolts & more with Brad Gerstner

Nvidia will continue to perform strongly and will beat its financial guidance/consensus estimates for all remaining quarters of 2023 ("the balance of the year").

“You know, I think a lot of the AI related stocks are at or within 10% of our end of year price target. And when you when you see that, take something like Nvidia, if I'm playing at home or I'm a, you know, a professional investor, I may say, oh, I'm going to go sell some long dated calls to buy a little protection to the downside. I don't think there's any problems with Nvidia. I think they're going to continue to perform. I think their beat their numbers for the balance of the year.” View on YouTube

From mid‑2023 forward, interest rates will remain elevated (above the level investors "want") and stay high for an extended period rather than being cut soon, and the market bottoming process is nearly complete by mid‑2023.

“What have I said. Like a broken record. Rates are going to be higher than you want and they're going to be around for longer than you like. And now Powell is basically telling you the same thing. So. We're almost at the end of I think the bottoming though I don't agree with Druckenmiller I think he's wrong.” View on YouTube

The U.S. (and global) economy will not experience a "hard landing" recession in Q4 2023 because China will undertake trillions of dollars of stimulus, preventing such an outcome.

“And the reason there's not going to be a hard landing is you just saw China today basically say we're going to start to rip in trillions of dollars. They're going to stimulate the economy. You can't have a hard landing when China's printing trillions of dollars. It's not possible.” View on YouTube

For the remainder of the 2020s (through about 2030), both interest rates and inflation will remain "sticky" at relatively elevated levels, rather than returning quickly to the near‑zero‑rate, very‑low‑inflation environment of the 2010s.

“And so you just have to get prepared for rates just being sticky and inflation being sticky. And I think that that's probably the most reasonable base case for the rest of the decade.” View on YouTube

Over roughly the next 5–7 years from 2023, the bulk of web search "top of funnel" advertising and related economic value (currently dominated by traditional search) will be redistributed among new AI‑driven interfaces and competitors, rather than remaining concentrated in the existing ten‑blue‑links search model.

“And I will tell you that the conversation around that table, there's increasing confidence that whatever comes next, right, the top of the funnel is up for grabs. And all of those dollars over a period of five years, seven years is going to get redistributed.” View on YouTube

Within roughly six months from June 2023, Google will deploy AI capabilities that surpass ChatGPT‑4 in practice, and once integrated into its search/ad products this will increase advertisers' cost per click (CPC) due to improved targeting and user understanding.

“I think Google's not as far behind closed AI as we think... I think Google's going to be ChatGPT four. I'm saying it right here, right now. I think they're going to beat them, because I think that they're better at indexing all this information and understanding it than anybody on the planet, and they have the largest ad network. If they get this done in the next six months, I think it's going to increase the cost per click because they're going to know so much about each user.” View on YouTube

Within one week of June 16, 2023, there will be at least two AI-related startup funding rounds announced that are larger than Mistral AI's $105 million seed round, and these will be publicly reported as significantly larger financings ("monster rounds").

“So again, they're going to be a couple of monster rounds, I think announced next week like that are going to make this one look like kids play. Um, so a lot more of this is coming.” View on YouTube

By roughly December 2024 (18 months after June 2023), the compute cost required to train a large language model equivalent in capability to GPT-4 will fall from roughly $400 million to approximately $5–10 million, implying about a 40–80x reduction in training cost.

“So something that costs like we said, OpenAI spent $400 million training models for for GPT four. If they spent $400 million in the last couple of years, you could probably assume that doing the same training exercise could be done for 5 to $10 million. 18 months from now to generate the same model. That's a, you know, 100 x cost reduction.” View on YouTube

Between roughly December 2024 and June 2026 (18–36 months after June 2023), the cost of a given large AI model training run that would cost $100 million in mid-2023 will decline to about $1 million, i.e., around a 100x reduction in training cost over that period.

“Play this out. $100 million of capital spent training today is $1 million spent doing training in 18 months. Yeah, three years, 1818 to 36 months. Somewhere in that time frame is likely the time frame.” View on YouTube

At some point after this episode (June 2023), the All-In Podcast will host Mayor Francis Suarez as a guest on the show.

“and Frances. Mayor Frances is now in the race, so I guess we'll have him on the pod.” View on YouTube