Mark Sellers

Hedge fund manager. I buy stocks other people hate, then suffer for longer than I expected to. Usually I get it right.

[Mark Sellers @TheloniusM23]

@GaryMarcus Anyway here’s a current research paper released 3 days ago, with same basic conclusion:

https://t.co/WXSkGYhuAb

[Mark Sellers @TheloniusM23]

@GaryMarcus The amount of copium by the ai bull bois in response to this kind of evidence contrary to their worldview is astounding. Predictable, yes, but still astounding.

[Mark Sellers @TheloniusM23]

@Ross__Hendricks @Austen Crickets

[Gary Marcus @GaryMarcus]

The “AGI-is-near” community keeps committing the same logical fallacy over and over; I have seen it at least half a dozen times today alone. Every time there’s an advance, I see the same error.

Here’s how the fallacy works.

1. Someone pretends that all cognition is created equally. (Totally untrue.)

2. Whenever AI achieves success on some form of (fancy) cognition they want you to believe that success on all forms of AI is imminent.

I include three prominent examples from today below, each promising that some grand universal can be solved (“science”, “every discipline”, “every problem [people] face in life”), simply because there was an advance in a particular domain (today, math, other times coding, etc).

You don’t have to be a cognitive psychologist to realize that this inference just doesn’t follow. We all know that expertise in math doesn’t guarantee genius in all domains.

We all know, for example, people who are great at math or physics or programming but struggle with writing or understanding human relationships (and conversely great writers who are weak at math), etc.

It’s common to be good at some forms of cognition and not others. Expertise in one domain does not at all guarantee expertise in all or even most domains.

That’s precisely *why* people like Gardner and Sternberg developed multidimensional theories of intelligence, why the SAT tests math separately from verbal, etc.

Astra is (apparently - we still haven’t seen the methodology) great at math, or at least some forms of math, but that does not mean that it will avoid hallucinations or solve the reliability problems other GenAI systems have. It doesn’t even mean it will be able to read a PDF properly. It doesn’t mean it will be able to follow hard rules either. (Which should terrify you.)

In fact, the performance we saw today doesn’t even mean it can write a decent math proof; to the contrary, the mathematician @henryquantum has already given an example where clarity in the proof was lacking.

Whatever you do, please don’t get suckered into the “all cognition is alike” fallacy. It’s not all alike.

Astra is obviously good at some problems; but that doesn’t mean it will be good at problems that are hard to formalize. It doesn’t mean it will be magic. It doesn’t mean it’s AGI or ASI or any of that.

It’s *very* impressive. But I see no reason whatsoever to think it is AGI let alone ASI. If it can score even a 5/10 on my 2024 bet with Miles Brundage I will be surprised.

Enough with this fallacy.

[GURGAVIN @gurgavin]

ORACLE IS NOW SPENDING 100% OF IT’S ENTIRE REVENUE ON CAPEX

AM I THE ONLY ONE WHO THINKS ORACLE HAS GONE CRAZY

$ORCL

[Mark Sellers @TheloniusM23]

Me at the McDonald’s drive through:

MCD: Will you be using the mobile app today?

Me: No, but if you invest $6 to buy a 0.1% ownership interest in my car, I’ll use that money to order a Big Mac for $5.99.

MCD (checks with manager): Uh….yeah, we’ve got a deal. Please pull forward.

Sounds ridiculous, but this is literally what the Big 3 hyperscalers are doing with respect to OpenAI and Anthropic. NVIDIA doing it too all over the place. It’s nuts.

[Mark Sellers @TheloniusM23]

@Dr_Gingerballs How can people not see it’s just revenue they’re purchasing by “investing” (doesn’t count as an expense) in Anthropic and then Anthropic uses that $ to buy cloud compute? I can’t understand how people aren’t penalizing them for this. It’s insane.

[Mark Sellers @TheloniusM23]

@Ross__Hendricks I think more and more people are starting to see it. They’re even allowing Zitron on Bloomberg and CNBC lately. Surely *some* of the viewers must be willing to listen…maybe?

[Mark Sellers @TheloniusM23]

The Big 3 hyperscalers have done multiple vendor financing deals with just two “customers.” And their stock prices go UP after announcing it. How do people not see how this is bad? It’s an accounting loophole! It’s fake revenue. The cloud backlog at these Big 3 scalers isn’t real

[Mark Sellers @TheloniusM23]

Amazon uses $50B of its cash to buy stock in Anthropic (which Amazon books as an “equity investment nudge nudge wink wink” rather than an expense), and Anthropic then turns around and uses that same $50B to buy cloud compute from Amazon. (1/2)

[Mark Sellers @TheloniusM23]

@patrick_oshag “Warren” is still alive and may be for several more years. So was Munger predicting an early demise for his friend Warren? Or is this a made up quote? Has to be one or the other I think?

[Mark Sellers @TheloniusM23]

@GaryMarcus I’d give it 60/40 odds on Son/Ellison. I agree Son is the most likely but Ellison is not far behind. Both are leveraged to the absolute hilt on a bad bet.

[Ashpreet Bedi @ashpreetbedi]

Everyone knows I'm super AI-pilled. But one trend I'm noticing as I talk to more and more companies: the personal productivity gains are not translating into organizational growth and efficiency as expected.

It's an odd dichotomy. Individuals are more productive. Engineers are writing an insane amount of code. But the gains aren't showing up in the numbers yet.

Not sure how universal this is.

[Mark Sellers @TheloniusM23]

@Ross__Hendricks Though I do agree every money manager should internalize this paper and the book written about it, and should think about it in theory all the time

[Mark Sellers @TheloniusM23]

@Ross__Hendricks If you run the Kelly Formula and have a confidence level of, say, 70% that a stock will be a winner, the formula will tell you to bet several hundred % on that bet. It was developed for Blackjack, not the stock market. It also doesn’t take multi-bet correlation into account.

[Mark Sellers @TheloniusM23]

@edzitron When he asked about whether the societal productivity gains justify the capex, you might have answered “where’s the productivity bump from Ai? None of it is showing up yet in the economic numbers. Why not?”

[Mark Sellers @TheloniusM23]

@GaryMarcus Two young CEOs getting far too high on their own supply.

[zerohedge @zerohedge]

The SPV bubble is batshit insane

GOOGL is $826BN and now Meta said it has already committed almost $700 billion in future AI spending (BBG).

Across all hypers, off BS commitments are now $3+ trillion, double in one quarter.
GLTA