‘You think you know everything!” My late mother used to say that to me, back when I was a 12-year old smart aleck.
I didn’t know everything. No one does. But I could mask it, even then, with confidence. Intuition, what the English call a “brain wave,” can substitute for knowledge. You don’t get serious advances without intuition.
Which brings me to Artificial Intelligence and Large Language Models. LLMs know everything, but in many ways, they know nothing. Their neural networks lack intuition. There are no cognitive leaps.
This doesn’t make them useless. It’s good to have someone next to you with all the background you need, when you are trying to come to a decision. It’s what secretaries and aides have been doing since bureaucracy began.
But it’s not intelligence.
Intelligence is the ability to learn, combined with the ability and willingness to make an educated guess. If that weren’t the case, we would be stuck with obvious conclusions drawn only from the evidence at hand.
OpenAI and the Tech Bros pushing it are lying to us. They’re grifting. They don’t have the solution they claim they have, and I don’t think they’re going to get it using the current approach.
We need new algorithms, not just new software but new approaches to using software and data.
The Coming Crash
There are valuable markets for digital assistants. There will also be great markets for cloud gaming, once all those Nvidia processors become available for it.
But that doesn’t mean there won’t be a crash before then. The difference between what’s real and what’s hype must be adjusted. That’s what markets are for.
Theranos claimed they had a testing system that would work with a single drop of blood. They didn’t. They were working on one. That’s why Elizabeth Holmes is in jail right now.
The difference between Holmes and Sam Altman is that there’s no regulatory authority that can call Altman on his bullshit. There’s only the market, which as tech bros like Chamath Palihapitiya proved years ago, is open to manipulation by confidence men. (Maybe we’ll know AI is real when one starts a SPAC.)
The timing of the crash will be revealed by numbers. The Cloud Czars’ failures to (so far) generate revenue from AI are masked by their growth in advertising. But at some point, the gap between their capex expenses and their expected earnings will become obvious.
I’m thinking that will happen when third quarter earnings come out, in late October. But I could be wrong.
I could be bullshitting.
Or I could be making an intuitive leap. Just because I don’t know everything doesn’t mean I’m not intelligent.
