Speaking of which, I noticed that the big market reversal during the beginning of the Iran war happened right around when ESLR requirements were due to relax. Is this a transmission mechanism for that? Did some of the big brokerages run a big promotion (0% APR on margin debt!) or something?
That chart (and many from FRED) is unfortunately not useful without viewing it on a logarithmic scale. Things can look quite parabolic in a linear space, but exponential growth is quite normal in many financial contexts.
Click "edit graph" and change units to "natural log". Now look again. You'll see that the growth in margin loans is absolutely normal, and actually has only recently recovered from the dip caused by the 2008 financial crisis.
Great tip, but I didn't see "natural log" specifically. Perhaps "Compounded rate of change" is most applicable? That's still mostly above 0 historically, indicating margin usage is ever-increasing. The helpful graph would be margin usage as a weighted percentage of market participation.
No, but it is about the, very expected, shitshow caused by the AI hype. When companies think cramming AI in any random place is a magic bullet for replacing workers, shit breaks. Because AI is dumb af. LLMs are just the shittiest of the ways to do it.
Do you honestly think this particular "cram AI in everything" isn't related to the current AI hype? Or that AI applications and companies providing things like this won't crash right along with the general llm AI hype and leveraged investments?
Nah, the kind of AI they were using has been around for a long time (AlexNet is 13 years old), and before that they would have used classical computer vision.
The only thing that happened here is that they didn't check it worked before firing everyone. That can happen with any automated system, AI or not.
I am referring to the hype causing it to be adopted now instead of a decade ago, not the specific implementation. The hype causing people to adopt AI of all kinds hastily without proper validation.
Oh, yeah, that's cooking lol: https://fred.stlouisfed.org/series/BOGZ1FL663067003Q
Speaking of which, I noticed that the big market reversal during the beginning of the Iran war happened right around when ESLR requirements were due to relax. Is this a transmission mechanism for that? Did some of the big brokerages run a big promotion (0% APR on margin debt!) or something?
That chart (and many from FRED) is unfortunately not useful without viewing it on a logarithmic scale. Things can look quite parabolic in a linear space, but exponential growth is quite normal in many financial contexts.
Click "edit graph" and change units to "natural log". Now look again. You'll see that the growth in margin loans is absolutely normal, and actually has only recently recovered from the dip caused by the 2008 financial crisis.
> change units to "natural log"
Great tip, but I didn't see "natural log" specifically. Perhaps "Compounded rate of change" is most applicable? That's still mostly above 0 historically, indicating margin usage is ever-increasing. The helpful graph would be margin usage as a weighted percentage of market participation.
Natural log is in there - I just tried it myself. It's the second selection from the bottom in the "Units" dropdown.
$600+ billion doesn't seem like much. I was expecting more tbh.
Margin debt (debt used to leverage stock buying) Is near record highs, and most of it is flowing into AI stocks ...
Ruh roh. https://www.bbc.com/news/articles/cgrkd41n2v9o
That's not about LLMs.
No, but it is about the, very expected, shitshow caused by the AI hype. When companies think cramming AI in any random place is a magic bullet for replacing workers, shit breaks. Because AI is dumb af. LLMs are just the shittiest of the ways to do it.
Do you honestly think this particular "cram AI in everything" isn't related to the current AI hype? Or that AI applications and companies providing things like this won't crash right along with the general llm AI hype and leveraged investments?
Nah, the kind of AI they were using has been around for a long time (AlexNet is 13 years old), and before that they would have used classical computer vision.
The only thing that happened here is that they didn't check it worked before firing everyone. That can happen with any automated system, AI or not.
I am referring to the hype causing it to be adopted now instead of a decade ago, not the specific implementation. The hype causing people to adopt AI of all kinds hastily without proper validation.