Can you comment on the possibility that the frequency of black swan events is related to algo (increasingly influenced by AI) trading? In 1987 (grad school...yes I am old), someone I can't remember the name of came to Stanford to give a seminar IN THE ENGINEERING DEPARTMENT on short term forecasting of the stock market (eigen-modes and such) and how to exploit 8 minutes of future performance to make a pile of money. The discussion period was fixated on the notion of this question: What fraction of the trading following such a strategy would lead to feedback (stabilizing or destabilizing) that could then "drive" the market?
I don't know about you but wherever I click these days I get about 50% ads to hook retail investors into AI guided training. You can say I must have clicked one of them to get that frequency but in fact I have skipped them all. Given that frequency, I have to figure they ARE getting "takers" for this (ads cost money) and this makes me wonder how much of the story about retail investors supporting the bubble (told as a confidence measure when all the studies say otherwise) while smart money has pulled back is driven by AI guided "algo" trading? Of course AI is "learning" its own feedback but, if its objective function is to capture short term gains without a significant penalty (like ALL investors have) for sharp downturns this seems like it has to end badly.
So, I will ask some curious questions (that are 38 years old...probably older than half of the retail market):
1) what fraction of retail trading do you think is currently AI guided? (is there any data on it)
2) what fraction do you think is the "tipping" point- where investing on principles and reasoning becomes meaningless because all that matters is what AI thinks of the signals?
Put another way, you mention watching for these 3 and 4 sigma events that are expected every 63 years but already they are a couple times a year and, if my concern is valid, we should expect them to be more and more frequent to the point that, if nothing is done to de-incentivize this (some sort of trade friction...seems what we have is insufficient), we are at the mercy of watching the markets either self destruct (crash or other) or just become a game of trying to catch black swans (at which point that name is meaningless if not already).
I follow your linking the big events to liquidity reactions to credit markets. My question was more for the rapid trades on movement over minutes that are not due to credit markets on that time scale.
Thanks. I am guilty of going off-topic and mixing up the issues. Do you know the best substack or other venue to discuss AI-guided training's role in sustaining the bubbles?
Googling "AI-guided stock trading influence on the bubble" provides an interesting response. The search (in CHROME) provides lots of SPONSORED products to DO that kind of trading but few articles (even after skipping past the sponsored stuff) that discuss its influence on the market. I say interesting because the AI Overview (Gemini) prepares a better version of my concerns at the top of this thread. Whether that is reasoning or sycophant is not clear to me which is why I am asking if you or others know where humans are talking about this. I suspect Gemini found it to "build" this response for me but search is not very effective in showing articles or venues. Yes there are "links" to its points but mousing over those AGAIN leads more to products to robo trade than to discussion. Click money.
After doing some readig, I suppose I should not be surprised that what Claude found was for institutional algo trading. I have some concern about that but, in the past few months, I am more concerned about a huge number of smaller/newbie investors drawn into AI-guided or flat out AI-Robo trading. From what I have read so far, this is too new to have been studied and likely sufficiently different as to have new and poorly understood pathologies.
Again, asking Claude:
Are there any studies of the systemic market risk from the Gen Z use of AI-guided, or AI-only stock picking, or fully automated AI trading systems?
mumbles around a while before closing with:
However, there appears to be no specific research examining:
Whether Gen Z's concentrated AI usage creates unique systemic risks
How Gen Z's preference for similar AI platforms might contribute to market correlation
The systemic impact of a demographically concentrated group using similar AI trading strategies
Why This Research Gap Matters
Given that Gen Z shows both high AI adoption (41%) and frequent trading behavior, their concentrated use of similar AI tools could theoretically create new forms of systemic risk through:
Coordinated AI-driven trading decisions
Similar algorithmic responses to market events
Potential for rapid, synchronized market movements
This represents a significant gap in current financial stability research, as regulators and academics have not yet specifically studied whether generational clustering around AI trading tools creates unique systemic vulnerabilities.
END of Claude
Again, this is either sycophant or authentic reasoning that I am asking an important/valid question. In my experience, human conversations call out BS when they see it with a higher frequency than AI does. Maybe there is a product opportunity. NoBSAI.
Drat. I had hoped NOT to interrupt your day at the ballpark with your kiddo. I already had a Claude account and this prompt:
Are there academic or other studies of the systemic risks of ai-guided stock trading's influence on the stock market?
Produced some stuff. I will dig into the rabbit hole a bit but I tend to like real people discussions (or papers) over AI summaries that I worry are trying too hard to please me by inferring the answers I am looking for from the way I ask the question. I am not saying it is not valuable but trusting an AI summary will only come after I can learn and verify from human sources.
Claude's summary at the end said:
The research consistently shows that while AI improves market efficiency, it introduces new forms of systemic risk that require careful monitoring and regulation.
I used that to form a new question:
What monitoring and regulation is in place to insure that the new forms of systemic risk from AI-guided trading do not distort or shock the stock market?
Summary of response:
While foundational monitoring exists, regulators are still developing AI-specific frameworks to address the novel systemic risks posed by intelligent, adaptive trading systems that can exhibit emergent behaviors beyond traditional algorithmic trading.
Of course there is a lot more detail but my summary is that there were efforts in the SEC in December of 2024 but I think we can assume that much of what the SEC was doing was shut down soon after. While I totally support deregulation that was holding back growth, this is one area where cutting regulations might be a bit dangerous. I think it is not too hard to imagine a mob with pitchforks and torches if a crash has a credible smoking gun on AI-guided trading with or without a link to the SECs (in) action. You probably know that there was steady decline in % stock ownership from 2008-2020 that is only now back up in the >60% of population. I openly worry that Gen Z represents a lot of that growth and they have no memory of 2008, thinking every crash is a dip to buy like the Covid and 2022 and we will be back to new highs within a year.
It’s pretty clear the AI-guided trading apps are targeting Gen Z FOMO and Boomers who fear they have far less than they need to avoid retirement poverty and are hoping for a few 1000 percent stock picks to rescue them. Asking Claud about that
Are there any studies that show the fraction of retail stock traders that are using ai-guided or AI-only stock trading?
Summary:
What's Missing
There appears to be a lack of comprehensive studies specifically measuring what fraction of retail traders are using:
Fully automated AI trading systems
AI-guided day trading platforms
AI-only stock picking services
The 20% figure for AI research tool usage and 28% for robo-advisor preference are the closest available metrics, but they don't capture pure AI-guided active trading by retail investors.
END OF Claud
I think I am convinced that 28% (which is not quite clear) can significantly alter the market and it is just a matter of time before smart money pulls the rug—Buffet already has mostly though not explicitly stating for this reason. This plus the consumer debt (which might also be influenced by AI-guided or AI-only retail trading ..is it trading or gambling at that point?) means tick, tick, tick…..
Well, with that happiness, I'll have to add my two favorite baseball players, the problem is they're both tied for first!
I'm from Houston so I'll say Nolan Ryan first and then mention my other #1 is (drum roll please) Cal Ripken Jr. Bless them both for longevity! Can you believe 2,632 straight games, all with Baltimore?
Thanks Garrett. Great stuff.
Can you comment on the possibility that the frequency of black swan events is related to algo (increasingly influenced by AI) trading? In 1987 (grad school...yes I am old), someone I can't remember the name of came to Stanford to give a seminar IN THE ENGINEERING DEPARTMENT on short term forecasting of the stock market (eigen-modes and such) and how to exploit 8 minutes of future performance to make a pile of money. The discussion period was fixated on the notion of this question: What fraction of the trading following such a strategy would lead to feedback (stabilizing or destabilizing) that could then "drive" the market?
I don't know about you but wherever I click these days I get about 50% ads to hook retail investors into AI guided training. You can say I must have clicked one of them to get that frequency but in fact I have skipped them all. Given that frequency, I have to figure they ARE getting "takers" for this (ads cost money) and this makes me wonder how much of the story about retail investors supporting the bubble (told as a confidence measure when all the studies say otherwise) while smart money has pulled back is driven by AI guided "algo" trading? Of course AI is "learning" its own feedback but, if its objective function is to capture short term gains without a significant penalty (like ALL investors have) for sharp downturns this seems like it has to end badly.
So, I will ask some curious questions (that are 38 years old...probably older than half of the retail market):
1) what fraction of retail trading do you think is currently AI guided? (is there any data on it)
2) what fraction do you think is the "tipping" point- where investing on principles and reasoning becomes meaningless because all that matters is what AI thinks of the signals?
Put another way, you mention watching for these 3 and 4 sigma events that are expected every 63 years but already they are a couple times a year and, if my concern is valid, we should expect them to be more and more frequent to the point that, if nothing is done to de-incentivize this (some sort of trade friction...seems what we have is insufficient), we are at the mercy of watching the markets either self destruct (crash or other) or just become a game of trying to catch black swans (at which point that name is meaningless if not already).
Liquidity driven… freezing in credit markets…
I follow your linking the big events to liquidity reactions to credit markets. My question was more for the rapid trades on movement over minutes that are not due to credit markets on that time scale.
Just pay attention to the 20 day daily ema on the FNGB. That’s your signal something is wrong.
It’s always linked back to the bond market.
The 20 day is where algos start to misbehave. So yea to your question.
Thanks. I am guilty of going off-topic and mixing up the issues. Do you know the best substack or other venue to discuss AI-guided training's role in sustaining the bubbles?
Googling "AI-guided stock trading influence on the bubble" provides an interesting response. The search (in CHROME) provides lots of SPONSORED products to DO that kind of trading but few articles (even after skipping past the sponsored stuff) that discuss its influence on the market. I say interesting because the AI Overview (Gemini) prepares a better version of my concerns at the top of this thread. Whether that is reasoning or sycophant is not clear to me which is why I am asking if you or others know where humans are talking about this. I suspect Gemini found it to "build" this response for me but search is not very effective in showing articles or venues. Yes there are "links" to its points but mousing over those AGAIN leads more to products to robo trade than to discussion. Click money.
I’m at a ball game and I’ll try to remember…
But I would look at academic studies on the subject. Ironically… ask ChatGPT for the answer. I’ll do that later… but get the academic studies.
They are out there
After doing some readig, I suppose I should not be surprised that what Claude found was for institutional algo trading. I have some concern about that but, in the past few months, I am more concerned about a huge number of smaller/newbie investors drawn into AI-guided or flat out AI-Robo trading. From what I have read so far, this is too new to have been studied and likely sufficiently different as to have new and poorly understood pathologies.
Again, asking Claude:
Are there any studies of the systemic market risk from the Gen Z use of AI-guided, or AI-only stock picking, or fully automated AI trading systems?
mumbles around a while before closing with:
However, there appears to be no specific research examining:
Whether Gen Z's concentrated AI usage creates unique systemic risks
How Gen Z's preference for similar AI platforms might contribute to market correlation
The systemic impact of a demographically concentrated group using similar AI trading strategies
Why This Research Gap Matters
Given that Gen Z shows both high AI adoption (41%) and frequent trading behavior, their concentrated use of similar AI tools could theoretically create new forms of systemic risk through:
Coordinated AI-driven trading decisions
Similar algorithmic responses to market events
Potential for rapid, synchronized market movements
This represents a significant gap in current financial stability research, as regulators and academics have not yet specifically studied whether generational clustering around AI trading tools creates unique systemic vulnerabilities.
END of Claude
Again, this is either sycophant or authentic reasoning that I am asking an important/valid question. In my experience, human conversations call out BS when they see it with a higher frequency than AI does. Maybe there is a product opportunity. NoBSAI.
Drat. I had hoped NOT to interrupt your day at the ballpark with your kiddo. I already had a Claude account and this prompt:
Are there academic or other studies of the systemic risks of ai-guided stock trading's influence on the stock market?
Produced some stuff. I will dig into the rabbit hole a bit but I tend to like real people discussions (or papers) over AI summaries that I worry are trying too hard to please me by inferring the answers I am looking for from the way I ask the question. I am not saying it is not valuable but trusting an AI summary will only come after I can learn and verify from human sources.
Claude's summary at the end said:
The research consistently shows that while AI improves market efficiency, it introduces new forms of systemic risk that require careful monitoring and regulation.
I used that to form a new question:
What monitoring and regulation is in place to insure that the new forms of systemic risk from AI-guided trading do not distort or shock the stock market?
Summary of response:
While foundational monitoring exists, regulators are still developing AI-specific frameworks to address the novel systemic risks posed by intelligent, adaptive trading systems that can exhibit emergent behaviors beyond traditional algorithmic trading.
Of course there is a lot more detail but my summary is that there were efforts in the SEC in December of 2024 but I think we can assume that much of what the SEC was doing was shut down soon after. While I totally support deregulation that was holding back growth, this is one area where cutting regulations might be a bit dangerous. I think it is not too hard to imagine a mob with pitchforks and torches if a crash has a credible smoking gun on AI-guided trading with or without a link to the SECs (in) action. You probably know that there was steady decline in % stock ownership from 2008-2020 that is only now back up in the >60% of population. I openly worry that Gen Z represents a lot of that growth and they have no memory of 2008, thinking every crash is a dip to buy like the Covid and 2022 and we will be back to new highs within a year.
It’s pretty clear the AI-guided trading apps are targeting Gen Z FOMO and Boomers who fear they have far less than they need to avoid retirement poverty and are hoping for a few 1000 percent stock picks to rescue them. Asking Claud about that
Are there any studies that show the fraction of retail stock traders that are using ai-guided or AI-only stock trading?
Summary:
What's Missing
There appears to be a lack of comprehensive studies specifically measuring what fraction of retail traders are using:
Fully automated AI trading systems
AI-guided day trading platforms
AI-only stock picking services
The 20% figure for AI research tool usage and 28% for robo-advisor preference are the closest available metrics, but they don't capture pure AI-guided active trading by retail investors.
END OF Claud
I think I am convinced that 28% (which is not quite clear) can significantly alter the market and it is just a matter of time before smart money pulls the rug—Buffet already has mostly though not explicitly stating for this reason. This plus the consumer debt (which might also be influenced by AI-guided or AI-only retail trading ..is it trading or gambling at that point?) means tick, tick, tick…..
Well, with that happiness, I'll have to add my two favorite baseball players, the problem is they're both tied for first!
I'm from Houston so I'll say Nolan Ryan first and then mention my other #1 is (drum roll please) Cal Ripken Jr. Bless them both for longevity! Can you believe 2,632 straight games, all with Baltimore?
Bonds.
I have a great collection of Nolan Ryan baseball cards. Including his real first appearance on a card in a Mets team picture card. Love that guy.
That's a keeper!
What to do?