AI Busted
Are we in the mist of an AI Bubble thats about to BURST?
All we hear about these days is how AI is going to change everything—transform the world, revolutionize jobs. And guess what? That change is happening right now…
I mean these are just some of the headlines we are being inundated with:
Where exactly is this change happening ?
The main claims focus on white-collar jobs being automated. Companies are attributing layoffs to AI, or at least that's what the headlines suggest.
But even with these claims, earnings calls and public statements made by companies indicate that these layoffs aren’t due to AI. For example, “A UPS spokesperson later said AI is not replacing workers, and that executives did not make an explicit connection between AI and the permanent layoffs on the company’s earnings call.” (Above Link)
So we have a huge wave of white collar layoffs happening and its all due to AI, yet its not due to AI.
I mean look at this graph. ChatGPT release date wasn’t even released until end of 2022. Yet software jobs had already peaked and returned to baseline. So ChatGPT / LLMs couldn’t have caused all of these layoffs? That also only leaves ~1year for companies to actually implement AI. Not many companies have the sophistication to do so.
We're in a situation similar to the dotcom bust of 2000. The internet is definitely transformative, but real change takes time—it doesn't happen overnight. The internet didn't really take off until smartphones became affordable and people could access it from anywhere. I mean pets.com reached $400 million valuation while selling products at about ~60% loss. The only reason pets.com reached this valuation was the internet trend stick that was slapped onto it. Exactly what we are seeing with the AI bumper sticker companies are slapping onto their company.
Look at the # companies saying AI in earning calls. Just following the herd and using buzzwords to generate stock interest. (Source)
Right now, most new AI advancements are just companies wrapping features that call upon large language models (LLMs), usually from OpenAI.
SideNote: Other forms of AI have already existed for years, so why weren’t predictive models laying off people before?
Current LLM Issues
Chevy's AI customer service messed up big time by selling a $76,000 item for just $1. (Source) This slip-up shows just how easy it is to trick our current AI systems. While this might seem like a harmless mistake, it actually points to a bigger issue: when these large language models (LLMs) have access to sensitive information, they can be easily fooled into giving it away.
Think of an LLM like a 4-year-old with a photographic memory of every piece of text ever written. This kid knows tons of words and can put them together in a way that sounds right. But they don't have the critical thinking skills to make good decisions. Even though they can talk the talk, they don't really understand the deeper meaning behind what they're saying.
I mean look at this Bloomberg Automation (AI) from 6/3/24. Yes, there was technically a drop in price due to a technical issue, but this is something a human writer would easily catch. The price dropped by over $600k but the AI has no reference of why that is crazy. It lacks the context and experience and the ability to connect with all of the data it has trained on. Source
Here is the NYSE technical issue that lead to many stocks literally dropping to worthless. $GOLD, $BMO, $BRK.A (Berkshire) all experienced drops of 99%.
Unexpected GME Connection
That leads to another question. Why is it that when all of the glitches happen, GME (Gamestop) seems to be at the center of the discussion? Shouldn’t GME have 0 correlation with the broader stock market and how exchanges are operating??
It has long been a theory among "APEs" that when the stock market crashes, GameStop will shoot up in price. This theory is supported by GME's beta, which is around -0.3. A negative beta means that when the market goes up, the stock tends to go down, and vice versa. See the picture version below.
When we dig deeper into this situation, there's a lot of unusual swap activity involving GME that seems to be at play. A swap is a type of derivative contract. One example is a Portfolio Swap, where an investor pays someone to manage a basket of stocks or assets. The manager then passes any gains or losses along to the investor. So, while I don't own the stocks, I still get the profits or losses from them.
Firms that blew up during the 2021 GME rise, like Archegos, had massive swap positions that went the wrong way. Another problem with swaps is their ability to hide your position and leverage. Archegos could go to multiple banks and open these swap contracts, continually leveraging itself without the other banks knowing about the counterparty risk. This commented document on the SEC website does a pretty good job of detailing this issue. (Source)
Maybe this is just a technical issue, but it seems GME has a connection to these issue. These hidden swaps/derivative contracts could play a massive role in the market moving forward.
Note: This was meant to be an article just on AI and how overblown it has become.
Thanks For Reading!
To Read Past Newsletters Click here.
If you have any source you wish to see in our lineup, please reach out and let us know. We continually look to incorporate more independent sources to our weekly wrap-up.