AI Stock Investing Update #3 (Almost Done)

Off the top, no danger of AI catching up to the S&P 500 in an up market, as I had hoped at the end of my last update. Since then, the S&P is up 9% and AI is only up 7.5%. But wait, a glimmer of light!

On November 3rd, I got my sell signal, for SWAV. I saw that despite it being down from when I bought it, it had risen substantially in the past couple of days. Then I noticed that there was an earnings call scheduled for the next day, and I recalled that often stocks rise going into an earnings call, then fall after the earnings are released. Something about the triumph of hope over experience.

To my pleasant surprise, the AI was right. The day after the earnings call the stock plunged steeply, and I had already sold.

That marked twice in a row that the AI warned me before a steep plunge in the stock price. Both times the overall result was poor (I lost money), but would’ve been much worse off without the warning.

Note that I sold because of my own interpretation of a weak buy from the AI (two days in a row of 7 out of 10), not because the AI actually told me to sell. Here are Midjourney’s best four responses to my prompt: Businessman giving a thumbs down in the NASDAQ stock market --ar 16:9

This seems eerily similar to most buy ratings from stock market pundits, including AI. ‘Buy’ means sell depending on the way it’s said. Notice how the first guy can’t make eye contact? Exactly.

But still, a little bit of advance intel on a stock plunge could be worthwhile despite the mediocrity of AI’s overall stock picking. Maybe even worth $20 a month?

Then two weeks later I got the sell signal on SE, the one stock in my AI portfolio that had been doing ok. Checking the stock price, I groaned, as the stock had already fallen 20%, again after an earnings call, but this time no advance warning from the AI. Fuck.

At the same time I got sell signals on PODD and CHWY, and saw that both had plunged after earnings calls, again with no warning from the AI. That last glimmer of usefulness was thoroughly snuffed.

I’m going to continue the experiment to its bitter three month end on December 12, 2023, in the hopes of further learning, but as I sell stocks now I’m not replacing them with new suggestions from the AI. Five times bitten, sixth time shy.

Here’s the early lesson though, stay away from AI, at least in an endeavour as complex as stock picking.

1. Original post on AI stock investing (includes links to earlier updates)