So ChatGPT completely changed the investing game and not too many people are catching on quite yet.

INTERESTED IN AI? – Check out our list of the best AI newsletters here

I’ll give you one example.

There is a channel on YouTube called Dumb Money which talks about stock trading.

Chris Camillo, one of the hosts, turned $20,000 into $10 million without any special software or formal education.

How do you ask?

He used networks like Twitter and what people are talking about on it, to gain an informational edge on the big investors on wall street.

What I had been doing was essentially looking at Twitter, looking for keywords on Twitter that were trending, um, and trying to. Understand changes in conversational volume around granular topics that were meaningful to publicly traded companies.

But it was a really difficult thing to do, just eyeballing it. So I was like, my dream, my dream as a social urban investor would be, To have a software platform that would ingest the entire Twitter fire hose, changes in conversational volume year over year, quarter over quarter, like in the two weeks prior to every iPhone release.

How many people are speaking about getting a new iPhone versus the last year and the year before that, and the year before that. But what I realized during the course of my investment,  is that when done crack, nothing else matters. Besides, besides the observational piece, the, so the arming of information, the arming of social information, I don’t look at stock price at all.

I don’t look at any fundamentals. None a single one. I don’t look at. Any technicals, and the reason why is because that will kind of skew my decisions on when it comes to the information I’m training. All I care about is was I able to identify some change that is relevant to that company that will either move the revenue needle or move the perception needle for that company once the rest of the world.

Appreciated or learned about that change, and for some odd reason, the rest of the world doesn’t see it yet. I might be an hour early, a day early a week early, a year early. 

Here is an example of one of their trades:

One of Camillo’s favorite trades was when he discovered a trend of children making slime at home. You may remember this if you had kids a while back.

Camillo said:

“It was happening globally with kids. Kids all out of nowhere started to make slime. It became a big crafting trend for kids, where they would create all kinds of different coloured slime and different textured slime just to play with at home,” 

While there wasn’t a pure-play slime company he could invest in, he found that white glue was the main ingredient in DIY slime, which was predominantly manufactured by Elmer’s. 

“When I started researching Elmer’s glue, I realized that they were selling out globally.”

Camillo promptly made an investment in the business’ parent company (Newell Brands [NWL]), which, sure enough, saw its revenue spike 50%, despite it only making up 1–2% of its total revenue. 

Camillo made close to 200% on the leveraged trade.

“That’s a really interesting example where the whole world could have seen this. There were 10s of millions, if not hundreds of millions, of families globally, where there were kids making slime. How many were able to make the connection between that and an ingredient of slime and the company that was investable?”

Spotting something like this before the big funds come in, can bring in massive amounts of money.

But competing against hedge funds meant that they would be able to use software and an army of interns to scour the info on the web, analyze it and make trades on it before you could.

Here is where ChatGPT comes in.

I’m going to ask it to analyze the sentiment of a Tweet:

We will ask it.

Decide whether a Tweet’s sentiment is positive, neutral, or negative.

Tweet: This new software is scary good

Sentiment:

Positive, it says.

A better way would be to rank how positive, let’s ask it to rank it 1 to 10.

Rank how positive this Tweet is on a scale of 1 to 10.  1 being extremely negative, 5 being neutral and 10 being extremely positive

Tweet: This is by far the best sauce I’ve ever had in my life

Sentiment: 10

That’s pretty good.

Let’s try another

“After reading that book, I felt like it had some good moments, but overall I could have done without reading it.

Sentiment”

5

I would agree with that.

“This wasn’t the best movie in the franchise, but it was up there”

7

Yeah, I mean, I would more or less agree with all of those.

Now it’s important to understand here that you can spend time to make it better at understanding Tweets and what they mean. 

If there is some way of speaking that a particular audience will use, that maybe ChatGPT won’t pick up on, you can start pulling in Tweets for it to analyze and then correct it, it will learn and over time it will be accurate in rating them.

I actually at this point asked ChatGPT if it can detect sarcasm and it replied that it’s difficult to detect sarcasm in writing without having the context or tone of voice.

Which that is true, even for humans, a tweet for example can sound legitimate, if you don’t have the context.

I actually spend a long time asking it to rate certain statements to see if they were sarcastic.

The interesting thing is that it gets it. For some statements, it will say “this might be sarcastic if it was made in reference to this” etc.

But it tends to default to saying  “it’s hard to know for sure” quite a bit.

My point is that you could easily build multiple run thoughts to rank how positive certain statements were, weed out sarcasm etc

You could replicate a human being combing through these tweets and accurately rating them.

That’s a legitimate use of this software right now.

You just need to figure out what data you want to extract.

Ok, so that’s fun. But…

How do we use it?

Well, Another thing that ChatGPT can do is write Python Scripts.

Python is a VERY powerful language for collecting data and organizing it.

I’ve always had learning Python, on my list of things I wanted to do, because of it’s ability to collect data from the internet, automate certain tasks in your browser, organize a bunch of data and throw into an excel spreadsheet etc

Using ChatGPT, we can start creating python scripts that collect data from Twitter, use ChatGPT to analyze the sentiment and then use Python to log that sentiment into some sort of a chart that tracks it over time.

And again this can be done without you being able to code.

This tech was available to hedge funds years ago, but it took many highly trained and highly paid computer scientists and data scientists to run this process.

Now, it’s possible to replicate by practically anyone willing to take the time to do it.

Mark my words, in 2023 we will see a headline that says something along the lines of “14 year old kid builds a trading algo with ChatGPT that makes thousands or millions of dollars” or something like that.

All the tech is there, the tech skills needed to create something like this went from someone with 4 plus years of college level computer science to almost anyone really.

An article just came out saying NYC is banning ChatGPT from it’s networks, because kids are already using it to write homework assignments including coding.

And the applications are limitless.

There are publicly available information on oil tanker movements across the oceans, we have publicly available information about plane flights.

The revenue of those companies is directly tied to those ships and planes moving around.

You see a particular oil shipping company’s ship never leaving port? It’s movement is down 90% from last quarter and it’s stock didn’t respond to that news?

If every person can harvest that data and analyze it with the help of AI, that will completely change how trading is done.

We are about to see some wacky things going on in the investment world.

If you are interested to know more about where AI is going and how normal everyday people will be empowered to use it, click subscribe, I’d love to be your guide on this journey.

My name is Wes Roth, thank you for watching!