I Asked AI to Help Me Beat the Stock Market

As a red blooded American millennial male, I have a tendency to log onto Robinhood and make some degenerate stock investing moves. I hope you can relate with this, but that's not necessarily good. How can I use AI and AI vibe coding tools to help me develop a stock investing strategy so I can beat the market without risking my family's financial future? Well, today I did that. I just built it in repl dot. And if you're not familiar with Vibe coding, it just means using natural language, like the language I'm using right now to code apps, because most of us don't know how to code. So I'm pretty hyped about what I just built. And I want to show you, and I want to show you how I got there so you could build something like this for yourself. And at the end of this, I'm going to give you about a dozen really cool prompts that you can use in ChatGPT or any other ChatGPT clone out there to find cool investing strategies that fit your risk profile or income level. So let me start by telling you what assumption I led with. I led with the assumption that companies that still have the founder as the leader and companies that enjoy network effects, and I'll get to what that is in a minute. Outperform the market. And by outperform the market, I mean provide better returns than just investing in the entire S&P 500. Now, what are network effects? Well, network effects are when a company gets more valuable with every new user that they add. For instance, eBay. EBay is a buy sell marketplace. The more sellers they have, the more legitimate their marketplace appears to buyers. Therefore, the more buyers flood to the market and it just grows and grows from there. Facebook. The more friends you have on Facebook, the more likely you are to be on Facebook. That's a network effect. It's a moat. It makes companies very resilient. Tesla has their supercharger network. That's a network effect. Other companies can use their superchargers. Tesla owners rely on them. StubHub has network effects because it's also a marketplace, like ebay. You get the idea. So if a company has network effects and is still led by the founder, then it will outperform the market. That was my thesis. So Replit is one of my favorite vibe coding tools. And they just came out with this new thing called Agent 3, where you can Vibe code a tool and then it will start stress testing your tool, breaking it, fixing it, breaking it, fixing it, all while you just sit there and wait Whereas before, you would have to break it and then figure out how to prompt it in the right way to get it to fix itself. Those days are gone. It's amazing. So let me share my screen, show you how I built what I built, show you the final product, and then give you these prompts so you can develop your own thesis for stock market investing. All right, my first prompt was I want a portfolio tracker that I can benchmark against the S and P. In other words, that I can compare against the stock market as a whole and upload via spreadsheet all of my historical trades. Here is the spreadsheet in question. I want to see 1 day, 1 week, 1 month, 1 year, 5 year, 10 year, 20 year performance in a sleek chart like Robin Hood shows. Here's the spreadsheet. So to get this spreadsheet, I went to ChatGPT, I used deep research and I said, find me all of the companies in the S&P 500 that have both network effects and are still founder led. There were 26 of them. And then I said, make a spreadsheet that would show what my returns would look like today in 2025, if on the day those companies went public, I invested $100 in that company and $100 in the S and P in the market. So then I can track over time how much money would I have today. Because really, we're talking about investing $2,600, 26 companies times $100. So one scenario is $100 in each of the 26 companies, and the other scenario is $100 in the S&P 500 on the same day that each of those companies went public. So it spit out this spreadsheet. And then I uploaded the spreadsheet into repple it because that's how I started this project. I wanted to just visualize this data that my spreadsheet was showing me. But then it evolved into something much cooler. So I did. I uploaded the spreadsheet, I hard coded it in the app and it showed me something kind of like this, like this chart right here, which is really cool. But I had no way of knowing if it was actually accurate. So it gave me a rough draft of my app that told me everything it was doing. And it worked. One shot prompted it. It worked right off the get go. But then I realized, okay, I need more data than this. I need to know what days all of these companies went public. I need to know what their share price was when it went public. And that was a lot more tedious to find But I did. I eventually found it. Most of it was in Wikipedia. And then I had to supplement it with some other AI tools. And I ended up with this Google sheet right here. I'll share this with you as well. Called all S&P 500 stocks. You can see the name of the company, the ticker, the IPO price, how many shares you would need to buy. $100 worth, today's price. And if you look at the formula right here, I'm just using the Google Finance formula. So this updates every minute or two. The current market cap, the PE ratio, the year it was founded, the date it was added to the S&P 500, because that's usually not the same day that it had an IPO and the exchange it's on if it's founder led, if it has network effects. And then I did something really, really cool that I'm proud of. I use this tool called GPT for Sheets right here. Okay. And GPT for Sheets is what it sounds like. You can basically use any AI tool inside of a Google sheet. And so over here I had a prompt that said, tell me on a scale of 1 to 10. And then I inserted the name of the company, how much, in this case, conagra Brands enjoys network effects. One is not at all. Maybe they just sell widgets. There's no network effects there. 10 is very strong network effects. And then I did basically the same prompt for founder led. You would think that it's binary. Either a company's led by its founder. It's not. But it's not. Because sometimes the company's founder sits on the board and they have a strong influence over the board or over the company, but they're not technically the founder. Sometimes like Meta. Mark Zuckerberg is the literal founder and the literal CEO of Meta. So that's a 10 out of 10. So I wanted to rank these. So it did. I just drug it down. It cost me like 15 bucks in credits. And I got this sliding scale of how much a company has network effects and how much they are founder led or not. So then I took the editable link for this Google sheet and I went back to Replit and I said, hey, forget that spreadsheet I uploaded with only 26 stocks. Here's the whole S&P 500. Here's what each of the columns mean. I want this to be your database that you pull from when displaying data. And it was like, no problem. It did it for me. Exactly. Because not only did I want to see what stock should I invest in but what does it look like? How much does my thesis beat the market? And I was pretty shocked. I also wanted to see like should I invest in companies that have a 5/plus network effect score? Only 8/5 + founder et cetera. A mix of the both. How does market cap come into play? I kind of overcomplicated it a little bit but. But I was really happy with what came out. I also wanted to see a time weighted returns and unweighted returns. Let's say I invested in one stock that fits these criteria 10 years ago and one that fits it 15 years ago. Well, I want to see a weighted average return of the S and P If I invested $100 on each of those two dates that are five years apart and an unweighted like what is the S&P 500 in total done over the last 15 years. So I wanted to see both of those on the chart. Here was my prompt. Okay, we need to revamp the app. Ignore the stocks in the upload entirely, pull the data from this live Google sheet and you can see I told it what each column meant. Now it can surmise what each column means. But I don't want to assume that it knows. I would rather get ahead of it. Take a few more extra minutes to tell it just to reduce the potential for error. And then I did little things like I want to comma in all of my numbers. I only want one decimal point on my percentages. And it fixed all those really easily. I wanted to check the math of the chart and show me the stocks at the bottom of all this so I can make sure that it was filtering correctly. And I was shocked at how accurate it was just from the jump. And so at the end of the day to see the final product, this probably took 45 minutes of prompting. All in all, boom, right here it's called NEFL Stocks NEFL Network Effects Founder led. And if I rank it, let's go 8 to 10 on network effects 8 to 10 on founder led. I won't touch the PE ratios or the market cap ranges. The default is a 10 year time horizon and I can change that on this dropdown right here. And it's loading. You can see right here six companies match your filters. Five of those six have valid IPO price data because some of them I just wasn't able to get all the data for. And now look at this. The S and P has returned 9% over this time frame weighted and 12.5% unweighted. And this portfolio would have returned 22.4%, which is over twice the weighted return. Crazy. The red line is the S and P, the green is this portfolio. And then it shows you the six stocks that it filtered here. And then you can verify that the IPO prices were correct on all those. How cool is that? All right, as promised, if you want to find cool stocks to invest in, here are some of my favorite prompts. Or, you know, teach a man to fish. Yada, yada, yada. If you want to come up with your own frameworks to pick your own stocks to invest in, here are the best prompts. Number one, show me small micro caps with moat, high ROIC return on invested capital, but low analyst coverage. In other words, really small, obscure, unknown companies that analysts are not covering. Which this is kind of similar to the thesis that I had with John McAfee when I launched that crypto company. Companies that had a lot of hype but very little market cap were about to go up in value. People that had low hype but a high market cap were about to go down in value. Number two, companies where insiders are buying materially over the last six months, but stock is down over 20%. Wow. I love that one. Holy cow. Because it's public information if an insider is buying a stock. So what stock are all the employees buying because they know something we don't know, but they're still down, AKA the stock hasn't been reflected of these impending awesome things yet. Number three, stocks with high free cash flow, yields of over X percent but PE price to earnings below peer median. In other words, stocks that are outperforming their peers that aren't reflected yet in their share price. Number four, businesses in structural secular tailwinds that the market is ignoring. This is going to be like a new regulation or demographic shift, tariffs, etc. Number five, industries undergoing disruption. List incumbents with weak balance sheets versus disruptors with strong balance sheets. This is like a valuation mismatch. Kind of like Nvidia, the chip company. Everyone's investing in them, everyone knows them, it's the most valuable company in the world. And then you've got amd, which is a direct competitor of Nvidia, but they're not as hot, they're not as sexy, but they're still doing awesome things with AI. Number seven, look for companies where short interest is high but fundamentals are improving. In other words, investors are predicting it's going to go down in value, but the underlying fundamentals of the business are looking stronger than ever. Number seven, companies with recurring revenue, customer stickiness operating leverage but market giving little credit. It's kind of self explanatory. The next one find under followed stocks with improving margins and growth inflection list stocks trading below their private market value or liquidation value. This happened with Coinbase like three years ago. Coinbase had more cash on their balance sheet than they were worth at as a company. That's a fact. And I looked at it and I was like huh? And some people were tweeting about it. It was a thing. But crypto was still risky in the eyes of many investors today. And then so I took that risk. I went in hard on Coinbase and it's my biggest holding and my best performing holding right now, which is a great little pair. Companies with high barriers to entry scale advantages, but currently in boring sectors. Iren, I really like that stock right now because it's a data center stock that used to be in bitcoin mining and now is in AI. Very high barrier to entry, riding an insanely large tidal wave and is pretty boring and unsexy. All right, that should be plenty to get you going. Also, you know, I couldn't help myself. I went and bought a domain name, neflstocks. Com. I put this free website behind an email form. So if you want to use it for free, it's completely free, there's nothing to pay for. You just got to put in your email and you can play around with it and maybe one day I'll build more stuff on this.