In partnership with

The Gold standard for AI news

AI keeps coming up at work, but you still don't get it?

That's exactly why 1M+ professionals working at Google, Meta, and OpenAI read Superhuman AI daily.

Here's what you get:

• Daily AI news that matters for your career - Filtered from 1000s of sources so you know what affects your industry.

• Step-by-step tutorials you can use immediately - Real prompts and workflows that solve actual business problems.

• New AI tools tested and reviewed - We try everything to deliver tools that drive real results.

• All in just 3 minutes a day

Join 1M+ pros

I Gave ChatGPT $500 of Real Money to Invest in Stocks

OpenAI says that their new GPT-5 model is a “PhD level intelligence.” It’s supposed to be good at a wide range of “useful” tasks.

Early this month, I decided to test it on the extremely “useful” task of making me a ton of money by investing in the stock market.

So, I gave the model $500 of actual money, and asked it to pick 5 stocks that would make as much money as possible in the next 6 months.

I documented the entire process in my most recent article in Fast Company.

How is it Doing So Far?

We’re now three weeks into my experiment. So, how are ChatGPT’s picks doing so far?

Here’s a video sharing more about the experiment—including a live update on the bot’s portfolio performance after 3 weeks. Spoiler alert: the results are dramatic so far. (Watch on YouTube).

Disclaimer: This is an experiment, and is not intended to provide investment or financial advice. Consult a professional advisor before making any investment. Do not trade based on the information provided here.

This is a fun experiment. But it also provides the perfect opportunity to test two new aspects of GPT-5, instruction following and safe completions.

Instruction following refers to the model’s ability to understand the user’s intent and follow it faithfully. GPT-5’s decision to provide me with risky stock picks instead of waffling or providing a wussy non-answer indicates that it’s quite good at instruction following.

Safe completions is OpenAI’s new framework for answering potentially damaging queries. Basically, it’s a method for allow the company’s models to respond with general information rather than refusing to answer sensitive queries.

I was worried that the “safe completions” framework would steer ChatGPT towards avoiding my stock picking question. Perhaps it would refuse to select specific stocks, for example.

In fact, it was very happy to design a risky portfolio for me. That indicates that—for all the talk about safety—OpenAI isn’t putting overly restrictive guardrails on their models. My best guess is that they’re leaning into safety primarily on medical/mental health queries, and leaving most business/finance topics alone.

That’s good news if you plan to use ChatGPT for financial analysis or business tasks—not just crazy investing experiments!