As an AI enthusiast, I've long been aware of the incredible capabilities of ChatGPT. However, it continues to make me smile as I find new ways to use it that reinforce just how versatile and indispensable it is, even outside the office.
This week the reinforcement came from a friendly Masters pool. As someone that loves golf but doesn't have the time to watch each tournament, I was looking for help in making the right picks. Since Gen AI has occupied much of my brain cycles lately, it felt natural to ask ChatGPT how it could help.
My use of Gen AI has really evolved over the last couple of years. In the beginning of my relationship with it, I used a lot of modified Google search queries wrapped with the unnecessary "Please" and "Thank you". Now, my chat's with GPT are much more collaborative and conversational. Just like with Google, producing an effective prompt is the differentiator between a quality and an exceptional output. To get help with the Masters, the conversation started with "I need to guess the winner of the Masters. Outline a process where we can work together to build a solution that can model the Masters golf Tournament and who will win." In response, it provided an extended strategy for us to follow.
We (ChatGPT and I) worked together; my role being something like a Project Manager and GPT putting on different hats to be the experts needed at each turn. I understand the nuance of golf along with GPT's limits, which enable me to keep the AI within the appropriate bounds and redirect it as needed. GPT, on the other hand, understands the deeply technical concepts and can quickly apply those concepts in ways a human cannot.
Much like my direct reports in a daily standup, as GPT reached a hurdle (for instance, it didn't have 2023-24 season data on the current players), I provided it with a solution (I copy and pasted data tables from the PGA stats sheets). As it started to apply statistics, I noticed it wasn't considering things like weather (as a human, I know that the tournament will receive rain this week and historically that means lower scores), so I asked it to go do more research on what factors, outside of the players themselves, can play into scoring during the week of the Masters. Of note, post research, it added weather without me explicitly directing it to.
In about 60 mins, ChatGPT had built a model for evaluating the 2024 Masters, predicting Scottie Scheffler to win this year (not a huge surprise) with John Rahm in a close second. However, if you’ve participated in many golf office pools, you’ll know that it’s not just about picking the winner. You’re presented with a handful of groupings of players, and the winner of the office pool is the person that can pick the player that finishes the highest out of each grouping.
To finish the analysis, I updated ChatGPT of our final intent; “I am going to paste a group of players into chat. Evaluate these players against each other and pick the person that you estimate to finish the highest out of that group”. I then copy and pasted the brackets of players into GPT for it to pick each of the 6 bracket winners. So, what did I pick this year with the help of GPT? Here are my selections after 90 mins, start to finish, with GPT’s help:
- Jon Rahm (Sorry, I just couldn’t pick Scottie because of friendly banter my son and I have back and forth about him.)
- Tony Finau
- Sahith Theegala
- Byeong-Hun An
- Denny McCarthy
- Harris English
- Tiebreaker - What will the winning golfer's score be at the end of tournament, not including playoff holes? 15 under par
This exercise truly excites me because it continues to highlight Gen AI’s potential as a sandbox for innovation and collaboration. As someone deeply entrenched in the AI landscape, I see firsthand the endless possibilities it offers for exploring new ideas and pushing the boundaries of what individuals can produce (with the help of AI). Yet, I recognize that many are still unaware of the full spectrum of capabilities that Gen AI brings to the table. That's why I'm passionate about sharing my experiences and insights, encouraging others to utilize the multiplying power of AI.
About the Author
Chris has been interested in what we all now refer to as AI for over ten years. In 2013, he published his first research journal article on the topic. He now helps companies implement these progressive systems. Chris' posts try to explain these topics in a way that any business decision maker (technical or nontechnical) can leverage.