How I Use Chat GPT

And light assessment of where we are in the innovation cycle

November 30, 2022

Throughout 2022, I experimented with DALL-E to generate images for bedtime stories for Van. This experience, even before the release of Chat GPT, made me excited about its potential.

When OpenAI released Chat GPT on 11/30/22, it was a Wednesday. The following Monday, I asked teammates if they had played around with the service over the weekend. Reactions were mixed; some were skeptical, others were blown away.

I was in the blown away camp. And still am.

Mid-2024 & how GPT plays a part in my daily life

I’ve observed three primary use cases in my personal life (e.g. dad to 2 young kids, husband, living in Atlanta) and 3 primary use cases in my business life (e.g. early stage startup founder).

Personal Use Cases

1. Search/discovery for random life questions:

   1. Examples: “removing a tick from a toddler safely,” “how to cook a tuna steak on the stove,” “how does Atlanta’s park system compare in number and acreage to other large metros in the U.S.?”

2. Search for business related information:

   1. Examples: “what was the inflection that helped catalyze companies like lyft & uber?,” “how are rental properties taxed?,” “how quickly did KiniksaOne get to $100M in revenue?”

3. Teacher/Tutor for code writing:

   1. Example: “I want to create an app that manages local farmer’s markets. I want it to be written in Ruby. Let’s start from scratch. Give me a step by step guide, one step at a time.”

Business Use Cases

1. Content Marketing

   1. We've created GPTs that transcribe call and podcast transcripts into blog articles, helping us build a consistent content engine.

   2. I use Chat GPT to improve the readability of LinkedIn posts. I paste in a draft and ask GPT to grade the post on a score of 1-5 for utility and readability. If I’m impressed with Chat GPT’s grading and explanation, I sometimes prompt it to incorporate its proposed suggestions by generating a revision. While I find the grading unimpressive 50% of the time, I’ve noticed improvement.

2. Product Knowledge & Sales Engineering

   1. I created a GPT by uploading our developer docs and lightly training it to be an expert Sales Engineer/Knowledge Base for Schematic. This helps me quickly answer questions from customers and prospects.

3. Thought partnership & clarity creation

   1. I regularly use Chat GPT as a sparring partner to clarify my thinking. For simple tasks, like reviewing a sales email for precision and concision, it’s a quick thought partner.

   2. For more complex explorations, I prompt Chat GPT to pressure-test my thinking against a variety of mental models. I share definitions of models like Occam’s Razor, Inversion, and First Principles, then ask it to use these models to clarify my thinking on specific topics.

Where are we in the innovation cycle?

Over the last ~18 months, I’ve wondered whether we’re living through Carlota Perez’s installation period, or deployment period. The purpose of asking this question is to explore how to think about AI’s investment frontier as an entrepreneur or investor.

Perez argues that there are two main periods of every technological revolution: installation and deployment. Within these periods are 4 distinct phases: irruption, frenzy, synergy, and maturity. 

Chat GPT argues that we’re in the Deployment Period, as AI began as academic research in the mid-20th century and is now integrated into business processes and consumer products.

My sense is that we’re in the Installation Period, and more specifically, the Frenzy phase. Frenzies usher in a wave of speculation as financial capital, afraid of missing out, rushes in to capitalize on the new, adjacent possibilities created in the irruption phase.

David Cahn has an interesting article about navigating what he views as an AI bubble, and he makes the good point that speculative frenzies are necessary and those who remain level-headed can build important companies.