On Becoming an “AI Product Manager” — Week 1

I sat down on a Friday evening in 2019 months before the pandemic was more realized as global — at that point just a blip in the news among many other things — and I committed to myself to transform some of the latent anxiety I had been feeling about my life’s trajectory into something more structured and intentional.

I was coming off of a strange transition. I put a lot of my effort and thinking into a project about community building that I still put some effort and thinking into. But by 2019 I had gone through a major burnout and transitioned through relationship , employment, and residential changes and community building, while still an aspiration, felt quite far from being a reality. And so I was back to putting time and effort into thinking about the future I wanted to live in — my personal microcosm of meaning and making and meaning-making.

I set a bunch of goals and I feel shy about sharing them but here they are:

Short Term (1–3 years…2020–2023):

Personal Goals

• Produce 3 Art Projects ✅
• Start an Adæpt Webshop & Service ✅
• Travel to Surf & Write — 2 Countries

Professional Goals

• Become a Product Manager ✅
• Product Manager of AI/Data Products
• Become a Certified Cloud Architect
• Start Masters at USC ✅

Mid Term (4–12 years…2024–2032):

Personal Goals

• Write a book
• Creative Talent Agent & Partner
• Building equity on Land

Professional Goals

• Mg. Dir. of Data Eng. & AI Products
• 3 streams of income — Product, Mgmt, Legal
• Legal Work — IP/M&A/Contracts/Regulation

Long Term (13–21 years…2033–2057):

Personal Goals

• Community Steward, Elder, Advisor
• Helping to raise children
• Traveling & Creating

Professional Goals

• Productive Firm Steward, Elder, Advisor
• Teaching, Researching, Guiding Lab-Studio
• Partner in Managing Project Financing

Retirement (~2058+)

Personal Goals

• Community Steward, Elder, Advisor
• Helping to raise children
• Traveling & Creating

Professional Goals

• Board member of Product Co-Op
• Speaker & Writer
• Strategic Partner for AI Products

This list is different from its first incarnation in 2019; at that point I hadn’t thought of applying/attending USC for the M.Sc. in IDBT. I was imagining that the next step in education I would take would be in legal — I think the idea there has been that a law degree and the license that follows it enables a set of abilities that are rare and sought after and that would be interesting — but the thing that stands out and kept coming up has been the desire to work as a product manager and then the desire to ultimately work on AI Products.

In 2021 I hit the goal of becoming a product manager and for that I’m grateful. And while I wish that I had done more between 2019 and now to deepen my abilities and understanding of all the things that constitute AI, I’m here in 05/05/2023 with a resolve to follow through on this goal.

I took a machine learning course in my grad program and using models, playing with the data to get the models to respond in ways I wanted them to, and reporting back on what those responses were like was all a rewarding exercise. I’m feeling excited about digging deep into “not knowing” again. And so, recently I’ve signed up for a couple courses to push me along in this goal:

1. Dan Shipper (from Every) is teaching a course on building chatbots with the Open AI API

2. Jon Reifschneider’s course on AI Product Management via Coursera

As I take these courses, I’ll complement this knowledge with other things I’m learning and hope to coalesce them into principles, practices, projects, and thoughts. This will be the first in a series of posts I do about my journey becoming an AI Product Manager.

So — reporting back from Week 1 of this Path:

Session 1 of Dan’s course was great; I’m excited to be a part of the Every community and alongside the learners in this cohort. We went over some high ground on what to expect and how to operate. We’ll be leaning on ChatGPT as much as possible for learning to do the coding we need to do (in the same way + alongside using traditional references such as Google and StackOverflow)

We’re also using Replit, which I’m just getting into and love and recommend. Here’s the code I made with ChatGPT’s help (it mostly wrote the code):

import openai
import os

# Set the API key
openai.api_key = os.getenv('OPENAI_API_KEY')

def chat_with_gpt(prompt, conversation_history):
    # Add the user's message to the conversation history
    conversation_history.append({"role": "user", "content": prompt})

    # Generate a response using the conversation history
    response = openai.ChatCompletion.create(
        model='gpt-3.5-turbo',
        messages=conversation_history,
        max_tokens=150,
        n=1,
        stop=None,
        temperature=0.8,
    )

    # Extract the response text and add it to the conversation history
    response_text = response.choices[0].message['content'].strip()
    conversation_history.append({"role": "assistant", "content": response_text})

    return response_text

def main():
    print("Welcome to the GPT-3.5 chatbot! You can start chatting now. If you want to exit, type 'quit'.")
    
    # Initialize an empty conversation history
    conversation_history = [{"role": "system", "content": "You are chatting with a GPT-3.5 powered AI."}]
    
    while True:
        user_input = input("You: ")
        
        if user_input.lower() == "quit":
            print("Goodbye!")
            break
        
        response = chat_with_gpt(user_input, conversation_history)
        print("GPT-3.5: " + response)

if __name__ == "__main__":
    main()

Throw this into your Replit, put your OpenAI key into your Replit Secrets (get a CC linked to your OpenAI account) and bam you can start messing with the API yourself 🤙

Week 1 of Jon’s AI Product Management Class has been cool as well. Going over the history of AI, I appreciate how far back this endeavor of building “cognitive machines” goes

And it’s also given some great context on the relationships between the different fields that make up AI:

My aim is to understand “the how” as well as I can, so that as I start becoming better at defining “the why” as a product manager. Moreover, I want to understand the UX/UI principles that are applicable to building products that surface or leverage AI — and how these principles may be different depending on the extent to which AI is used and how central it is to the product interface.