Skip to content
Provolove
TwitterLinkedIn

Learning to AI

— Tech, AI — 4 min read

TL;DR - I’m trying (and struggling) to learn to use AI effectively in my work which has reminded me of my “Learn to Code” journey from a decade ago. I’ll be attempting to document this process here on my blog. If AI is anything like the hype would have us believe it’ll write the whole series for me.

The Hype 🚀

The AI hype has been going strong for over a year now, promising overnight industrial revolution or societal collapse, maybe a little bit of each. Generally, I think of myself as an early adopter but I've found myself a bit behind the curve on this one. However, as I've been diving into articles about the surge in funding, the copyright debates, and every SaaS platform touting their new "AI assistant" features, it's reminded me of the 'Learn to Code' fever of the early 2010s.

I'm thinking about the days when the next “killer app” was what all the buzz was about. Back then, app stores were booming, and coding bootcamps were popping up everywhere. It was the era when everyone believed that with the right idea and some coding skills, you could become a billionaire as an indie developer. Learning to code seemed like acquiring a super power which would unlock untold creative abilities and a limitless career. Starting a company with a “strong technical founder” on the team was at best naive and at worst a death sentence in the pitch room.

It’s not a perfect comparison, but the AI as a cofounder mentality and the near-requirement of integrating AI into any product to stand a chance landing VC funding feels familiar in this way.

Rewind ⏪

I attended a “learn to code” bootcamp in 2015 and loved it. My skills progressed much quicker than they had on my own and I made wonderful friends. However, I soon realized that while coding was valuable for my career, it wasn't a guaranteed path to success. Just as learning to write any language doesn’t immediately lead to writing novels, learning to code was a foundational skill. The outcomes of applying that skill depend on interest, motivation and talent. Many of my peers became successful Software Engineers over the years, others charted their course as indie developers and some transitioned into related roles such as Product Management 👋, Solutions or Technical Sales and UX Design.

I have no doubt that AI will have a monumental impact on founders, developers, and workers across various fields. However, I believe that individuals will need to invest time and effort to discover how to work with these tools and ultimately develop an ability to use AI to enhance their own unique human skills.

AI for Product Managers 🤖

For the past few months, I've been grappling with integrating AI into my role as a Product Manager, feeling a bit lost at times. Much of my work involves processing and analyzing information, which seems tailor-made for AI. Tasks like writing Jira tickets, summarizing performance results, and creating product documentation, while not my favorite, are essential and time-consuming. It feels like AI could easily handle 30-50% of my workload if I could just figure out how to best leverage it!

Despite numerous attempts, most of my efforts to offload tasks onto our soon-to-be superior species have fallen short. However, I'm undeterred. Over the next three months, I'm determined to push myself to at minimum a vague existential concern for my job security.

The Problem

My observations lead me to beleive that I have been too narrow in my approach in working with AI tools. Likely because I also haven’t familiarized myself with the basic concepts. I’ve been using only a handful of tools, primarily as glorified autocomplete or search replacements without experimenting or learning enough.

Tools

So far, these are the tools I’ve attempted to use regularly.

  • Notion AI: I use Notion for everything from personal notes to building product roadmaps and drafting developer documentation. Using Notion’s built in AI has been underwhelming, but again probably just because I’ve been asking it to complete sections of text for me…
  • ChatGPT: I’ve experimented with a variety of document drafting requests, to mixed results, and more generic search related items, though I prefer Bing/Copilot as its use of references makes fact-checking easier.
  • Microsoft Bing/Copilot: Not to be confused with Github Copilot, I have found Copilot to be an excellent search replacement as it blends summaries and references to other sources well.

Goals

Reading posts like this one from the brilliant (and hilarious) Casey Newton, lead me to believe that many people are getting much more productive contribution from AI tools than I am. I’m trying to catch up:

  1. Discover at least 1 frequent task of mine that can consistently be completed faster with AI when compared to without.
  2. Reduce my own time spent on selected tasks by 30% using AI tools

The Plan

Even while writing this, I realized “maybe I should be using AI right now” and after asking ChatGPT to review some paragraphs and make some editing suggestions I pivoted to:

Can you outline a plan for me to systemically experiment with and learn how to use AI for my work over the course of 3 months?

This yielded surprisingly sensible suggestions. What’dya know 😏 

Naturally, I made some modifications but I did find it quite helpful actually!

3-Month AI Integration Plan for Product Management:

Step 1: Understanding AI

  • Research generative AI tooling concepts
  • Research popular tools and resources

Step 2: Experimentation

  • Identify AI use cases in your workflow.
  • Integrate AI tools for selected use cases and gather feedback.

Step 3: Optimization and Evaluation

  • Refine AI integration based on feedback.
  • Evaluate impact and plan for future AI integration and learning.

Hypothesis

By approaching AI tools similar to how I approached coding languages or frameworks when I was learning to code, I will learn new skills adapting my workflows and work style to be more productive, leading to more output with equal or less input (time spent working).

Summary

I’m going to spend the next 3 months attempting to “learn to AI”. This will be broken into steps with the bulk of the time spent on experimentation (more on this later). My hope is that at the very least I will learn some new skills; possibly I will be able to spend less time on content-intensive tasks that I don’t enjoy and be able to redirect this saved effort to more interesting tasks. Or who knows, if Sam Altman is right maybe this will lead to my next startup becoming a “killer app” 😜.