First 90 Days Reflection
Twists, turns and idea maze navigation with demand as my new north star
I started my most recent venture exactly three months ago on October 13th. I’ve learned a lot along the way!
I’ve written 35 posts about my journey, kicked off with the initial vision of an AI-Powered Learning System.
As we stand on the cusp of an incredible technology revolution, AI offers unprecedented ways to interact with and retain knowledge. I’m building a purpose-driven AI-powered Learning System, designed to enhance how knowledge workers learn and retain critical information more effectively.
I still believe this but also have learned an immense amount about startup building along the way.
I had these three goals 90-days ago.
Find Serious Learners
Get Smarter on EdTech
Build and Release an MVP
How did I do?
Find Serious Learners - Grade D
Despite my best effort I couldn’t find a community of serious learners that vibed with my vision. It’s not that there weren’t serious learners out there. There are!
The problem is serious learners already have the tools at their disposal to build a great learning system, including a system powered by AI.
My co-founder, Matt and I, went deep into learning tech to figure out where we could play. We eventually ended up piecing together a really great system with open source tools. I posted about my setup on LinkedIn where I use Logseq to write down the things I’m learning then sync with Msty via iCloud. Msty is a powerful chat client that can link to any LLM API.
Every day I talk to my learning journal and ask it to quiz me. It’s a great work flow and gets me about 90% of the way to my initial vision.
Get Smarter on EdTech - Grade B
I felt good about my efforts here. I spent time with many EdTech pros and learned about the space.
I mostly learned to avoid EdTech, or at least the boundaries drawn around it. There are businesses to build here, but I’m not certain I’m the right person to build them.
Build and Release an MVP - Grade D
We didn’t release a product but boy did I have fun learning how to code with AI. I’ve since learned there’s a phrase called “Chat Driven Programming” to describe the paradigm shifting approach these AI tools require. I’m a full embracer of this new paradigm.
I used Github Copilot, Cursor and Windsurf to supercharge my development efforts. We made great progress on one idea we called Personalized Learning Feed, a social media inspired feed of the things you are trying to learn. There’s a sea changing happening for writing software and I’ve had a front seat to it.
I have a workflow I love using Storybook for frontend UI testing. I’m working at a higher level of abstraction then I did pre-AI days coding. I barely look at the code. I focus my efforts on getting the right test harness in place for the AI then instructing the AI to iterate and iterate until it gets the product right.
Why didn’t we see this product through? The net of it was there was little demand pulse for what we were building. We came close to just seeing it through for the sake of releasing something. For a few days I had a crisis trying to make the decision but ultimately settled on, if we’re going to build a business we need to find something with real demand.
Discovering Demand-Side Selling in PMF Camp
I wrote this post, More Idea Maze Navigation, to describe my thought process. “How fast can we get to revenue?” kept ringing in my ears and I eventually decided to make an investment in Rob Snyder’s PMF Camp to learn a better methodology.
My time in PMF camp was game changing as I wrote about in Demand is all that matters. It’s a two week course with a lifetime of alumni access. Every Wednesday alumni meet and Rob goes through a live tear down of a founder’s sales pitch. Wednesday’s meeting is one of my most important meetings of the week because it helps me recenter on the Demand-Side sales methodology Rob teaches in PMF camp.
Out of PMF camp my MVPs became 5 slides to test demand. We aren’t building anymore in the traditional sense. We’re searching, trying to find the elusive Venn diagram overlap of demand (what people want solved) and what we can deliver.
This search for demand is the crux of being an entrepreneur. We’re demand searchers, trying to find a way to solve the problems in the world. Not just any problems though. The problems people want solved right now!
One of Rob’s most impactful mental models is imagining that everyone you talk to has a Trello board of projects on their plate. If you ask them about their pain points, they can list off a litany. But, and this is the key point, they won’t try to solve them (read: pay money) unless they move to a column called “critical / most get done projects.”
The goal of demand searching is to find a project in that column then build a solution for it. If you build for anything else, it’s going to be a grind pushing up a hill to convince people to buy.
I’ve seen demand at Lucid and VS Code. I think there was some demand at OSPInsight (sold to IQGeo) and Elements for a certain segment of customers. In the later two cases, I think there was much room for improvement in dialing in the when of the customers circumstances that led to demand (when vs. what is a key principle of demand-side selling).
The Next 90-Days
All this brings me to today. I’m 90 days in. I’m working on demand ideas and refining them by having many, many conversations with people who actually have the Trello board in their head.
My current demand idea is to help SMB owners get their team to use AI.
AI is a fast moving space and it’s hard, even if you are fully committed to learning about new AI tools as your full time job (as I am), to keep up.
Add to the hype around it: does it even work? Where do we even start?
My thought is the fast moving and sometimes over hyped nature of AI is leaving an opportunity for someone to help SMB owners train and implement AI for their teams.
Call it an AI consulting play for now. I have a handful of clients I’m working with to get actual revenue to help them solve their AI problems.
I like this approach because I can earn real revenue and get deep with a handful of clients to understand their demand and where AI does and doesn’t work.
We’ll see where it evolves. Here’s to the next 90 days!
...
#HappyLearning and #HappyBuilding