Passion Problems
Oct 23rd - Dec 13th, 2023
It’s been over a month since my last update. I have no excuses, we’re just going to jump back into it and see what I’ve been up to. Since my last post, I’ve researched home inspections, trained a few GPTs, learned how to install hardwood floors, hosted Thanksgiving, upgraded my Midjourney subscription, went to my first Hawks game, continued reading (Overbooked, The Creative Act, Going Infinite), and began building the first version of a new idea. More below.
I caught up with my friend Eli a little while back and he reminded me of something I thought I’d share. When deciding to build a company, it’s not just about finding an idea, it’s about solving a problem you’re passionate about. This is because a problem can be solved in a variety of different ways, leaving room for exploration, and iteration. A good way to figure out what you’re passionate about is by figuring out how you spend your time. This thinking helped me reframe my process and inspired a new problem I’ve decided to explore next.
After shutting down Seasons, I needed a break from New York. In August 2022, I bought my first home in Atlanta and began what became a complete home renovation. It was the first time I ever worked with a lender, put in an offer, negotiated a sale price, booked a home inspection, and sat through a closing. There were so many things I wish I had known then that I know now. With advancements in AI and LLMs, there seem to be some interesting opportunities related to buying, selling, renovating, and maintaining a residential or commercial property.
The same week I closed on the house, I scheduled walkthroughs with 6 different contractors and started to get estimates. The entire renovation took a little over 5 months and went relatively smoothly. I used a combination of Notion for material tracking, Instagram to curate inspiration, Procreate to sketch visuals, and iMessage to communicate with contractors. After we started the demo, we ran into a few surprises that weren’t caught during our home inspection. Asbestos, plumbing issues, roofing leaks, you name it. In hindsight, we couldn’t be happier with how it turned out and would do it all over again in a heartbeat.
Fast forward, I recently had dinner with a friend who works in construction and we started talking about AI + LLMs and how they may impact real estate. He also recently started a renovation project and discovered a bunch of issues in his home. We ended up talking about inspection reports and how they’re not standardized, done almost entirely in PDF format, and don’t include estimates on how much it could cost to fix issues identified in a report. Meanwhile, 77% of home buyers in the US get an inspection done before closing. Here’s an example of how my report looked:
Around this time, OpenAI had just released GPTs. I began uploading a handful of inspection reports I gathered and after a few weeks of training, I was able to summarize the data into something more visual and easy to understand. It also ranked repairs based on severity using a 1 to 10 scale and gave you an estimate based on property zip code, fixed material costs, and experience of labor. I also grouped everything into essential vs cosmetic repairs to help prioritize projects. After getting the GPT to a good spot, I began working on the concept for Terrace. I built everything using Webflow, Midjourney, Photoshop, Notion, and Figma.
After I finish the form uploader and connect payments later month, I’m going to test it out by taking out some ads, cold-calling a few home inspection companies, and partnering with as many real estate agents as I can. After getting some feedback, I’ll figure out whether or not it makes sense to scale or move on to the next idea. My goal is to crank through at least one of these projects per month, for the next 3 months, building components that I can reuse. I’ll try to hold myself accountable and provide updates along the way but if you read these and don’t hear from me, please give me a nudge.
More soon. ✌🏼
