Early Lessons Learned
I have been working on Hobbes full-time for about 2 months now. I thought it may be a good time to provide a first progress update, specifically on what I have learned so far.
- Humans are awesome. This may be obvious, but I’ve been blown away by how helpful everyone has been. From total strangers who are excited to offer help because you are trying to build something on your own, to companies that are willing to jump on calls to explain how their product works and how they can help me. From other founders who spend time answering my dumb questions and helping make connections, to services that are willing to offer discounts & credits just because I am just getting started. My bank - yes, my bank - reached out to share resources on VC networks and refining pitch decks. Above all, all of you who read these emails each week and respond, answer my calls for help, and offer advice & suggestions. You have been awesome. Thank you!
- AI is incredible. The basics - ChatGPT and Gemini - have been very helpful. I use them several times all day and I can’t remember the last time I used regular Google search. The deep research feature in these products has itself saved me weeks of work. Midjourney is great for visualizing concepts and Granola works so well for meeting notes. As a non-developer, Replit and Windsurf have been fantastic to quickly prototype concepts to make sure what I want to build is technically feasible. The combination of Google Scholar and NotebookLM has worked wonderfully for research. I don’t yet use AI to write documents and emails yet because I find the writing soulless. If I had started working on Hobbes even just two years ago, I would have spent at least 2-3x more time to make the same progress.
- Process over portfolio. In the past month, I have spoken to two dozen contractors and software development companies across illustrators, UX designers, and software engineers. I’ve got my fingers burned picking people based on impressive looking portfolios / projects. My biggest lesson learned here is to interview people and ask them about their process - how they work, how they organize their thinking, how they specifically work with startups, and what tools they use - and spend less time on the outputs they produce. Everyone you speak to promises the moon, but asking probing questions on how they work and think really helps find the gems.
- Quality vs cost follows a bell curve. When I shared the same project requirements with multiple software engineering firms, the quotes varied from $5K to $60K, and project timelines ranged from two months to twelve months. I discovered that both the very low and the very high bids often overlooked key aspects of the scope, lacked technical expertise, or missed robust processes.
- Self-doubt is real. Every day, I wake up excited to work on Hobbes. I get to work on a problem I find fascinating, free from the distractions of endless emails and unnecessary meetings. I get to learn something new almost each day. Yet, these are mixed in with moments of self doubt when I question whether I’m tackling the right problem or if weeks of effort have been misdirected. I’m still learning to embrace this self-doubt.
In the next couple of weeks, I will begin working with a software company to develop Hobbes, aiming for a late-summer / early-fall launch. Thank you for coming along on this adventure!