October 6, 2024

Tanagram Updates: October 2024

I expected to feel a revelation when I started working on Tanagram full-time, but that hasn't seemed to have happened. The motions and day-to-day experience felt the same as they'd been when I was still full-time at Stripe, although the subjects of my thoughts changed from the esoteric sequencing of cash movement to existential questions like "what should my product actually look like?" and "is it too early to hire someone?"

I do notice new feelings though; the rollercoaster is real. One day, I got feedback that some mocks I'd been working on, about which I was very excited earlier that day, didn't actually seem useful. I went home feeling dejected. The next day, I mocked up some other ideas, and got a 5¹, and I had to double-check that I was seeing it correctly. Earlier in the week I was wondering if it was too early for me to hire someone, given that I didn't know what the product should be. By the end of that week, I almost felt like I needed someone immediately to help build a few different promising ideas.

So maybe the change in feelings isn't a revelation, but more of a gradual fade. I'm ok with that.

I spent much of September doing customer discovery calls — chatting with developers across a range of seniority and company sizes to learn about the problems they run into while working with code. After over 60 conversations so far, patterns are emerging: developers struggle with disparate architectures, understanding the sprawling impacts of a change, tracing and navigating across services and repos, understanding unfamiliar code, and performing syntactic migrations/refactors.

Over the past week, I drew up some ideas attempting to address some of these problems. I've included the promising ones below. I'm curious to know how useful each of these would be for you in your work, on a scale of 1 = don't care, to 5 = need this now. Hit reply and let me know.

I also spent a bit of time thinking about what I'm looking for in a founding engineer, and I think I have a decent idea (roughly, someone with 2–4 years of experience building "real" software (outside of class projects), with experience or at least a keen interest in developer tools, editors, compilers, etc). If you know someone who'd be a good fit for the role, please let me know.