in San Francisco.
I build
serious tools and silly hacks, mostly at early-stage startups.
Recently at
Daybreak Health. I'm often
found at @kkuchta@ruby.social.
I’ve noticed two things recently:
So: epistemologically speaking, I think my best path towards having a well-informed opinion about what AI-based coding is and is not good for is to “Get Gud” as the kids would say.
kttt.io is my first pass at that.
One of the coolest things about working at a tiny startup is getting to make foundational technical decisions and seeing how they play out. A couple years in startup time is as long as a life-age of the earth like a decade in normal business time. It lets you get feedback on how those decisions worked out as the company + team scale.
I was the first eng hire at Daybreak Health. 3 years later I was managing a team of 8 in a 70-person company. Here are the decisions I’m glad of and here are the ones I regret.
How do you build software that will last more than a decade with no maintenance?
So, it turns out that css background-images don’t get loaded until the relevant selector is triggered.
Many people might say “neat!”
I used it to build a bi-directional CSS-only async chat.
I challenge you to reliably tell the difference between AWS and Totes-not-amazon.