CodeTrain
AI coding training on your own codebase
Visit Website
July 16, 2026
Launched an AI tutor that refuses to write code. Bootstrapped, $0 marketing, day 3.
-
2
The idea becomes interesting the moment it refuses to optimize for speed. I'd keep validating whether engineering teams are buying better learning outcomes or lower review overhead. If you can consistently reduce the cost of onboarding and code review, that's a much stronger business than simply being an AI tutor.
5 Comments
One thing I like here is how quickly the bottleneck changed once you had real users.
Before launch, the obvious question is "How do I get people in?" After launch, the evidence can point somewhere completely different. That's one of the few advantages small products have—you can let reality replace assumptions before you've invested months optimizing the wrong constraint.
That is a good way of thinking about it, and I am trying to handle this more like a love project than a commercial product, even though it has huge enterprise/teams value
That's an interesting tension. Sometimes keeping something as a love project protects exactly what makes it valuable, while enterprise demand can gradually pull it toward a very different product.
Curious how you're thinking about that trade-off right now.
A deep question for sure, and the genuine answer is probably no idea! lol I honestly don't mind solving that bridge once I hit it. I will say, I am passionate enough about helping others use AI in ways that is good for your brain that if I could replace my day job with taking car of and maintaining CodeTrain it would be like a dream job for me.
That’s actually the interesting part — not having the answer yet is probably where the real tension starts.
I’d be interested in continuing this conversation outside the thread. What’s the best email to reach you on?