Building a university-shaped ed-tech platform in public
What exists today and what's still on the drawing board
I sat down with the questions I wish someone had asked me before I registered another domain and called it a company. Here’s what I actually mean when I say I’m building Hearth & Code — in public, on a budget, with more architecture docs than revenue.
I walked away from a software career in 2020 for mental health reasons. I didn’t walk away from education, or from technology, or from the feeling that computer science ought to be part of how we grow up — especially now, when AI agents are changing what it means to work with and design software.
For a long time I was more the dreamer than the builder. I want to change that. Hearth & Code is where three things I care about coalesce: games, education, and the craft of building software. Not a content library. Not a bootcamp tunnel. Something closer to the depth I got at university — RPI, CS and Psychology — without the debt, and structured enough that progress feels like a real quest, with artifacts you actually earned along the way.
I’m not pretending that’s a trivial thing to build, but I’m starting anyway.
Who this is for
If you land on hearthandcode.dev at midnight, tired and stuck, I’m building this for you if you’re any of these:
The burnt-out dev who knows a bunch of tools but not enough about systems
The tutorial-hell veteran who’s got more courses under their belt than stars in the galaxy, but still can’t design a thing end-to-end
The eager newcomer who wants a foot in the door without pressure or being talked down to.
Learning doesn’t have to be lonely. I’d like a spot where we gather around something warm and work through hard problems together — not performing expertise, not grinding in silence.
What broke for me
I’ve done it all — MOOCs, bootcamps, online courses, and a full CS degree. The pattern was the same: good at teaching one piece of technology, weak at teaching you how to learn on your own and stay current.
Nobody walked me through how to hold a larger system in my head — components, interfaces, API design, architecture, tying theory to something shippable. I was trying to learn how to build. What I kept finding was instruction on how to use the tools required to build.
That’s the gap I’m aiming at. University-shaped depth. Practical spine. Personalized path. Not a generic curriculum stamped out for everyone.
What exists today (honestly)
I need to be blunt: there is no product yet.
What exists:
Brand and voice guides — the hearth isn’t just a decoration; it’s how the atmosphere should feel
Domains and social handles
Architecture and pedagogy docs — directions, not a shipped app
A knowledge archive (xanastros-archive) that ingests open-access material and tracks licenses — the corpus I want curriculum to grow from
Agent orchestration repos — how I’m multiplying a solo founder’s output
A landing page ready to deploy
What’s still a drawing: the gamified platform, the AI tutor, accounts, progress, assessment — all of it. If you’re looking for a signup flow that teaches you Python tonight, this isn’t it yet. If you’re looking for an honest build log, pull up a chair.
Why the hearth
The name matters. In the oldest version of us, people gathered around a fire and taught each other — stories, skills, warnings, hope. That time felt sacred to me when I imagine it. I wanted a brand that captures that essence: warm, safe, no judgment for what you don’t know yet.
Hearth & Code is not just a platform — it’s also a community idea. I don’t see enough developer spaces that combine real mentorship energy with solidarity — not leaderboard performance, not hot takes, just people getting better together.
What I’d get wrong with the AI tutor
Everyone has an AI tutor now. Mine would betray the project if it became generic — rehearsed explanations with no personality, no connection to your goals.
It should challenge you to think, not dump answers. It should know where you’re trying to go and let curiosity branch off the path when that’s useful. A rigid pipeline with no room to explore is just another course in a chat box, and there’s enough of those.
Building on a fixed budget
I’m not funded. I’m not paying myself. That changes every decision.
I won’t form an LLC until there’s recurring revenue or a grant that requires it. I self-host when it’s $0. I don’t take on monthly bills for vibes. I’ve formed companies before that cost money and produced nothing — I’m not doing that again.
Slow, for me, isn’t missing ambition. It’s a promise that this has to earn its way before I pay for paperwork.
How I’m different from Boot.dev and the rest
Boot.dev and Frontend Masters are good resources. Use them. I’m not trying to replace a focused backend path or a professional workshop library.
Where I’m placing my bet: personalization, theory alongside implementation, and gamification with actual depth — progress that reflects your journey, not a generic checklist. If you only want a fast path to React or Go, I’m not your first stop. If you want the why and the architecture and the path from concept to working code, I think Hearth & Code will eventually be the right choice for you.
What I won’t publish (yet)
I’m building in public because the process is worth documenting — especially for solo founders wondering if AI tooling makes this more possible than it used to be.
Some things stay private: agent orchestration that powers personalization, and core service layers that are the actual moat. Architecture decisions, bugs, pivots, and technical lessons — those I’ll share when they help someone without giving away the farm.
How I use AI (and what I won’t pretend)
I’m not going to shy away from the fact that I use AI tooling — heavily. Cursor, multi-agent workflows, research pipelines, draft generation — just to name a few.
For a solo founder with limited energy, I don’t view this as cheating. I view it as multiplying my productivity, effort, and output. I could never hope to design something this ambitious without the guidance, input, and support of AI tooling along the way.
Hearth & Code isn’t “ChatGPT with a game layer on top.” The tutor should adapt to you and challenge you, not offer generic, bland explanations. The curriculum should grow from pedagogical choices and a license-aware knowledge base — not whatever a model might have hallucinated this week.
My rule for anything public: AI drafts, I edit and review. I add personal touches and reshape the language until it matches my voice. If it doesn’t sound like something I would write or say out loud, it doesn’t go out.
My rule for the product: AI assists design and implementation, but I own the architecture, the assessment logic, and what learners actually experience.
You’ll see scaffolding in these posts. As someone who tends to be verbose and say in twenty words what could have been said in ten, that works well for me. You’ll also see the corrections — the thought patterns, the personal reflections, the decisions that go into building and marketing a product. I’m building a platform to teach people how to learn with AI, not how to get AI to do the thinking for them. I try to practice that in everything I use AI for.
That’s the transparency: I won’t hide the tools, and I won’t hide the human editor either.
What you can do
If this resonates:
Join the waitlist at hearthandcode.dev
Reply with one honest thing — what helped you learn, what didn’t, what you wish existed
I believe people want to learn. They don’t always know how. Your answers shape what I build first.
What ships next week
This post — live. The landing page — deployed at hearthandcode.dev with email signup and a link here. That’s the artifact. The platform is still docs and architecture; I won’t dress that up as a launch.
After that: one public artifact per week — a post, a deploy, a spec worth reading. If I miss a week, I’ll say why.
Why 500 subscribers
I’m aiming for 500 on the list before I treat audience as validated. Not because 500 is magic — because it’s enough signal that I should shift from “document the build” to “ship an MVP people can touch.” It also deepens my own commitment. I have low-energy weeks. Knowing someone is reading changes what I do with a Tuesday.
— Scott
Pull up a chair: hearthandcode.dev · build log on Substack