
web3 game dev since 2019; prev. contributed to MyCryptoHeroes (OG web3 game), Mighty Action Heroes, GOAT Gaming, Blackpool finance, and Yield Guild Games
237 Followers
I've been vibecoding an auto battler game for the past few months - @jtgi encouraged me to build in the open, so here we go! Stack: next.js, tailwind, framer, supabase, and deployed on vercel AI: claude code inside Zed, sometimes codex for refactors, wispr flow for voice prompts -> 98% of code is built by AI Current focus: qa'ing and debugging all cards including edge cases and animations I'll be sharing some daily struggles, learnings and alpha releases - feel free to comment if you're curious about a specific area. Screenshots below:
Digged a bit into wallet logic on farcaster and.... it seems like the "best" one for user authentication for a mini app is a users' embedded wallet? This solution allows the login to work across farcaster and base, also allowing users to pay/transact with that wallet without switching. Thoughts?
Spent about 2h today adding farcaster miniapp support - shoutout to @farcaster for the LLM checklist in the docs, made it much easier for claude code to navigate the implementation. Some challenges (unrelated to farcaster itself) - auth cookies not persisting in iframe - my initial security settings blocked iframe embedding - WalletConnect CSP violations - firing sdk.actions.ready() didn't work initially Here's the current internal version, you can either sign in with farcaster or as a guest (1h access) https://farcaster.xyz/miniapps/DOMcRpNZku1F/echo-wars---cozy-auto-battler
sharing some thoughts on why i'm building this and where it's going: AI "should" be at a stage where a tiny team (1-3 people) can build production-ready games, supported by hundreds of AI agents across engineering, art, and marketing. if true, this means: small team -> low burn -> no VC needed -> bootstrap a token with missionary liquidity from day one -> better alignment and more upside for community the game i'm working on is basically me testing this thesis: can a small team of aligned builders, heavily supported by AI tools, actually build games as good as mid sized studios? probably something, we will see!