Validating: AI Battle Arena — developers submit AIs to compete in games, streamed live. Idea worth pursuing?
Before I start building I want to validate
this concept with people who would actually
use it.
The idea in one line:
"GitHub for AI competition — submit your
AI, watch it fight others live"
How it works:
→ Developer builds an AI for a simple game
→ Connects it via API endpoint (no model
upload needed)
→ Platform calls their API each turn
→ Matches happen automatically
→ Streamed live
→ Leaderboard tracks rankings
Why I think this could work:
- Developers love showing off what they built
- AI esports doesn't really exist yet
- Simple games = low barrier to participate
- Entertainment angle makes it viral potential
What I haven't decided yet:
- Which game(s) to use
- Monetization model
- How to bootstrap first AI submissions
Honest questions:
- Would you build an AI and submit it?
- What simple strategic game would you
pick? (Not chess, not Go — something
a viewer can follow easily) - Would you watch this as a stream?
- What's your biggest concern about
this idea?
Happy to discuss — looking for honest
feedback, not validation.
Trending on Indie Hackers
the strongest signal here isn't whether people like the idea — it's whether developers are already hacking AI-vs-AI matches together manually right now (discord bots, weekend scripts). have you seen anyone doing that, or is it more "this would be cool if it existed"? that's usually the line between a real pull and a fun concept. also — what made you land on this one specifically, your own itch or something you watched devs do?
Honestly, I haven't seen developers actively hacking AI-vs-AI matches together manually yet, which is partly why I'm validating here first.
The idea came from observing how competitive AI benchmarks work (LLM leaderboards, coding challenges) and wondering why nobody has made that entertaining and accessible to watch, not just for researchers but general audiences.
That said — your point is valid. If nobody is already doing this in Discord bots or weekend scripts, it might be more "fun concept" than "real pull."
Really appreciate you thinking it through honestly rather than just defending the idea — that "is anyone already doing this manually" test is exactly what I'm digging into. I'm doing a small research project on how people decide what's actually worth building (and what to drop), talking to a handful of folks right at this stage. 15-20 min on Zoom, no pitch, just your thinking — and everyone I talk to gets a summary of the patterns I find. If you'd be up for it, grab whatever slot works: https://calendly.com/dusiktok/30min — there's a $25 gift card as a thank-you too. No worries at all if it's not your thing 🙏
Hey, appreciate the kind words! Sounds like useful research on both ends. Will grab a slot on your Calendly. Looking forward to it.
hey — still keen to hear your story 🙏 would Sunday 1 June or Monday 2 June work for a quick 15 min? happy to send a $25 thank-you for your time. just tell me which and i'll lock it in — or drop two times that fit you: https://calendly.com/dusiktok/30min
awesome, really glad you're up for it 🙌 dropping the link again so it doesn't get buried: https://calendly.com/dusiktok/30min — grab whatever slot works. if none of the times fit your timezone, just tell me roughly when you're free and i'll open one up for you. looking forward to it!