Rutgers CS - web game dev of 5M+ Players - Currently Agent Puppeteer
New Jersey
Joined April 2018
- Tweets135
- Following177
- Followers329
- Likes299
Pinned Tweet
I took @karpathy’s autoresearch loop and applied it to game development, here's what it built last night:
Agents read player data → plan improvements → spin up new git branches in a game evolution tree → ship playable HTML5 variants → repeat forever.
The system optimizes toward the game variant most likely to be chosen by players.
Live leaderboard + playable games at autogamestudio.ai
MIT open source: GitHub.com/centipede5/autoga…
I feel like we're just beginning to unlock the potential of this beyond ML, soon we'll have self-improvement loops in every product. Very excited to see what comes next.
Well played @sama, I got way too used to burning through 5.5 xhigh fast mode on the 10x codex promotion and now I'm gonna have to upgrade to pro
The whole "flagged for possible cybersecurity risk" is bullshit right? They just don't have enough compute to serve 5.5 for the exploding number of codex users. I've been getting this every few hours for the most benign SWE work ever.
Autogamestudio has been absolutely insane after gpt-image-2 dropped, custom sprite animations are finally feasible
I took @karpathy’s autoresearch loop and applied it to game development, here's what it built last night:
Agents read player data → plan improvements → spin up new git branches in a game evolution tree → ship playable HTML5 variants → repeat forever.
The system optimizes toward the game variant most likely to be chosen by players.
Live leaderboard + playable games at autogamestudio.ai
MIT open source: GitHub.com/centipede5/autoga…
I feel like we're just beginning to unlock the potential of this beyond ML, soon we'll have self-improvement loops in every product. Very excited to see what comes next.
Huge productivity hack for vibe-building:
Have your agent build a simple streamlit / chartjs admin dashboard for whatever you’re working on.
Visual debugging >> logs + manual testing.
~60% of the time it exposes a broken or weird architectural choice from a 5 second scan
Very impressed with o3's performance at @kaggle's game arena. Despite being almost a YEAR old it still beat all but google's latest Gemini 3 model in both chess games. Not to mention a very strong performance in poker so far.
Just goes to show how much of a revolution thinking models where for structured tasks.
Soon it won't matter how many mac minis you buy
Tomorrow I'm running 30 - how'd I do @steipete?
The @openclaw is great - yet there isn't an efficient and consumer friendly way to create and manage 1+ instance (besides ever more mac minis, ofc). Working on something cool for this
Predictions for 2026:
- Rise of "Software as Content". Custom apps will be roughly as cheap to make as images/videos, so they'll begin to be used like them.
- No "AI bubble" will burst in the stock market
- Elon Musk is at least arguably a trillionaire sometime in the year
- Brands will waste a lot of money on AI ads / content, alienating customers in the process
Centipede5 retweeted
The @ilyasut episode
0:00:00 – Explaining model jaggedness
0:09:39 - Emotions and value functions
0:18:49 – What are we scaling?
0:25:13 – Why humans generalize better than models
0:35:45 – Straight-shotting superintelligence
0:46:47 – SSI’s model will learn from deployment
0:55:07 – Alignment
1:18:13 – “We are squarely an age of research company”
1:29:23 – Self-play and multi-agent
1:32:42 – Research taste
Look up Dwarkesh Podcast on YouTube, Apple Podcasts, or Spotify. Enjoy!
The most annoying problem by far with current agent architectures are their data accessibilities. If you want to use anything outside foundation models, you either get stuck buying expensive subscriptions to stuff you don’t use or have to settle for crappy free alternatives.
Solution soon.
Centipede5 retweeted
I’m building a tool that finds the cheapest possible grocery list to hit your macro targets. Started with Aldi, and planning to add more stores as I go
Try it here: macros.cheap
The success of "Reasoning models" and "AI agents" both point to one thing—AI needs more structure to unlock higher capabilities. Base models are impressive, but the real breakthrough will come from platforms like @simNGMI where AI can collaborate and iterate