
i want to give you credits to make cool shit. plz take my credits, i beg of you.
✓devrel @modal

i want to give you credits to make cool shit. plz take my credits, i beg of you.


GTA 6 is out in 2 days. People call it the last pre-AI blockbuster. I've been excited about what AI-powered games might look like, so I built a small one on @modal: Sandbox Village. Every NPC thinks with an open model on our Model APIs and has its own Modal Sandbox to build in. Talk to them, or watch them talk to each other. Inspired by Smallville paper.

The winner of our “best use of Modal” challenge at the latest @HackIterate’s AI Research Hackathon! A general framework for recursively self-improving RL to prevent reward hacking and produce stronger policies.

this project caught my eye the moment they showed their trained LLM for Street Fighter III! x.com/ajhinh/status/21130368… amongst the auto-research submissions at the hackathon, i was particularly impressed by the generality and branching nature of their framework. more to come from these folks 👀

Over the weekend @mankydankybanky and I trained AIs to play Rocket League, Street Fighter, and Tetris, and won @modal’s challenge at SF’s AI Research Hackathon for it. We built a recursively self improving RL framework where AI agents supervise training, spot failures, and run experiments to fix them with the end goal of training stronger models with less time and compute. All we need is an arbitrary environment and objective, and our framework handles the rest. Train your own AI on our framework using the repo in the thread.

And by using shared endpoints, usage is billed per token! modal.com/docs/guide/shared-… Get up and running in minutes 🚀



Modal Endpoints are now available in @NousResearch Hermes Agent hermes plugins install modal

Benchmaxx with us this weekend at Vals Hacks! We will be on-site with compute credits and engineering support. Register here: vals.ai/hackathon

RT @modal: Modal's inference engineering team will be at @sgl_project SGLang Summit! Meet us on the expo floor the rest of this week and…

A new superhuman Polytopia bot arrives: "PLATO" by @JackYuan35 - 3.2M param CNN + transformer hybrid: 4 conv blocks into 6 attention blocks - Pure self-play PPO, 31B samples - 700 H100-hours on @modal Play here: polyzero.app

when i started climbing in 2021 at Movement SF, i always looked forward to being challenged by the sets put up by Summer, co-owner of CCC. now, i'm so excited to help sponsor their first ever competition, Ungraded! for those looking to compete, a $20k prize pool awaits...

:bufo-not-enjoying-the-security-apocalypse: :laugh-cry-scream: :party-dead:

ever wanted to hear what an LLM sounds like? turns out, they’re pretty good at controlling the synth models manually! here’s Ave Maria in 4-part harmony built with Kimi K3.1 Pro on @modal