Sabitlenmiş Tweet
Introducing Flash.
An open source Python SDK for building cloud-native AI apps. Define your hardware, functions, and dependencies in code. No Docker, no config files.
github.com/runpod/flash
English
Runpod
645 posts
Runpod
@runpod
AI Infrastructure developers trust 💜 https://t.co/PWVZocfp6D
Katılım Mart 2022
8 Takip Edilen7.7K Takipçiler
We're the compute partner for OpenAI's Parameter Golf challenge. Need credits? Apply at the link below!
OpenAI@OpenAI
Are you up for a challenge? openai.com/parameter-golf
English
Runpod is @OpenAI 's infrastructure partner for Parameter Golf, the first challenge in the Model Craft series.
Train the best language model that fits in a 16MB artifact in under 10 minutes on 8×H100s.
Together with OpenAI, we’re distributing up to $1M in credits across the challenge period to support experimentation and broaden participation.
Enter the challenge and request credits 👇
English
Introducing Flash.
An open source Python SDK for building cloud-native AI apps. Define your hardware, functions, and dependencies in code. No Docker, no config files.
github.com/runpod/flash
English
Runpod retweetledi
The brilliance of @karpathy is being able to distill vastly complex concepts and make them simple to understand and implement at a small scale.
All it took was Claude Code and $10 on @runpod to spin up a single H100, and I had a world class ML researcher working on autopilot.
I'm taking the general concept of autoresearch and applying it to an inference pipeline I've been working on (no GPU needed thankfully). Everything is so fun now.
Andrej Karpathy@karpathy
I packaged up the "autoresearch" project into a new self-contained minimal repo if people would like to play over the weekend. It's basically nanochat LLM training core stripped down to a single-GPU, one file version of ~630 lines of code, then: - the human iterates on the prompt (.md) - the AI agent iterates on the training code (.py) The goal is to engineer your agents to make the fastest research progress indefinitely and without any of your own involvement. In the image, every dot is a complete LLM training run that lasts exactly 5 minutes. The agent works in an autonomous loop on a git feature branch and accumulates git commits to the training script as it finds better settings (of lower validation loss by the end) of the neural network architecture, the optimizer, all the hyperparameters, etc. You can imagine comparing the research progress of different prompts, different agents, etc. github.com/karpathy/autor… Part code, part sci-fi, and a pinch of psychosis :)
English
openclaw for training robots...? (not exactly... will share more updates this weekend…)
next, excited to play with @theworldlabs
ps,: @runpod you beauty, what would i do without you!!
English
Runpod is officially HIPAA and GDPR compliant. Healthcare orgs and EU companies can now train and deploy AI models on our GPU cloud with enterprise-grade security protections for sensitive data. 🔒
🔗 runpod.io/press/runpod-a…
English
Runpod retweetledi
AI teams don’t need “cheap GPUs” - they need infrastructure that lets them move now.
Runpod is an AI infrastructure platform built for speed and on-demand access:
↳ Spin up serverless endpoints for inference and APIs without managing infra
↳ Multi-node instant clusters available on demand for large models (not weeks of provisioning)
↳ API-first workflow that feels like it was designed by developers for developers
For researchers and startup teams under deadlines:
↳ No buying hardware or waiting on traditional cloud queues
↳ Launch experiments, RFP prototypes, or production inference fast
↳ Pay-as-you-go so you don’t commit before you validate
It scales with your project:
↳ Start small for testing and experimentation
↳ Access a wide GPU range (5090s, A40s, H100s) as needs grow
↳ Move to serverless auto-scaling when usage spikes
↳ Skip ops overhead and focus on shipping models
The real value isn’t GPUs - it’s removing the friction between idea and deployment.
Worth exploring for your next model or inference pipeline?
Check it out here: fandf.co/4qU2Ab4
—
Thanks to @runpod for sponsoring this post.
English
Serverless vs. Pods — how to pick the right deployment model for your AI workload 👇
Serverless: Scales to zero, pay-per-second, no infra management. Best for stateless inference and bursty traffic.
Pods: Dedicated GPUs, full control, persistent storage. Best for training, fine-tuning, and interactive dev.
The move? Build in Pods. Scale with Serverless.
Teams fine-tune in Pods, then containerize and deploy inference to Serverless for production.
Match the model to the workload, and don't default to one approach.
Follow us for weekly AI infrastructure breakdowns 😉
English
Runpod retweetledi
Learning a dance or workout from videos is tough. It's passive, there's no feedback, and you don't know something's gone wrong till your back starts paining.
So at @hackwithtrees, we built JiggleWiggle to solve this problem. It can convert any YouTube video or live Zoom session into an interactive coach.
Built on @Render using @HeyGen @Zoom @Runpod @OpenAI @Modal
English
Introducing Lab: A full-stack platform for training your own agentic models
Build, evaluate and train on your own environments at scale without managing the underlying infrastructure.
Giving everyone their own frontier AI lab.
English