We're excited to partner with @Cognition_Labs @Mercor_AI @CoreWeave and @AnthropicAI to host an inference-time compute hackathon, featuring >$60K in cash prizes and >1 exaflop of free compute.

We believe today's models significantly underutilize inference-time compute and search. With better algorithms and much faster hardware, we can push this scaling law by several orders of magnitude. We're particularly interested in research that accelerates parallel reasoning, high batch speculative decoding, text diffusion, and adaptive scaling.

Introducing Oasis: the first playable AI-generated game. We partnered with @DecartAI to build a real-time, interactive world model that runs >10x faster on Sohu. We're open-sourcing the model architecture, weights, and research. Here's how it works (and a demo you can play!):

You can play Oasis here: oasis.decart.ai Learn more about the model and partnership here: etched.com/blog-posts/oasis

Tens of thousands of people are using Oasis, the world's first AI-generated game. Try it yourself at oasis.decart.ai! It's trained on purely open-source data from VPT, OpenAI's open-source Minecraft dataset (MIT license)!

GPUs aren’t getting better, they’re just getting bigger. In the past four years, compute density (TFLOPS/mm^2) has only improved by ~15%. Next-gen GPUs (NVIDIA B200, AMD MI300X, Intel Gaudi 3, AWS Trainium2, etc.) are now counting two chips as one card to “double” their performance. With Moore’s law slowing, the only way to improve performance is specialization.‍ The economics of scale are changing Today, AI models cost $1B to train and will be used for $10B in inference. At this scale, a 1% improvement would justify a $50-100M custom chip project. ASICs are 10-100x faster than GPUs. When bitcoin miners hit the market in 2014, it became cheaper to throw out GPUs than to use them to mine bitcoin. With billions of dollars on the line, the same is happening for AI.