Hardware
Great report from Google and MIT. Co-designing ML models and the accelerators is an exciting research direction. This report introduces Coco…
Great report from Google and MIT. Co-designing ML models and the accelerators is an exciting research direction. This report introduces Coco, an agent platform TPU architects use for hardware and mode
Great report from Google and MIT. Co-designing ML models and the accelerators is an exciting research direction. This report introduces Coco, an agent platform TPU architects use for hardware and model co-design. Each decision depends on hundreds of gigabytes of fresh simulation sweeps that no model saw during pretraining. Coco registers every sweep in a normalized relational database, so each number an agent reports comes from a SQL query. Agents compose typed tools on their own, run recurring analyses such as comparing systems at matched execution configurations, and read the UI's navigation state as context. The near-term goal is halving the time architects spend setting up simulations and getting insights from them. Paper: https://arxiv.org/abs/2610.02376 Chat with Paper: https://academy.dair.ai/papers/coco-an-agentic-copilot-for-the-hardware-software-co-design-lifecycle-2610.02376
Related
- Coco: An Agentic Copilot for the Hardware--Software Co-Design Lifecycle
- Google’s AI agent platform takes pole position but work remains
- MDTransformer: A Hardware-Software Co-Design of Mode-Division Photonic Transformer Accelerator with Inverse-Designed Coherent Crossbar
- ViM-Q: Scalable Algorithm-Hardware Co-Design for Vision Mamba Model Inference on FPGA
Source: DAIR.AI (X) | 2026-10-06