New Tool Radar

  Computable Review: Buy GPU Compute for the Exact Weeks You Need As artificial intelligence continues to grow, access to powerful GPUs has become one of the most important parts of building and running modern AI systems. Training models, running inference, processing large datasets, and developing AI applications can all require substantial GPU capacity. Computable takes a different approach to this problem. Instead of treating GPU infrastructure only as traditional cloud rental, Computable presents itself as an exchange for physical compute , allowing users to buy GPU capacity for specific calendar weeks. What Is Computable? Computable is a GPU compute marketplace designed to give teams more flexibility when purchasing high-performance computing capacity. The platform currently focuses heavily on dedicated NVIDIA H100 infrastructure and lets users select specific weeks for their compute requirements. The company is part of Y Combinator's Summer 2026 batch and is liste...

  Decawork Review 2026: The AI Agent Control Plane for IT Teams AI agents are becoming part of everyday business operations. Marketing teams can build content agents, recruiting teams can create candidate-screening workflows, finance teams can automate reconciliation, and engineering teams can deploy coding agents. But as companies deploy more AI agents, a new problem appears: who controls all of them? An AI agent may have access to company data, APIs, internal applications, customer information, or other business systems. If every team creates agents independently, IT and security teams can quickly lose visibility into what exists, who owns it, what it can access, and what actions it is taking. Decawork is designed to address this problem. Instead of building another AI agent, Decawork positions itself as an AI agent control plane for IT teams . It provides one place to deploy, govern, monitor, maintain, and audit internal AI agents built across different tools a...