Sakana AI just unveiled "TreeQuest," a clever new method that gets multiple AI models to work together like a well-coordinated team—outperforming solo models by 30%. Think of it like assembling a group of specialists, each chipping in their strengths, guided by a decision-making system (inspired by game-playing AIs) to tackle tough problems more effectively.

Dust, a promising AI startup, just hit $6 million in annual revenue by tackling a big frustration in enterprise tech: AI that doesn’t just talk but actually does things. Their agents automate real workflows—like pulling data, updating systems, or triggering actions—using Anthropic’s Claude models and a protocol called MCP. Basically, they’re making AI less chatty and more useful for businesses tired of hype without results.

A German research lab has unveiled a turbocharged version of the DeepSeek R1-0528 model that reportedly runs 200% faster than its predecessor. While details are still emerging, this breakthrough could significantly speed up AI processing for applications ranging from data analysis to real-time decision-making. The development has sparked lively discussion among tech enthusiasts, though some are waiting for independent benchmarks.

Editor’s Note: Faster AI models like this could unlock new possibilities in industries where speed is critical—think autonomous vehicles, financial trading, or emergency response systems. But as with any lab announcement, the real test comes when it hits the real world. If the claims hold up, this could be a game-changer for businesses relying on rapid AI computations.

Capital One is developing AI agents designed to boost car sales by mimicking the bank’s own decision-making processes. Their head of AI foundations shared details at a tech conference, suggesting these tools could streamline auto financing with smarter, faster assistance—almost like having a digital salesperson trained by the bank itself.

Big tech players like Atlassian, Intuit, and AWS are already prepping for a future where AI agents—not just humans—handle API calls and automate workflows. At a Women in AI event, their tech leaders shared practical insights from real-world AI agent deployments, hinting at how businesses should adapt.

Chevron slashed a major data processing task from 30 days down to just one by moving to the cloud—and they’re sharing the hard numbers to prove it. Their system now handles massive amounts of data across cloud and edge computing, keeping critical operations running smoothly.