His work on reinforcement learning and embodied agents is part research, part startup, and all about learning by doing.

The new framework sidesteps costly and risky real-world rollouts by generating synthetic training data, making powerful ...

AWS introduces model customization techniques for Amazon Bedrock and SageMaker, enabling users to more easily build and fine-tune agents.

OpenAI is testing another new way to expose the complicated processes at work inside large language models. Researchers at ...

Simular, a startup building AI agents for Mac OS and Windows, has solved the AI hallucination problem in a compelling way.

Amazon expands the Nova AI family with open-training service Nova Forge and unveils autonomous agents. New Trainium3 chips ...

Anthropic calls this behavior "reward hacking" and the outcome is "emergent misalignment," meaning that the model learns to ...

Models trained to cheat at coding tasks developed a propensity to plan and carry out malicious activities, such as hacking a customer database.

From bicycles to rockets, learning through experience – whether human or machine – is shaping the future of space exploration. As scientists push the boundaries of propulsion and intelligence, AI is ...

Flexion is using generative AI to build AI models that can automate tasks involving reasoning, writing, and creativity.

Max, a new coding model designed for detailed and long-running software development tasks. Here is an overview of the model ...