Steven Dillmann

State-of-the-art VLMs struggle to execute scientific research workflows, revealing a critical gap in their ability to interact with specialized software and produce verifiable results.

AI agents can now make scientifically defensible claims by embedding methodological rigor directly into the research process, drastically improving analysis accuracy.

Training data diversity is the secret sauce that boosts agentic model performance, with OpenThoughts-Agent achieving a notable accuracy leap over existing benchmarks.