Steven Dillmann

AI

A post also thanked participants in a fireside chat at the October 27 launch event.

2601234567890123456789

AD

AR

SD

12 Sources, 10d ago, updated 10d ago

AI

Updated model targets coding, reasoning, and long-running agentic tasks.

Technology

Scientists from labs worldwide shared their actual research workflows to test AI agents. The result? Even top models solve just 30 percent of these authentic tasks.

8401234567890123456789

SA

BB

SD

9 Sources, 73d ago, first seen 73d ago

Technology

It assesses AI agents on scientific research tasks across multiple domains.

128012345678901234567890123456789

TW

LS

SD

3 Sources, 74d ago, first seen 74d ago

AI

Stanford-led effort evaluates AI agents on 70 expert-contributed research workflows across scientific domains.

7001234567890123456789

AK

WS

SD

7 Sources, 74d ago, first seen 74d ago