My life & career
AI assistance can be useful, but critical judgment and expert review remain essential in demanding tasks.
THE YOTTABIT SCALE INDEX / Automated scientific discovery
38.8%
AI science agents still have a long way to go.
A leading science agent scored 38.8% on a particular evaluation versus 83.5% for PhD experts, as summarized in the 2026 AI Index.
2026 report · Stanford AI Index
THE YOTTABIT PERSPECTIVE
What does it mean to me?
Big change is fascinating. Its implications are what matter.
My business
Organizations should test systems on complete workflows and real failure modes, not just impressive examples.
My industry
Scientific labs and regulated industries need reproducibility, auditability and human oversight.
FROM JIM CARROLL’S WORK
Jim’s innovation approach combines experimentation with guardrails and evidence of outcomes.
Meet the futurist behind YottaBit ↗WHAT COULD I DO MONDAY MORNING?
Before automating an expert task, benchmark the complete result against a skilled practitioner.
THE EVIDENCE BEHIND THE STORY + METHODOLOGY & SOURCES
What the number really measures
Scientific agent benchmark vs PhD baseline, expressed as %.
The research foundation records: PaperArena 38.8% vs 83.5% expert baseline, 2026 chapter.
Where the claim has limits
Benchmark-specific; measures tasks not scientific productivity
Comparisons and historical rates must be independently checked against definitions and original measurements before an audited chart is published.