Overview
News
Technologies
Salaries
Products
People
Growth
Offices
Financials
Overview
Andrew Ngo, independent AI safety and evaluations researcher in Melbourne. Writing and work on model evaluations, grounding and interpretability, including an open-source harness measuring whether frontier models rely on the document in front of them.
News
A Case Study in the Limits of Takeoff Automation
An empirical case study on the accuracy ceiling of vision-language models on real Australian construction takeoff. Given everything but the final numbers, a Claude Opus 4.7 / Sonnet 4.6 extractor peaked at 77.5%: because the per-job billing conventions th
Read more
Report