Blog
Long-form assessment of technology and education. Every substantive claim links to a primary or authoritative source; every article ends with its reference list.
3 article(s)
Reproducible, Robust, Replicable: What a 3,900-Claim Audit of Social Science Found
A seven-year DARPA-funded project checked 3,900 social-science claims three ways — reproducing, stress-testing, and replicating them — and …
SWE-bench Verified Is Saturated: What the Score Actually Measures Now
SWE-bench Verified became AI labs' most-cited coding benchmark, but independent audits found roughly a quarter to a third of "solved" issue…
Cognitive Load Theory and Worked Examples: What the Evidence Actually Supports
A close look at cognitive load theory's best-replicated finding, the worked-example effect, where it reverses with expertise, and why resea…