There's a big problem with Altman's working out here. The statistical effect for 1:1 tutoring is 30+ years old, only applied to two contexts with small samples, referred to *human tutoring* not AI tutors, and - the kicker - *has never been replicated* 3/ www.educationnext.org/two-sigma-tu...
An experimental intervention in the 1980s raised certain test scores by two standard deviations. It wasn’t just tutoring, and it’s never been replicated, but it continues to inspire.
AI tutors are imaginary tech built on pseudo-psychological simplifications of debunked psychological statistical effects, glossed with entrepreneurial confidence and masses of investor speculation 4/
Learning scientists do care about this and the education research community is leading rigorous evaluation of tools that use advanced computation, from psychometrics to AI. See rori.ai/learning-out... for an early example.