Venture capital technology is mostly about replacing expensive processes with human input by cheap substitutes that are 75% as good but hoping scale will compensate. Sometimes 10x at 75% is really is better than 1x at 100%. Sometimes it’s not, because 75% isn’t good enough, and it might as well be 10x at zero.
Education is closer to the latter. VCs think they can remove the humans and the scale and cheapness (the latter in theory, at least) will compensate, but people are social animals, and students lose trust in the state when it wastes their time, so education by machine turns out not to work well at all, despite the fervent belief held in a few hundred Bay Area tech companies that it should.