Why acceleration is important
As gameable credentials lose signal, genuine early understanding compounds, and acceleration should be about substance rather than grades.
Society has slowly been moving away from objective, verifiable skills towards more nebulous, subjectively-earned credentials. For a while, Ivy League schools became standardized test-optional and wholly dependent upon the opinion of admissions officers about the “holistic person.” Even after standardized tests were reinstated, there is still a widespread sense that there’s a gamification of extracurriculars going on. Grades have been steadily inflating over time. In New York City city-wide gifted and talented public schools, admissions now only require 1) a recommendation from a teacher and 2) winning a random lottery, no tests in sight.
As we discussed in our first article, verifiable skills are now the only high-signal measure of human skill in the age of AI. The danger with subjective admissions to traditional credentials is that they are all gameable with AI. Essays and homework can be completed with AI. Extracurriculars can be boosted with AI – suddenly there’s a bunch of student-run charities cropping up, self-published books by middle schoolers, etc. These traditional credentials are losing signal, hence the popular opinion that college isn’t necessary, that the Horowitz Andreessen school can skip select high school grads straight to “building,” that school grades don’t matter.
If traditional, gameable credentials don’t matter, what does? As we said before, verifiable skills are going to be more valuable than ever – chess competitions, programming contests, math competitions. Unlike traditional schooling, there are very few age gates. There are always competitions for younger and younger kids. Students can always demonstrate their competence, whenever they acquire it. “Building” will also economically matter – ownership of income-generating businesses. If AI democratizes having PhD-level domain-specific knowledge, and rewards having generalized decision-making and logic skill, then kids don’t need to wait until they’re 18 or 22 to become employed. This makes early capability much more productive and useful than before.
Dismantling the traditional age-gated credentials means that accelerated learning becomes more important, and should be more encouraged than before. But, acceleration should look different than it used to. Now, foundational skills that are the training ground for judgment and decision-making should be acquired for its own sake, instead of for a teacher-pleasing grade. If substantive learning is more important, that means genuinely understanding subjects early is more important than earning an A or skipping a grade. And if only substantive learning is important, that means genuinely understanding a prerequisite early gets you compounded advantages of genuinely understanding a subsequent subject earlier too. Now, students aren’t Goodharting their school tests by doing whatever will get the A’s with a particular teacher. They will have to either 1) enter verifiable skill contests to verify their genuine understanding or 2) evaluate themselves, with the only incentive being that they now fully own their consequences of not acquiring foundational knowledge.
This new paradigm should be viewed as freeing. Now, instead of catering to subjective evaluators, students can remove all the artificial administrative gates, and acquire foundational skills for the purpose of developing generalizable intelligence. Once a child’s mastery is there, there’s no reason anymore to stop them.