Another benefit of heterodidacticism, related to your (1), is that it provides more of a sanity check: it helps keep you from talking yourself into crankery/craziness and helps you understand why the normies are normie. The large fraction of autodidacts who are cranks should give any aspiring autodidact caution. It also highlights the importance of the viewpoint diversity you mention: without that, heterodidacticism becomes epistemically closed cultishness that socially reinforces rather than countering crankery, as we tragically see with QAnon and the like.
I think it's also worth distinguishing between 1:1 tutoring and group learning as varieties of heterodidacticism. Tutoring has much less of the downsides you list, and a lot of distinct upsides (and note that professor office hours informally count as tutoring, and you can think of chavruta as a sort of mutual tutoring too). I am not sure whether I believe Erik Hoel's argument that the decline of aristocratic tutoring is why we don't have as many geniuses these days-- I am not even sure I believe the premise about the decline in genius!-- but there is enough substance to his arguments to make me think we should, at the current margin, do much more tutoring: https://erikhoel.substack.com/p/why-we-stopped-making-einsteins
As an engineer and inventor, I think that hetero will lead to more creative ideas and solutions. The most creative group will include a group of diverse masters plus a group of curious neophytes who are eager to ask questions. The neophytes will push the masters to consider possibilities that would otherwise be missed.
It seems like there's also a motivational advantage to each approach. Other people who you love and/or respect can motivate you to engage seriously with a text that you might have passed over otherwise—while the autodidact gets to remain constantly engaged by the question that interests her (rather than something, possibly less motivating, that is externally imposed).
Another advantage of heterodidacticism can be a sense of scale. The autodidact is in danger of both monomania and also just getting lost, never putting anything together in a coherent way. But with other people around, you have some common questions or texts to orient your search around, which can be a fruitful amount of structure.
But it ultimately seems a little strange to reduce the comparison to a question of advantage; each mode has goods that are internal to it, that don't just cash out as how much you learn or produce. Friendship, or even a sort of Arendtian politics, can be part of the good of heterodidactic learning. And autodidacticism probably has its own internal good, something like unfettered intellectual exploration.
Another benefit of heterodidacticism, related to your (1), is that it provides more of a sanity check: it helps keep you from talking yourself into crankery/craziness and helps you understand why the normies are normie. The large fraction of autodidacts who are cranks should give any aspiring autodidact caution. It also highlights the importance of the viewpoint diversity you mention: without that, heterodidacticism becomes epistemically closed cultishness that socially reinforces rather than countering crankery, as we tragically see with QAnon and the like.
I think it's also worth distinguishing between 1:1 tutoring and group learning as varieties of heterodidacticism. Tutoring has much less of the downsides you list, and a lot of distinct upsides (and note that professor office hours informally count as tutoring, and you can think of chavruta as a sort of mutual tutoring too). I am not sure whether I believe Erik Hoel's argument that the decline of aristocratic tutoring is why we don't have as many geniuses these days-- I am not even sure I believe the premise about the decline in genius!-- but there is enough substance to his arguments to make me think we should, at the current margin, do much more tutoring: https://erikhoel.substack.com/p/why-we-stopped-making-einsteins
As an engineer and inventor, I think that hetero will lead to more creative ideas and solutions. The most creative group will include a group of diverse masters plus a group of curious neophytes who are eager to ask questions. The neophytes will push the masters to consider possibilities that would otherwise be missed.
It seems like there's also a motivational advantage to each approach. Other people who you love and/or respect can motivate you to engage seriously with a text that you might have passed over otherwise—while the autodidact gets to remain constantly engaged by the question that interests her (rather than something, possibly less motivating, that is externally imposed).
Another advantage of heterodidacticism can be a sense of scale. The autodidact is in danger of both monomania and also just getting lost, never putting anything together in a coherent way. But with other people around, you have some common questions or texts to orient your search around, which can be a fruitful amount of structure.
But it ultimately seems a little strange to reduce the comparison to a question of advantage; each mode has goods that are internal to it, that don't just cash out as how much you learn or produce. Friendship, or even a sort of Arendtian politics, can be part of the good of heterodidactic learning. And autodidacticism probably has its own internal good, something like unfettered intellectual exploration.