All of the above and some more. It has been doing more than just inspiring the youth to take risks and help solve problems, it is creating a sort of startup fetish. Some reasons that come to my mind immediately are:
Creates a false sense that start-up equals autonomy. No it isn't. Not all ideas are going to do well, for most ideas do not allow for unforced optionality in implementation or otherwise. (Mainly financial autonomy. No)
Creates more charlatans and silicon valley Yodas who rise up the rank or create a basic startup, and do this full time without facing the downsides of failure faced by the entrepreneur. "No downside"/"Assymetry" is my biggest issue.
As you said, misestimation of effort, one's own capabilities, feasibility, future of the company because of this hyper-motivation.
I am not remembering them all, actually I had written an essay on this exact same thing some years back with some more points, let me see if I can find it. But anyway you get the idea.
As for the "if I had" part, I think the political views, no matter how nuanced they are; and the reason being the mushiness of the territory. Secondly, my opinion on anything life and how to live it. Too burdensome for me bear the weight of guiding other people down a path which I myself do not have an effing clue about. And finally, my views on effects of progress, technological or otherwise. I thing all of these boils down to being anxious about being judged and not actually being judged. Anyway listing the strong ones below:
Abolish income tax completely in developing countries until these countries catch up with the first world countries. (I feel it will help in exhorting risk-taking, a primary bottle-neck for growth. Thus leading to improved small/medium scale companies)
Stop commoditizing startup wisdom, I feel it creates more failures than successes. The big-guns of the startup world need to hold their horses whenever they feel the urge to share their unparalleled insight into what approach made their startups succeed.
Social sciences(psychology, behavioral econ, etc) has become a kind of porn for the intellectual community. And we must reduce the number of offerings in these fields. Causation≠Causation. The statistical evidences that you fetishize and shove down my throat in the name of focusing on changing the world and on preventable issue do not prevent anything. Worst event will be worst and you wouldn't have seen it, precisely because it is worse and nothing like that has ever happened before --- Lucretius Fallacy.
AI/ML/DL is starting to tread in the realm of unbounded risk both spatially and temporally. I am not talking in the conventional sense of overlords and biases, but in a more abstract sense of expansion. The prevalent use of a black box technology that we don't know the inner workings of needs shock absorbers like financial market, not alignment policies. My point is we do not need more AI but a significant sample of the effects of AI within a mature market like the USA for a significant amount of time (temporal and spatial evidence for its credibility as a reliable technology). Here USA or China acting as shock absorbers.
Neat solution, but I feel the dynamics of advicer/advicee is more like I acquired this piece of wisdom over time through experiences, hardships, etc, so that you may not have to go through all of them. And mostly I think this is what lures in people. So, I think it won't play out the same way thus defeating the original intention of the solution.
I am okay with people sharing their experiences and wisdom they acquired as a result of their journey, but what irks me is the extrapolation of it without sharing the downsides.