I'm pretty sure!
You're right about the initial natural history of HIV: the virus is incredibly active during the initial infection (this is due to the time it takes to develop an efficient immune response). The virus definitely can be detected in these scenarios, by picking up on a certain part of a protein the virus makes, but what happens if someone doesn't go to their GP? Or a sexual health clinic? Because they don't get symptoms, are particularly stoical, or just happen to suffer from them less severely? What happens if the GP misses it? The symptoms will go away, like you said, and this person won't be aware they have HIV. If we test them again, we won't pick up on it! This is because the antibodies (which is what we test after the initial stage) take a while to reach a detectable level in the blood: now for most people, it happens pretty quickly (2-4 weeks), and you can detect those antibodies in the blood (you can see these figures, all explained in understandable terms,
here). But for some people, and nobody really knows why (but we know taking post-exposure medication makes it more likely!) it takes longer! A
vast majority of people will have detectable antibodies after 3 months (Wikipedia says 97%), but a very small percentage of people take much longer than that to be detectable (even cases in the literature as long as a year, 10 months
here, not sure about a percentage who have this, but it's small - <1% surely). It's these people that the blood donor service is worried about (since the rate of transmission from HIV-infected blood is incredibly huge), and it seems to be a method of hedging their bets (especially when you look at it combined with the low transmission risk of a single MSM exposure). There aren't any reported cases (or at least that I know of!) of HIV being undetectable for more than a year using modern methods, and certainly nothing I know to say that there's no difference in detectability between 1 year and 10 years, hence why I think extending the ban past that would be unreasonable, and took issue with the bit I highlighted.
I think the more pressing issue (rather than, should the ban be 3 months, 6 months, 1 year, etc.) is the inconsistency. An MSM male in a monogamous relationship who practises safe sex is at a lower risk than a promiscuous heterosexual man or woman, but the former is restricted from donating, and the latter isn't. The tough question to me is: how do we adjust the process to balance the risks properly?