I've also heard conflicting data on the natural immunity duration - many sources claim at least 3 months of protection, but I can't quickly find any that suggest immunity lasts no more than 3 months. Indirect evidence suggests immunity lasts longer, namely the persistently low re-infection rate - since the first wave is so long ago, you'd assume there was more reinfection by now if immunity lasted just 3 months, right? But maybe I'm reading this wrong; do you have a citation to back this surprising estimate?
Evidence is necessarily lacking, but at least there's no obvious cause for concern yet.
Vaccine immunity durability is not some kind of primary goal - it's several steps away from one. The risk we should be mitigating is the risk of continued pandemic spread, not the risk of short-lived immunity. That may seem like a pedantic point, but the thing is, it raises one very relevant question: how would our actions be different if we pessimistically assume vaccine-induced immunity only lasts (say) 6-9 months?
I'd argue it's likely that this makes very little difference to our actions at this point. At worst it means we'll need to scale vaccine production to a higher level, but even at that higher level the costs are trivial compared to the extra budgets already being spent to save the economy, which are themselves probably small fry compared to the actual economic damage; i.e. even in the "worst case" the costs are eminently doable; the US or EU acting alone could afford to finance vaccinating the entire world, and would even from an entirely selfish perspective probably be well served to do so - and of course, they don't need to, since other countries won't bother waiting for external help. (However, it perhaps does partially help explain Chinese vaccine diplomacy - not only is it truly helpful, it's also in their self-interest both epidemiological, and diplomatically).
But extreme pessimism isn't harmless; I think people start taking warnings less seriously when we cry wolf too frequently, and turn out to have been too cautious, i.e. wrong, in retrospect. There's little organizational benefit to assuming vaccine protection is short-lived, and it's a dishonest projection. We should be honest about the fact that we don't know how long immunity lasts, but certainly more than a few months, and potentially much longer; it could last for many years.
So sure; we shouldn't unnecessarily accept intolerable risks to society even when those are unlikely (but at least plausible) - but I don't think that's the case here.
But it's not even a binary: the most likely outcome (by far I think) is that strong immunity of the quality that keeps you even from becoming a symptom-less spreader for a few days won't be lasting, but infection will never again be as dangerous as it was the first time (or would have been the first time, without a vaccine). The immune system is a memory hierarchy.
This is plain wrong. Natural immunity for SARS-COV-1 was observed even 10 years after infection in some trials. Immunity against other coronaviruses also lasts longer than 2-3 months and actually there is strong indication that natural T-Cell immunity might last people much much longer and be sufficient on its own like it is also with other coronaviruses.
The only thing which is estimated to disappear after 3-6 months are antibodies in your bloodstream.
Plus having an immunity to one coronavirus doesn’t automatically buy you an immunity to another strain of coronavirus. We are already seeing several new strains.
Also having COVID-19 doesn’t automatically mean your body even develops an immunity. All sorts of caveats might prevent that from happening such as the infection being too mild to begin with (eg viral load).
And lastly, while on the topic of viral load, if your reinfection it too severe then you can still die from a reinfection even if you do have a immunity (immunity gives your body the resistance to fight it, but that doesn’t mean you’re invulnerable to the decease).
So while your points are valid, that doesn’t mean anyone who’s caught COVID-19 and recovered should be thinking they’re bullet proof (metaphorically speaking). At best, they could still be carriers. At worst they could still find their body unable to fight off any new infections.
But I do agree their risk is generally lower.
Also, natural immunity is only estimated to last around 2 or 3 months and offers no protection against other strains (some of which appear (statistically speaking) to affect younger folk more severely than the original SARA-CoV-2 strain).