Okay, SARS-CoV-2 is pretty different from SARS-2003 ("~76% amino acid identity in the spike protein"), this might be the reason it won't work. OTOH, I don't know how different HCoV-OC43 is from both SARS strains.

Two facts:

1. HCoV-OC43 (one of human coronaviruses causing common cold) can generate cross-reactive antibodies against SARS.
2. Immunity to HCoV-OC43 appears to wane appreciably within one year.

Here's the paper which mentions both of these facts. (The actual paper is not important, I expect these facts to be well-known to coronavirus researchers, if the paper itself is not terribly mistaken and if I haven't misread anything.)

Even if cross-immunity is mild, won't it make sense to intentionally infect people with HCoV-OC43? Downside seems quite small compared to the number of deaths, and intuitively it seems that "mild cross-immunity" = "less severe SARS-CoV-2 cases", which is extremely valuable.

I notice I'm confused, since these facts should be well-known to pretty much everyone who's working on the vaccine. What's the explanation for why it's not a good idea?

Possible explanations, but I'm probably missing something:

1. Vaccines which cause the actual illness are considered unethical. (Probably not? I don't expect humanity to be that stupid.)
2. Mass-producing HCoV-OC43 virus is too hard for some reason. (Possible? I don't know much about vaccine production, and I'm clueless about whether it's even possible to mass-produce and store a "live" virus; but this seems solvable through organized infection parties, etc.)
3. Researchers or medical organizations don't want to rely on expected utility. Related hypothesis: time and productivity wasted by infecting many people with HCoV-OC43 is too valuable, and infecting everyone with HCoV-OC43 at the same time would hurt economy too much. (I don't believe this, but I haven't really tried to estimate this. If the alternative would be "wait for the real vaccine which is just around the corner", then yes, let's wait, but if the alternative is waiting for 12-18 months, then it doesn't feel right.)
4. Maybe I don't understand what "mild immunity" means and it's not that valuable of a perk to intentionally cause it? (But the same paper I quoted talks about HCoV-OC43 importance for predicting future SARS-CoV-2 outbreaks.)
5. Maybe being infected with HCoV-OC43 is too risky because getting two viruses at the same time is dangerous? Or because it would confuse the situation and complicate diagnoses of the real SARS-CoV-2? (Maybe... If everyone is sick with common cold then it would help SARS-CoV-2 to spread since everyone would be sneezing and coughing. But it also seems like a question of good timing and at least worth considering.)

So, what am I missing here?

Yes! We have an English club each Saturday at 5 PM.

Whoa, for some reason I thought that LTF fund is not relevant to us, but looks like I was wrong. Thank you!

For context: in the last few months I applied for two CEA grants.

1. Community Building grants (in December, outside of a funding round, so they warned me that the bar will be higher); they decided not to fund and asked to reapply. In the current Feb 2019 round there's a $150,000 budget cap, and since there'd be a risk of Kocherga competing against EA Russia team (which is separate from Kocherga), I decided not to reapply this time.
2. I also applied to the EA Meta Fund, since it seemed like a closest match for what we're doing. They responded that they're not interested for now and that I should apply to the Community Building grant instead.

We could work more on improving our reputation on LW and EA forum (I have a few long posts in mind, e.g. our community building strategy which we've developed recently and which we're very hopeful about), but that's a costly strategy and there's a lot of uncertainty on whether that would be useful (for us or for the international community).