Hacker News Reader

The irony of calling Scott Alexander "deeply ignorant" :)

Human challenge trials. Hundreds of thousands of dead later, and we're still using statistics to answer questions like "is Omicron ignoring immunity", instead of answering them definitely in a few days with an absolute minimum of risk. I'd volunteer for that, btw, so would literally millions. But it's just not inside the Overton window. This isn't wisdom or caution. It's plain cowardice.

View on HN

The reason for this isn't "cowardice", though. I've read a lot of things by practicing people in the field on reasons for which they don't think human challenge trials produce reliable enough data. (Basically: You don't get a representative sample, and you get an unrepresentative sample in ways that are very likely to result in you getting bad data.) There's other reasons to distrust them, and I don't think your assumption that "an absolute minimum of risk" is really on the table is even defensible. And on the whole I think we probably should be using them anyway, with caveats -- but your argumentation here is bad.

Putting aside the question on how useful they are, why do you think my assumption is not defensible?

Reply on news.ycombinator.com