Monthly Archives: April 2020

Testing for Covid

The author describes the process involved in actively trying to get tested for the disease, given symptoms and history.

The upshot is: we have some idea of the mortality figures for this disease in the US (although there are indications that there are many uncounted deaths of people who never make it into a hospital). We have a very much more approximate idea of infection rates, because we aren’t testing people in a comprehensive way.

Until these two numbers are nailed down within an acceptable range of error, and we can surgically test everyone following contact tracing paths on demand, relaxing the worst-case self-isolation instructions seems like a huge gamble with people’s lives. As a heuristic, I’ve been assigning a confidence level based on the ratio of the two being in the range of 2% or so, based on mortality rates from other parts of the Western world and the more reliable mortality numbers. I see no evidence that either of these numbers can be trusted for the foreseeable future, so I don’t expect to be out of self-isolation anytime soon.

Absolute scale corrupts absolutely · Tailscale Blog

Source: Absolute scale corrupts absolutely · Tailscale Blog

Tailscale is interesting.  They’re pushing auth into the networking level, using WireGuard, so all connections of any type across the net are secure and restricted (if appropriate), and identity is associated with IP.  As one of the founders says, it’s like going back to working on LAN: you can do arbitrarily insecure things, because it’s locked inside a building that only a few people can muck with.  But it’s a LAN you can get to from anywhere and anything.

They also like to build things that are simple to use, by design.