It took a huge amount of effort from several people on the team, but we seem to be ahead of the ground truth curve with regards to the March 15th eligibility expansion. That was a bit of a nailbiter.
My favorite chart:
"What's the sudden drop?"
That's ~2 people reviewing two call centers of calls for one day for accuracy / consistency, making manual corrections, and then pushing the lot live a bit after midnight.
We know from experience that there is a productivity and accuracy curve with callers, so as our initial agents in the call centers season and as we add more agents, we'll have more calls than we know what to do with... nah, I'm pretty sure we know what to do with them.
I wonder how many billions of dollars every year are wasted due to friction in paying taxes.
Just at the payment step. Where you know how much you owe, why, and which authority you owe it to, but now you have to get them the money.
I just had a clock-definitely-running conversation with a Dangerous Professional where we discussed, for ten minutes, the *strategy* necessary to pay a particular government what I owe.
Good golly. This would be inexplicable if I didn’t work in payments. Since I do it’s a farce.
“Maybe you should just take the hit for late payment and physically walk cash into the office after the pandemic is over?”
“ARE WE ACTUALLY DISCUSSING THAT?!?”
One reason that the national conversation on vaccine availability isn't necessarily reflective of ground truth is that we've set up an auction where you pay not with money but with PMC-level brainsweat.
People who participate in national conversation have a lot of that to spend.
(This is subtly different from the PMC's diagnosis of the issue. The language of power can say what it will about root causes, but one side effect of being the language of power is that it is very, very, very good at getting what it wants out of bureaucratic processes.)
Imagine if there were figurative passwords being passed out right now to get a vaccine, and a great debate on whether publishing the passwords was justice-maximizing, conducted largely by people who professionally specialize in guessing passwords of complex systems.
The thing most technologists would do well to internalize is the notion of fraudogenic environments, where one can (unintentionally!) set up an incentive system which causes fraud to happen w/o there being a top-down directive to commit fraud.
This is against the intuitions of a lot of people, who think there is some seedy back room where four senior banksters in suits have developed a multi-billion dollar fraud and want them marched out in handcuffs.
It's Day 60 of VaccinateCA.com and it's a big day, but most important news first:
Today we called hundreds of pharmacies looking for the covid-19 vaccine and eligibility criteria, and published what we learned.
Tomorrow, March 15th, is the largest expansion of eligibility for California yet, to include workers in several industries plus individuals with pre-existing conditions which would place them at elevated risk.
This substantially increases the count of people eligible.
Describing eligibility in a tweet is hard so a sneak peek: vaccinateca.com/checklist ; we're publishing a tool tomorrow to help vaccine seekers (and people managing care for others) easily navigate the decision tree.
Have you ever found yourself doing something which you know is strikingly irrational and which you feel powerless to resist?
This is me when one of the banks says that I have 18 cents worth of points expiring on March 31st. "This here is a P0 drop-everything-and-resolve email."
And then I spent a few minutes reading through the various things I could convert those 20 expiring points and 100 non-expiring points into to maximize the conversion ratio, because darn it if I was going to leak 30 cents worth of margin by picking the wrong point ecosystem.
"This sounds even crazier than airline miles. What ecosystem works like this?"
There are numerous competing points systems in Japan and one thing that the less-well-distributed systems do is offer point conversion to more-well-distributed systems, often below par value.