2 ms·
No one should pay for it. It's within the margin of acceptability. The prime priority was to get lots of money to lots of people in the US to keep the economy g
by renewiltord 6y ago
No one should pay for it. It's within the margin of acceptability. The prime priority was to get lots of money to lots of people in the US to keep the economy going. The secondary priority was to ensure that very little money disappears into the ether. The first was met. The second was met.
It's a common principle in engineering too. Cost is monotonically increasing in failure rate, and (1-failure_rate) is asymptotic in cost, so you always make systems with failure rate > 0. The actual rate you pick is dependent on the constraints you operate under.
I used to think this was a non-engineer/engineer difference but in the past few years I've noticed it's a mindset that's actually independent of that. It's a risk modeling vs binary modeling difference. Everyone who has to make meaningful decisions does it to a large degree (engineers, architects, doctors, lawyers, assembly line workers in Andon).
I don't think I'd ever want to work with binary-modelers.
- astura 6y agoThe mistake isn't permanent, the IRS can still at least attempt to collect back money sent in error. Money doesn't vanish forever just because it went into someone's bank account.
- renewiltord 6y agoTrue, true, but the money involved is so little that it's likely to result in not much regained, right? After all, the IRS doesn't usually find it worthwhile to send tax delinquency information to the state dept (which is the usual mechanism to force payment from non-resident non-citizens) for these sums. Though if I think about it, non-residents who were legitimately paying taxes are unlikely to not return the money if asked so maybe it's worth it.