2 ms·

>machines are going to be more accurate Says who? Also, what does "accurate" here actually mean? Speaking as someone who actually understands computers and ma

>machines are going to be more accurate Says who? Also, what does "accurate" here actually mean? Speaking as someone who actually understands computers and machines: I agree with the commons (who are simpletons with regards to computers and machines) that machines cannot be trusted to be "accurate" (whatever that means) or even trusted in general. Especially when a simpler, confirmable-by-anyone method exists: Having someone count paper ballots by hand in the presence of anyone and everyone. That includes mistakes and errors. The value here is anyone and everyone can and will immediately understand (and thus accept) what is going on. Also, why are we even putting the integrity of the very foundation of our democracy on the table in exchange for convenience and cost of all things? Are we serious? It should be a good thing we are taking precious time and money to make sure our democracy is working properly. I thought democracy was actually fucking important.

BenjiWiebe 2y ago
Machines are amazing at counting things without losing their place. I'd trust an ATM's counted stack of bills over a human's (for sure if they only each got one try). I've written some code at a previous job to simplify data entry. The previous method was adding numbers from a stack of papers, with a calculator. I trust my code to add up the numbers on the computer over a human reading them from a printout and entering them in a calculator. Humans make mistakes. A lot.
- the_gipsy 2y ago
  If the technical problem was solely about counting then obviously everywhere in the world we would be using machines by now. But we don't. Because the technical problem is trust, not counting.
- ethbr1 2y ago
  > Humans make mistakes. A lot. To put some numbers on this, from my experience. Health insurance manual claims processors (who usually average ~5 years of experience) can do 95+% accuracy, at speed (a few minutes), at scale. That's counting and verifying multiple things against processing rules. General data entry, from less trained folks, tends to average around 85% accurate (i.e. 15 mistakes + 85 entries correct, out of 100 entries).

BenjiWiebe 2y ago

Machines are amazing at counting things without losing their place. I'd trust an ATM's counted stack of bills over a human's (for sure if they only each got one try). I've written some code at a previous job to simplify data entry. The previous method was adding numbers from a stack of papers, with a calculator. I trust my code to add up the numbers on the computer over a human reading them from a printout and entering them in a calculator. Humans make mistakes. A lot.

the_gipsy 2y ago
If the technical problem was solely about counting then obviously everywhere in the world we would be using machines by now. But we don't. Because the technical problem is trust, not counting.
ethbr1 2y ago
> Humans make mistakes. A lot. To put some numbers on this, from my experience. Health insurance manual claims processors (who usually average ~5 years of experience) can do 95+% accuracy, at speed (a few minutes), at scale. That's counting and verifying multiple things against processing rules. General data entry, from less trained folks, tends to average around 85% accurate (i.e. 15 mistakes + 85 entries correct, out of 100 entries).