[CorsiHockeyLeague]

Originally Posted by blankall

I run a small business, and none of those services can be automated. I need to answer the phone and speak to customers.

How long do you think it'll be before something like this can deal with 95% of those calls?

Originally Posted by blankall

I need to have a person track my finances. A computer can physically print the letter, but I still need someone to physically input the data into the computer.

What about this task makes you think that a computer can't do it just as well - if not better - than a human? What step in the process is so subtle that it can't be performed automatically? Obviously it isn't data entry.

I think you are vastly over-estimated the current abilities of automated technology. AI cannot respond to nuanced and dynamic situations. Real life scenarios require an actual human mind to navigate around.

I think you're vastly underestimating how far AI is advancing all the time, and how close we are to AI that exceeds human minds in all aspects - it's likely a matter of decades, and what we're talking about is far easier to accomplish. This is a talk from last week that's very optimistic but really puts what we're facing front and center.

We also are very far away from having robots who can climb into someone's apartment and remove a clog from a sink. We can barely build a robot that can walk in a straight line. We haven't even built a machine that replaces a simple plunger.

You've taken the one thing that I admitted you probably still need - a plumber - and created multiple examples based on that. The reality is, that plumber's entire business structure - accounting, scheduling, transportation, even diagnosis of problems - will be automated to a significant extent such that we need fewer people working in plumbing businesses, even if we still need plumbers. This is true in basically every small business context you could imagine.

Basically, my point is that there are few real life situations, run by small businesses, that you could totally automate.

And I completely disagree with that point, because it ignores a whole bunch of automation capability that currently exists, even if it's not broadly available to be used in those businesses yet. In ten years, it will be, and it'll be even more effective.