In fact, almost nothing that's actually _useful_ is. Anything that can be broken up into predictable low effort chunks is something that can be automated away.
>
If you can automate away programming, there's some people who would like to give you a lot of money.
A lot of the work that programmers do (which is not identical, but somewhat related to "a lot of programming") could be left out (which is not identical, but somewhat related to "automated"). The reason is that a lot of programming is working around the gigantic brokenness of the whole software ecosystem. Here there does exist an insane amount of possible cost savings.
Concerning the "some people who would like to give you a lot of money": I actually talked about my thoughts with a friend who knows a lot both about economy (he works as a business consultant) and programming. He clearly said that my ideas on this topic were really smart and thought through, but getting money is a lot more complicated (it demands both being a good salesman and having connections). So "some people who would like to give you a lot of money" is simply not true. :-(
> gigantic brokenness of the whole software ecosystem
If you mean the employing organization and its ignorance, I agree. For most organizations today, their value is locked up in technology and those who maintain it. The rest of the business (non-technical people) are essentially overpaid punching bags for everyone to smack around. In an ideal world programming would not be a special skill, but just another means of communication.
I don't think that's what the parent to your comment was about.
You can't automate away programming, but you can write one program that behaves equivalent to many that accepts configuration from people who don't know how to program.
I don't think he says predictable or low effort. Obviously you can't build and release a bridge in a week, nor a big software system. But you can still push yourself to move fast (a relative term), and these projects do in fact get broken down into steps, even if you can't predict all the steps from here to there, and even if you can't release something along the way.
In fact, almost nothing that's actually _useful_ is. Anything that can be broken up into predictable low effort chunks is something that can be automated away.