Hacker News Reader

Comment by atemerev | Hacker News Reader

This happened before, and this will happen again. There are no more typists, nor data entry personnel. Hardly any human travel agents. No bank tellers. Human translators are next in line.

However, much more positions in machine learning and data science.

Software _is_ eating the world. The only viable survival strategy is learning to code. I don't believe that not everyone can learn to code. I teach people to code, and I have yet to meet one that couldn't learn, assuming some general intelligence (about the same that you need to learn a foreign language).

View on HN

The coding skills I'm trained in will be handled by Copilot or its successor in a few years.

I haven't gotten around to faffing about with Python and Tensorflow or whatever yet.

Don't think of "coding" as writing javascript to make websites. Think of "coding" as "talking to computers".

As software eats the world, the value of people who can talk to computers will increase. Even if it takes fewer programmers to make a website, there will be more jobs for programmers to automate concept art production pipelines. You may not be writing javascript or python in fifteen years (I bet you will) but there will be code to tell the automation services what to do.

The interesting question is the general population becoming more tech savvy? Will this change in work encourage more students to learn how to code (whatever that looks like in twenty years)? Or will the demand for coders rise without a corresponding increase in supply?

Well, in software engineering, like in many modern professions, you don’t get trained once and then work for life. Learning is continuous, and this is not optional anymore.

... will there ever be a (better) Copilot for COBOL or PL1? This may be an escape hatch, just throw yourself at legacy mainframes, this may work for an additional 20-40 years. :D

Yes, Copilot works for COBOL. (Haven’t tested for PL/I, though).

Reply on news.ycombinator.com