Is programming still a valuable skill to learn?

[zbigi]

They are also really resentful of their new young rich neighbors who are snapping up the homes next door. Odd because that drives their real estate boom…

Why would they care about the real estate boom if they don't ever want to sell and move?

[Sclass]

Good point. Likely it’s bragging rights.

[ducknald_don]

Having all the capital does provide security even if it is illiquid.

[conwy]

I am a software engineer. Do not love it. Actually I pretty much hate it, and the field is so sedentary. And very capitalist hell which I also hate. But I feel a little better about it twice a month.

Haven't written code by hand in a while, just telling ChatGPT (i.e. correcting it and clarifying) how to do my job when it gets something wrong.

If your job gives you the kinds of problems that can be solved by ChatGPT, I can kinda see what you might hate it.

This is night and day from what I'm getting given in my job. The problems I get are tough and seem to push my cognition to the limit, but I enjoy the challenge, especially when I have an opportunity to solve a problem generally and systematically. I'm often implementing new workflows or processes into very large and complex code bases being actively used in a live environment and maintained by many teams (50+). It's a very engineer-heavy and fast growing organisation, so the deadlines can be tough, but on the flip side, I feel my capabilities expanding rapidly. Even comparing to 1 year ago I find I can produce a lot more in a lot less time.

For me the real passion is problem-solving, so writing code is just one part of it. I spend a lot of time reading discussions, reading and writing documentation, thinking and diagramming out problems and solutions, writing and reviewing code. Apart from the deadlines this is basically a dream job.

It's unclear to me how ChatGPT could do even a fraction of my job. Everything I do requires heavy contextual understanding of the problem domain, business processes, the code / systems and the users. Even if we could upload the whole codebase to ChatGPT (which would obviously be dangerous to the business) I doubt the answers it gives back would be meaningful. These aren't the kinds of problems that are "Googlable" in the sense that there's a clear, general, widely known solution. They're heavily domain and context specific problems. If the knowledge and solutions were public domain, the business could cease to exist, as it could easily be ripped off and cloned by a firm with a bigger marketing budget.

[conwy]

I'm still interested in this one.

Thinking algorithmically is as different from thinking analytically is as different from thinking synthetically as "processing" logically is from "processing" emotionally.

For example, much of standard education is dedicated to logical analysis (math101) and emotional synthesis (lit101). This leaves gaping holes in the standard education.

Having limited formal education, I don't know much about these topics jacob, but I admire you for having studied Physics.

I'm currently working through a Calculus online course as a pre-requisite to study Comp Sci or Engineering.

One think I've found about learning math - it seems you have to understand the supporting concepts for the area you're learning. E.g. to understand limits, I first have to understand functions. To understand derivatives, I have to understand limits. To understand Liebnitz notation I have to understand derivatives and limits. And if I missed some detail about fractions or exponents or order of operation back in high school, I have to relearn it. Learning college math has been tough but I feel I'm growing from it.

This is quite different from software development, where it seems you don't need to know the supporting concepts. E.g. you can use "printf" in C without understanding how strings work or what standard output is. Or you can implement an AWS Lambda function without knowing the internals of how AWS infrastructure actually executes your code.

[Viktor K]

If your job gives you the kinds of problems that can be solved by ChatGPT, I can kinda see what you might hate it.

That sounds even worse to me haha but I understand people enjoy different things.

Latest quotables from corporate:

• Yes I've got COVID but I'm going to try and still make some meetings today.
• Our team is going to have mandatory weekend standups until deadline.
• Our average is 10 tasks/week, but because of the deadline, were going to shoot for 25 this week.
• You can make up any late nights and weekends with some extra days off after the deadline.

I think coding in ERE terms is primarily valuable for building a nest egg. Coding small apps when free can help in a digital and connected world, though, and I've made some hobby apps before. However, only the simplest have both made it to production level and/or seen any real use since I wrote them. And these days, those can be hacked together with ChatGPT and a beginner level understanding of code.

But tough to prove how there's anything more valuable than income in the coding arena (or hobby/personal enjoyment as conwy mentioned).

[7Wannabe5]

In the low energy world of the future it might still be possible to, for instance, construct a more efficient irrigation system with water as flow and mechanical logic gates. As Ryan North outlines in his very fun book "How to Invent Everything", the only technologies that must be (re)invented prior to construction of an analog computer are Symbolic Logic and Forge/Kiln. So, even at that level it could still be a valuable skill.

@Viktor K;

It's interesting how the management tactics you describe align with what I am currently learning in Project Management 699. The theory is that most humans, inclusive of skilled technicians, estimate the time needed to complete a task at 90% likelihood of success level, so one method towards "crashing" a project towards completion would be to force the completion times closer to 50% likelihood of success level. Unfortunately, it may sometimes be the case that a project manager might conclude that if 9 days allotted = 90% likelihood of success then 5 days allotted = 50% likelihood of success. "The Goal" by Eliyahu Goldratt, a very popular (although already dated/old-fashioned) book in the odd genre known as "business novels" introduced the "theory of constraints" and some thoughts on human production-psychology that underpin tactics such as these.

Although I am sure it is terrible to be a worker in such a situation, it can be fun to play around with project management software and theories applied to your own life projects. For instance, could you "crash" your early retirement estimate?

[Viktor K]

@Viktor K;

It's interesting how the management tactics you describe align with what I am currently learning in Project Management 699. The theory is that most humans, inclusive of skilled technicians, estimate the time needed to complete a task at 90% likelihood of success level...

Valid, for sure. There's surely some soft skills or analytical skills or even hard skills in coding job that can apply, but maybe less so than some other careers e.g. electrician or carpenter. Trade off being the income

[AnalyticalEngine]

But tough to prove how there's anything more valuable than income in the coding arena (or hobby/personal enjoyment as conwy mentioned).

My personal experience with this after doing it as a hobby for ten years and then in corporate for 10 more years is that the hobby/start up apps are a lot harder to "score" with now because the market is 100x more saturated now than when the internet started. There's way more technology to learn, so coding jobs often become plug and play between 100 different libraries and old dlls and frameworks and build pipelines and APIs and databases more than actually writing lines of code. And unfortunately, bizarre Windows compatibility issues are something ChatGPT is bad at solving, even if it can write your entire REST API for you.

These are all just the continuation of trends that have been on-going for 10+ years. While actually interesting coding jobs might exist, they require networking or more skill than your generic corporate programming gig, and you can usually cash out a generic corporate programming gig with minimal effort.

That being said, I did teach CS in a university for awhile, and students who are there only for the money usually eventually drop out because programming is hard and boring enough that most people don't last unless they have a genuine interest in it.