Ask: Will coding ever be made redundant? | Hacker News Reader

Ask: Will coding ever be made redundant?

13 pointsgitgud12 comments

Some friends of mine (mechanical engineers) once asked me (software engineer) if I was worried about my career in the software industry, as they were certain that apps are just made in drag and drop editors and people didn't need to code any more... Thus programming was becoming redundant.

This seemed ridiculous at the time (still does) but I sometimes wonder if code will ever be made obsolete? what could replace it?

After looking at countless graphical implementations of programs and dealing with the frustration of GUI builders, I just can't see CODE ever going away. To me code is the essence of automation, the logic and brains of a system. It would appear you could replace everything else but that...

View on HN

The idea of drag-and-drop editors replacing code is laughable as far as I'm concerned. We might eventually see compilers evolve to the point where a specification written in natural language can be translated into a program automatically, but as the old joke goes at that point writing a sufficiently detailed specification is the programming.

This reduces to the question of: will technical innovation slow down?

Hand-crafted code, composed by thoughtful experts, is a really useful tool when exploring new technologies or when adapting technologies to new industries.

But as when a technology-industry fit matures, and innovation slows down, you can start working in larger grains that require less crafting and less expertise. The problem domains are understood, the tools are stable, and there’s an affordance for inefficiencies. People with industry expertise and vaguely literate in programming concepts can build the tools they need without the intervention of a professional engineer.

This maturation process has already happened a thousand times in various industries, and sometimes we see it disrupted by new innovation.

If there continues to be widespread innovation and disruption, we’ll still need lots and lots of professional programmers writing code.

If, instead, industry problems can genuinely be optimized or technical innovation can encounter practical limits, professional coding will narrow to smaller and smaller niches.

We’re probably still a long way from the latter scenario. I’d say your career as a software engineer is safe. Your daughter’s? Harder to say.

Yes and no.

No, we'll never see a complete replacement for coding. Unless we get literally god like AI that can code by magic, it's simply not possible to replace all programmers/developers/coders with a tool that lets people do the same thing with a drag and drop or WYSIWYG interface. It's simply too complicated for that, and the more ambitious edge cases simply can't be built like that.

At the same time however, a lot of people's needs are also quite modest, and a fairly simple system that could do a few things well could replace many of those needs. Look at the CMS market for example. For many sites and web apps, something like WordPress or Drupal can indeed be used to build them with minimal coding experience, just by adding in a few plugins and a custom theme. Same goes with online services like Shopify or Squarespace or what not. For many simple use cases, that's enough.

Another good example would probably be the Super Mario World tool 'Blockreator'. It's basically a GUI for building simple interactive elements for the game, which lets you drag and drop commands like 'if condition X do Y' as well as 'set whatever variable to this value' (like the timer, power up, coin count, etc). It's not gonna do everything a power user wants, but for many people who just want a few extras in game, it's a time saver.

So the answer is yes and no. Yes at the low end you'll likely see more and more services/tools/whatever crop up that make hand coding less necessary than it was in the past, and many individuals and small businesses won't need anything more complex than that.

But coding will always exist in some form, since the high end requires bespoke stuff that can't easily be done through some fancy tool or service.

Code is just instructions, so ... yeah humans can totally be replaced with something more efficient & better... but what does 'better' really mean? is the output lower latency? higher concurrency? did it do something amazing with little resources?

I've only been in the industry 20yr, but I've met a wide range of people with different goals that drive many factors of many outcomes; as long as people are in the picture stuff is going to be inefficient from 1 pov, while at the same time human diversity can be a huge pool of talent/skill/creativity that a code generator can't have by design (it's solving for efficiency).

It would be a mistake to think things are never going to change. I think it's more reasonable to plan for change and try to keep a head of it by being relevant in a field that interests you, and has demand.

Some aspects of coding as we think of it today probably will be made redundant and in fact have continuously been for quite some time now.

Higher-level programming languages, abstractions, libraries, IDEs: These are all tools that automated one or several aspects of coding

However, short of an AGI some tasks will still require a human being, particularly the translation of business requirements into code. Humans get this wrong frequently enough due to communication failures. I don't see how a machine without at least human-level intelligence could perform better at this.

Consider website builders. Have they made web developers redundant? Generally, website builders suck. They do not allow for the customisation and flexibility required. Static site generators, such as Jekyll, are pretty good for their limited scope, but tools such as Webs generally remain bad for any purpose.

I don't believe some kind of drag + drop editor will replace programming. Even if some kind of editor could, additionally, design underlying functionality and design well, programming also involves a certain style of thinking. Mathematical thinking, for example, is not only applicable in research mathematics. I don't think "business people" could design apps better than us, even if such a tool exists. Indeed, the converse has been true, however (software engineers have indeed managed to replace, to some extent, business people). Startups have thrived on the idea of not monetising early, I think that's something missing in traditional business.

I don't see software engineers being replaced by drag+drop editors, or the typical users of such editors, ever. Conversely, I believe software engineering is one of the most secure job fields there is (some others being mathematicians, scientists, journalists, etc.) - though I may have a bias in this respect.

What may be true, however, is software engineers being replaced, to some extent for some tasks, by their own creations. Although it may be an interesting discussion, it usually follows a lot of hypothesis and any evidence-based discussion on that matter tends to be moot.

Coding is a means of getting something done. They are just set of instructions that computers can process. So technically, there is an argument to be made that tomorrow we could have a new way of creating instructions for computers and not through human coding? Who knows. Just like when cars replaced horses, it did not change the fact that people still need to travel. Just the "how they travel" part changed.

Indeed, we could see industries change. Technology itself has changed many others intensively, who's to say it itself cannot be changed.

Though, unless we can create some self-refining process to create software automatically, some group of people must be creating it. Unless another group, other than software engineers, would be better and more able to apply their current skills to whatever coding becomes in the future, it is more likely that it will simply just be software engineers refining the way they work, rather than being replaced by a different group of professionals.

A friend told me a story from the early 80s. He had double majored in CS and Accounting. He was interviewing and the guy told him programming was doomed. EVERY thing he needed he could do with this new fancy spreadsheet. LOL

When I went to school I had to learn assembly language. Even then C was a good replacement for it, but I still see jobs for people who can program assembly. WTF...

Act of software engineering overall probably no, aspects of it yes.

Just like how industrial jobs are increasingly becoming "human manages machines doing labor" rather than "human does labor", I imagine the same will happen to coding. A computer won't be able to write all your code for you, but it will become more efficient to work in conjunction with one.

That's already happening with lots of existing tasks related to software engineering. (Searching through code, linting, etc.) ML will mean there will be more of them.

Yes, some parts of coding will be made redundant. It's already happening partially with APIs, everything doesn't have to be built in house anymore.

Dystopian, self-programming molochs making everything but themselves redundant? Enter deep reinforcement learning 2030.

Reply on news.ycombinator.com