Are we going into codeless future? | Hacker News Reader

Are we going into codeless future?

7 pointstsibelman5 comments

Yesterday I visited a Meetup hosted by Wix https://youtu.be/_Xf3M2s4KG0?t=3069

To summarise, existing development/packaging/hosting technics are not able to keep up with the velocity of business demands that Wix is facing, so they are trying to develop a platform that would be at the highest level of abstraction, developers writing only business logic code, with deployments times at the level of seconds and maximizing of the utilization of resource. The solution is cohosting functions written in JS by different people/groups inside the same process, providing higher-level Wix-specific APIs for interacting and deploying them without any dependencies or packaging.

While the individual choices they made can be debated, the more interesting question is the software world is going in the direction they took?

I think it connects to the NoCode/LowCode buzz that going on now for a couple of years. If yes, developers of tomorrow would be separated into two groups, one enormous consisting of today business operation people who are going to write in DSL for highly abstract platforms like Wix and another much smaller in numbers even smaller than today software development community who is going to develop and maintain these abstract platforms. What does it mean for our careers if we want to keep enjoying high salaries? Should we seek to dig deeper into the internals to be in that small group of platform maintainers, or the real money would be in the creative usage of business DSL?

I vaguely remember VB and PowerBuilder days from the start of my career. At the time, people who specialized in business-centric development in these environments became quickly irrelevant as technology moved to the internet. So I myself have a somewhat skeptical and cynical view of such efforts. I prefer to have highly transferable skills, and being a single company business logic developer has limited value to the outside world.

No. This has been going on since before COBOL. No code/low code is always just around the corner. Now we have the additional fantasy of AI writing code for us.

The more routine and mechanical parts of developing software will get automated away and packaged up as managed service; that's been happening for a long time too.

If businesses could meet their software needs with some off-the-shelf packages plugged together that would have happened a long time ago. In some domains and applications it has happened -- no one writes custom database managers anymore, but that was part of the job when I started (pre-RDBMS). Business software generates new requirements and needs, often faster than the business can find people to work on them.

Some people may get put of a job by new software tools, that happens when people over-specialize and fail to adapt. I've been programming for 40+ years and have heard about how my job is going away real soon now since I started. I'm not worried.

What does it mean for our careers if we want to keep enjoying high salaries?

It means there'll be a lot of downward pressure on salaries for people who implement apps and websites for clients. If you want to keep earning a lot you'll need to be building the tools that other developers use.

I don't really buy it though. No-code is great for things that have clear requirements and no innovation beyond what's been done already, but there's very little scope for building something entirely new without writing code. I strongly suspect no-code is only going to grow the tech industry, and there'll always be demand for other devs to work on the new innovations and more complex things.

Let me chime in as someone developing a no-code website builder. Wix are in a similar space as us in fact, though that doesn't matter for the point I am about to make:

No-code only works in a very limited context. Whether that will ever change, I do not know. But to make anything truly no-code, you immediately face a trade-off that's nearly impossible to overcome. You either:

a) Make your no-code solution more flexible by adding a gazillion options to everything, steadily driving up complexity, until learning your no-code solution is about as difficult as just learning the underlying thing it tries to magic away

b) Limit the remit of your application to the point where it does a few things relatively well, but admit that any customisation is unlikely to be possible and hope that your remit isn't too limited to impair usability

Particularly in the website space I feel that 'the thing that just works' hasn't been found yet, despite hundreds of millions in investment and advertising being thrown at that exact problem space. Which tells me it's a tough nut to crack.

And ultimately you will always need people maintaining the 'no code' solution. And those are more complex to maintain, meaning it takes more people to do it.

You might one day be able to replace a few of the simpler coding jobs, by building tools where people just sort of visually plug one API into another, but so long as I still see people doing jobs that could be automated by some simple bash scripting EVERYWHERE, I can tell you, programmers aren't going anywhere anytime soon.

This is a great explanation, thanks. Any no/low code tool that has the features and power to address real-world problems will get just as complicated as “real” coding.

WordPress is an example. Easy to put up a few blog pages, but anything else requires plugins and glue code and someone who knows how to make it work.

SQL was at one time presented as a tool for non-technical people to run ad hoc queries, supposedly saving time not waiting for programmers to write code. It turns out that mastering SQL is moderately hard, but understanding the relational model and a non-trivial schema requires real skill and experience.

Just look at the huge number of terrible spreadsheets in the wild — Excel is the most widely-used low/no code tool out there and almost no one can do anything but type into cells and maybe total a column. I taught my daughter advanced Excel and pivot tables for her job at a big non-profit, now she’s practically lay-off proof.

No. There will always be a demand for people to write actual code.

Reply on news.ycombinator.com