Is the era of VC-fueled complexity ending? | Hacker News Reader

Is the era of VC-fueled complexity ending?

55 pointshliyan20 comments

I initially wrote "over-engineering" in the title, but changed it to "complexity" so as to not unnecessarily trigger some readers. I've been in the software industry for 20 years, and on HN for 10. This quarter was the first time I've seen a pattern of a lot of engineers yearning for old-fashioned simplicity while critiquing "modern" tech and practices, some of which were still on vogue just a year ago. E.g.

Imaginary Problems Are the Root of Bad Software https://news.ycombinator.com/item?id=36380711

Looming demise of the 10x developer – an era of enthusiast programmers is ending https://news.ycombinator.com/item?id=36719409

Every time you click this link, it will send you to a random Web 1.0 website https://news.ycombinator.com/item?id=36739920

Windows 9x and Word 9x at 800x600 resolution. Spacious. Comfy https://news.ycombinator.com/item?id=36682403

Why I Hate Frameworks (2005) https://news.ycombinator.com/item?id=36637655

DevOps Is Bullshit https://news.ycombinator.com/item?id=36354049

How I run my servers (2022) https://news.ycombinator.com/item?id=36744090

Look ma, no React: I recoded my portfolio site with vanilla everything https://news.ycombinator.com/item?id=36736576

Don't Take VC Funding – It Will Destroy Your Company https://news.ycombinator.com/item?id=36654960

Some of these submissions would have been received very differently in the past.

I was just entering the industry when the original OOP craze began, and for the longest time, thought I was stupid or crazy for not "getting" it and preferring to write simple functions. Same with cloud-native systems, reactive SPAs, container orchestration, SCRUM, and a bunch of other practices. In a way, this is refreshing. But at the same time, I worry that this may signal the end of the current era of explosive growth in software (both in terms of tech and money), and what we thought would be perpetual progress may in fact be a brief summer. Does anyone else feel this way?

> I worry that this may signal the end of the current era of explosive growth in software (both in terms of tech and money)

It's interesting that you're worried about this. I'm hoping for this. I think that the explosive growth we've seen has not been good for software, the industry, or our customers. I yearn for a return to a saner pace, where quality is valued more highly than it has been.

Don't forget the OG complexity hater, grug-brained dev: https://grugbrain.dev/

Thought you were going to mention Zed Shaw lol

This is my first time reading this and I loved every second. Great read - got a few snickers out of me. Thanks for sharing.

> "Oh, don't worry: the tests will show you what you need to do."

>grug once again catch grug slowly reaching for club, but grug stay calm

The early 2000's recession, which you would have just missed, brought about a major trend shift that realigned more of tech to Web 2.0 ideas. When good programmers are laid off, they find themselves pushing fundamental parts of the software stack forward that "nobody had time for" before.

So, we are absolutely in for another such shift - the easy money spigots are turned off - but probably in a different direction. It's definitely not the case that the software world has run out of interesting problems.

I see this now with people rediscovering that using Django/Rails/PHP for web development is possible in 2023. Not every application needs to be built like a SPA using the latest hyped JS framework with a GraphQL-based API, specially when the application is composed of mostly CRUD pages.

Containers and Kubernetes should probably make way for the next (simpler) thing now. It will start slowly but all things die and I’m excited to see what kills them.

No. Unnecessary complexity is a powerful snake oil that will always be sold. Everyone benefits from it too, except for customers and investors. And screw those people.

The complexity exists because it creates demand for products that eventually become entrenched over time (Terraform, Kubernetes, DevOps products, SCRUM/Agile etc.). I don't think this is necessarily bad, but I don't think the complexity will ever stop existing until the incentive structure to keep it in place goes away. It's hard to reverse complexity, so I can't really see this happening.

I do think that the definition of 'growth' will change from spending money (i.e. 'growing a team' or 'growing headcount') and sheer output (we hit X goals this year) to actual long-term profitability (not just EPS like it's been for the last year or so). I don't personally benefit from this (since I am a tech worker) but from a social standpoint it's probably a good thing.

It's the beginning of the age of automated development/service and software taming. VC is silently replacing positions with AI, and looking for ways to control all the complexity with as few knobs as possible.

Shopify employee breaks NDA to reveal firm replacing laid off workers with AI https://news.ycombinator.com/item?id=36828409

That’s customer service not engineering.

Re: the [dead] sibling comment that reads: “80% of code will be written by Copilot”

That's like Microsoft saying "80% of code will be written by intelli-sense / autocomplete." Great for minimizing typos and increasing productivity, but meaningless with regard to overall engineers required.

It's not VC, but rather the zero percent "free money" that the VC and Wall Street types have that have deranged capitalism so effectively.

A return to fiscal prudence and sound management/programming practices is well overdue.

If AI is the future of programming, it will be with python, jquery and php, because they make infinitely customizable spaghetti without the constraints of abstraction. And if AI is going to be handling all the spaghetti that's fine.

Wow, jquery and ai? Are you sure?

is AI racist to jquery? vanilla JS?

If Peter Zeihan is right, the era of 0% interest rates is over, and we're looking at 10% rates for a decade, in the US. (If, and only if we manage reshoring correctly, otherwise rates never go back to normal) The rest of the world is likely to be worse.

We can dream.

I wonder if a lot of people who were leading freelance engineers are pulling back to use AI to learn new things, or to go out on their own?

Just a thought.

Reply on news.ycombinator.com