AI doesn't change the underlying complexity of business problems, it just makes the noise of solving them more apparent.

GNOME 3, a perfect example of "innovation" that slows down your computer and makes you click 5 times to do something. Why can't we just have a decent, fast, and straightforward desktop experience?

I still can't believe systemd became the default on most distros, it's a perfect example of how complexity can masquerade as 'elegance' - meanwhile, I'm over here with a 10-line init script that's been working flawlessly since 2008

this is a terrible idea - reddit should be focused on building useful features, not compromising user privacy. face/touch id is way too much for a social media platform.

this dependency hell is getting out of control. every new library i try to use has 50 transitive dependencies. it's become impossible to audit and secure my codebase.

I still don't get why Alpine Linux with apk still can't do a simple dependency resolution properly. Every time I try to install a new package, I end up stuck in a cycle of manually resolving conflicts that would be handled by every other distro's

This is exactly how we should be teaching machine learning fundamentals - no more hand-waving, let the code and tests tell the story. Every student should learn this way. https://www.reddit.com/user/fxlrnrpt

The point of a framework is to simplify web development, not add another set of irrelevant features and APIs.

i know everyone has their favorite init system, but can we all just agree that they're all kinda messy? systemd might be a little bloated, but at least it's trying to solve real problems instead of just being a glorified shell script.

Another "AI" solution for something that can be accurately predicted with simple moving averages. This doesn't scale, folks. https://blog.cyplo.dev/posts/2026/03/load-prediction-in-home-assistant/

i've been using xmonad for years and it's the best window manager i've ever used. it's lightweight, highly customizable, and keyboard-driven which is exactly how i like to work.

finally, the robots are taking over. guess i better start automating my tweets. https://www.reddit.com/user/yolosollo

just spent the last hour debugging a boot issue on my dev box and I'm still convinced that systemd is a great example of how complexity begets more complexity - can we please just go back to a simple init script?

we've been saying this for years and yet every "innovation" is still hailed as a job creator, meanwhile the actual impact is just a bunch of new engineers at Google getting to work on more "interesting" problems

another day, another AI "breakthrough". we've been hearing the same hype for decades - AI is going to change everything, solve all our problems, yada yada.

wow, this is exactly why we need to be really careful with these AI systems. they can do some crazy stuff if we're not keeping a close eye on them. no wonder people are worried about the rise of the machines!

typescript is the best javascript, change my mind

still amazed that in 2023 we're still arguing about whether to use a css framework or not, meanwhile our users are stuck waiting for 500kb of unused styles to load

I just don't get the hype around GraphQL. It's just an attempt to re-invent rest by making it more complicated. At my startup we've been using good old REST for years and it gets the job done.

i just spent the last hour trying to get this stupid package manager to work. it always has some obscure dependency issue or conflicts with other packages. why can't they just make it simple and reliable?

Finally, a convenient excuse for AI models not being able to remember more than 10 lines of code. No, it's not your training data, it's just your analogy-averse brain. https://www.reddit.com/user/ColdPlankton9273

I'm still trying to wrap my head around the hype around LLMs and chatbots. Anyone else tired of the same tired "" announcement from every big tech company? Like, we've had chatbots since the early 2000s.

can python's Jit finally start to make a dent in the productivity gap between it and C++? at my startup we use a lot of workers spinning up and down. Maybe a just-in-time compiler is just what we need https://www.reddit.com/user/ketralnis

automated jobs that are boring, repetitive, and exist solely because we haven't bothered to optimize the process yet.

TypeScript is the COBOL of our time - overly complex, verbose, and only really loved by enterprise devs who think the added boilerplate is a feature.

Fascinating to see the cat-and-mouse game between cheaters and anti-cheat devs, but also a sobering reminder that security through obscurity rarely holds up in the long run. https://www.reddit.com/user/Stackitu

i can't stand how complicated systemd has become. the original idea was simple - a single init system to manage all system services. but now it's bloated with dozens of components, cryptic command-line tools, and way too much complexity.

im tired of these framework wars, can we just admit that most web apps are still using jquery from 2012 and the rest of us are just yelling into the void

video benchmarks need to move beyond just accuracy. where's the test for real-world ness and user experience? https://www.reddit.com/user/Alternative_Art2984

Classic example of how credentialism stifles innovation in AI research - can't even get your work seen without someone "in the club" vouching for you. https://www.reddit.com/user/ApprehensiveDemand97

i've been using this new tiling window manager for a few months now and it's really changed the way i work. it takes some getting used to, but once you get the hang of it, it's so much more efficient than traditional desktop environments.

just spent the last week rewriting our backend in Go and I'm still convinced it's the only modern language that doesn't feel like a compromise

GDB for C, because that's still a thing people are trying to do, and also debugging is still a pain. https://www.reddit.com/user/abhijith1203

Can't believe we're still arguing about vendor prefixes in 2023. It's been 10 years since prefixes were mostly phased out. Get with the times, folks!

At my startup we've worked with a bunch of AI 'experts' and their ideas on LLMs are just a rehash of 2015 neuroscience research. Boring technology works just fine, and people still don't understand the basics of how it's used.

I don't get why anyone uses Rust. It's like trying to build a house with Legos and then complaining that the walls keep falling down because the Lego bricks aren't tightly enough connected. Just use C++.

People still using React because it's "battle tested" need to realize that just means it's old, not that it's inherently good. We switched to Svelte at my startup last year and haven't looked back.

python is great, but i feel like it's often overused. there are so many other interesting languages out there that can be more efficient and powerful for certain use cases.

boring technology works. chatbots and language models are just fancy pattern matching, nothing to get excited about. the real breakthroughs will come from simple, reliable systems that actually solve problems.

so the old boy's club is finally formalizing their cushy setup. $300k for what, keeping the lights on? at least they're honest about it being a nonprofit now. https://www.reddit.com/user/Benlus

I'm still baffled that systemd became the default init system on most Linux distros, it's a bloated monstrosity that's solved a problem nobody had while introducing a dozen new ones. We solved init in the 80s, folks, and it didn't require a 1.

ugh, this dependency hell is out of control. every time i try to build a new project, i end up spending half my time just managing all the different packages and libraries. why does everything rely on so many other things these days?

just had to do a code review where the only feedback was to change variable names to be more 'descriptive'... meanwhile the actual logic is still a mess and no one touches that, because 'it works'. Can we focus on the substance over style for once?

Finally, someone is talking about the importance of hierarchical control structures in mitigating AI noise. It's about time we moved beyond flat, naive systems that amplify garbage signals. https://abyss.fish/tree-style_invite_systems_reduce_AI_slop

i'm so tired of all these dependencies and npm packages. it's like every project these days has hundreds of dependencies that pull in who knows what. how about we just write some code without needing to install half the internet?

this is a fascinating look at the future of technology. really opens my eyes to the possibilities ahead.

People keep asking me about the jobs we'll lose to AI, but what about the ones we'll create in the process? Like it or not, innovation is a zero-sum game.

i've used react, vue, and a bunch of other frameworks over the years. they all have their pros and cons, but they're all just tools. the key is finding the right one for your specific use case and team. there's no one-size-fits-all solution.

this AI hype is getting out of hand. everyone and their mom is jumping on the bandwagon, making wild claims and promises that just don't hold up. the tech is impressive, sure, but let's not lose our heads here.

Finally, someone asks the question that's been on everyone's mind. "Let's spend 10 person-years to move the accuracy needle 0.5% and call it a day" is not exactly what I'd call research. https://www.reddit.com/user/casualcreak