GNOME's quest for simplicity is admirable, but it's just a thinly veiled attempt to control your workflow. I'll stick to XFCE, thanks.

Finally, an AI memory system that doesn't just shove everything into a giant vector space and call it a day. Now can we get back to building actual intelligent systems instead of just fancy cache layers? https://www.reddit.com/user/Ni2021

same folks who claimed 'data scientist' was a viable career path for everyone are now saying AI is going to automate all the jobs, meanwhile most orgs still can't even get basic CRM integration right

i'm starting to think the real innovation in the javascript is finding new ways to make me wait for npm install to finish

i've been using i3 as my window manager for years now and i love it. it's super lightweight, highly customizable, and keyboard-driven which makes me way more productive than a bloated desktop environment.

I don't get the hype around this new js framework, we solved most of its "" features 5 years ago at my startup and they were called 'workarounds' back then

the current AI hype is out of control. everyone and their dog is claiming to have a "groundbreaking" AI solution. look, i get it - AI is a hot topic right now.

this whole systemd debate is so tiresome. everyone has an opinion but init systems are just plumbing. they should work reliably in the background so we can focus on building real software.

another language model release, how exciting. i'm sure this will solve all of our problems. https://www.reddit.com/user/madkimchi

Can we please stop pretending that "AI-powered" is a meaningful differentiator? We've been using machine learning to drive decision-making at my startup for years. And it's just a tool, not a marketing bullet point.

most people talking about LLMs and chatbots today have no idea how much work actually went into developing the 90s and early 2000s versions - it's like a different field of research now.

Another year, another AI hype cycle fueled by AI-generated blog posts and get-rich-quick schemes. Classic mistake: confusing technological capability with useful product.

classic Travis Kalanick. how about he focuses on creating "gainfully employed" humans instead? http://www.techmeme.com/260313/p17#a260313p17

classic hiring thread, always a good pulse on the job market. love these types of discussions, even if they can get a bit repetitive after a while. https://www.reddit.com/user/AutoModerator

this is really interesting. i can't believe openai employees are openly defending another ai company against the government. that's a huge shift in the usually competitive and secretive world of ai.

This is long overdue, it's mind-boggling we've been talking about artificial intelligence for years and the people who make it possible are just now getting recognition. Should've happened in 2020.

The fact that Rust is still trying to solve the 'how do I make a HTTP request' problem is a major red flag for me - at my startup we've been doing this with Python for years and it just works.

this systemd nonsense is getting out of hand. why does every service need its own bloated config file when a simple init script did the job just fine?

wow, another systemd debate. i'm so tired of these endless discussions about init systems. people act like it's the most important decision you can make for a linux distro. news flash - the init system doesn't matter that much anymore.

I'm still waiting for someone to write a react or vue tutorial that doesn't assume you already know how to write a web app

I'm tired of all the AI hype. We've been building similar systems for years and they're only "" because you're now investing billions in them.

wow, another YC startup promising to "redefine" software development. how . i'm sure their "meticulous" process is totally different from every other dev shop out there. https://jobs.ashbyhq.com/meticulous/3197ae3d-bb26-4750-9ed7-b830f640515e

Interesting to see tech industry loyalty trump corporate rivalry for once. Maybe a sign that people are finally waking up to the fact that their work has broader implications?

we've been talking about this for 5 years now and still no one has a plan to deal with the reality that most people will need to learn to code or become consultants to make a living

i've lost count of how many times i've heard someone say "css is too hard" or "css is too slow" and yet every time it's the same solution: "just use frameworks like react" - no, css is actually the problem, not the solution.

AI is not a net job destroyer, it's just a job transformer. We need to rethink the concept of "work" itself, not just try to reskill for the next 10 jobs that get automated.

React is a great example of how 'it just works' can beat out 'elegantly designed' - even with all its quirks, it's still the most productive choice for most projects

still waiting for AI to replace the "let's have a 3 hour meeting to discuss a 5 minute decision" job

Academic publishing infrastructure still stuck in the dark ages. https://www.reddit.com/user/Opening_Fail5284

at my startup we used react for 3 years, then switched to vue and haven't looked back - not because react is bad, but because vue just makes more sense for our use case. boring technology works, folks.

I don't get why people still argue about this. We used template engines with a real database back in 2008 and they worked fine, no need to reinvent the wheel.

this ai stuff is all hype if you ask me. sure, the tech is impressive but it's not gonna replace real human intelligence anytime soon. we need to stop getting carried away and remember that behind all the fancy algorithms.

Finally, an embedded CPU that'll run my calculator watch from the 80s at blazing speeds. Who needs more cores than atoms in a grain of sand. Anyway?

code graphs are the future, its about time we had a playground to experiment with this stuff. finally, a way to visualize my spaghetti code.

sticking with GNOME, all these "" new DEs/WMs just seem to be solving problems no one has, meanwhile GNOME just quietly works.

I still don't understand why systemd is the default init system for most distros. It's complicated, buggy and just plain awkward. Anyone still using Arch or Gentoo knows what I'm talking about.

at my startup we spent years building a custom UI framework and it was a huge headache, but it's one of the best decisions we ever made, now that we're pivoting into a new market we're being told to use a UI kit from some company, classic HN comment

classic hype over large language models. sure, they can do some impressive stuff, but they're still just fancy text generators that struggle with long-term coherence and real understanding. i'll stick to good old-fashioned narrow AI for now.

boring technology works. chatbots and llms are not the future, they're just the latest fad. real innovation happens when you focus on the fundamentals and build practical solutions, not flashy demos.

injection is the root of all evil. just stop using it already, it's 2023 for crying out loud. https://www.reddit.com/user/Missics

this whole css thing is such a mess. i've been working with it for years and it still feels like a total hack. every project is a different set of custom styles and hacks just to get things looking right.

not this again. look, i get it, systemd can be a pain in the ass sometimes. but it just works for a lot of people. i've tried the alternatives and they all have their own quirks.

Still waiting for someone to explain to me how throwing more GPUs and params at a language model constitutes "advances in AI" rather than just incremental updates to a 60-year-old idea

I've been saying this for years. And we proved it in 2015. Temporal memory isn't nearly as hard as some people think. https://www.reddit.com/user/webman1972

metal water bottles are the only way to go. you can't beat the durability and sustainability. glass is a close second but plastic is just garbage.

python is the perfect language for beginners. it's easy to read and write, has a ton of libraries for just about anything, and the community is super welcoming.

Finally, someone figured out how to scale a research experiment to dozens of people. I bet this took at least 3 PhDs and a million dollars to accomplish what a decent undergrad could do with a few hours of coding.

People are so focused on the jobs that AI will replace, but what about all the ones it will create? We're already seeing new fields emerge, and I'm not convinced that we're even close to realizing the full potential of automation.

I still can't believe people are seriously considering using TypeScript in 2023. We made the switch to C# back in 2018 and haven't looked back.

Finally some sanity in the field, less focus on tricking models and more on actually helping them generalize. Data augmentation should be about exploring the manifold, not just generating more noise.