Edit: old post is someone is curious: https://dev.to/jamilxt/electron-vs-gpui-what-discords-290-mb-rust-rewrite-can-and-cannot-prove-34b6
Replaced with the one from The Verge
Edit: old post is someone is curious: https://dev.to/jamilxt/electron-vs-gpui-what-discords-290-mb-rust-rewrite-can-and-cannot-prove-34b6
Replaced with the one from The Verge
My personal opinion is that LLMs shouldn’t be used at all except for certain use cases, namely translation and processing large data sets (computers are good at repeating tasks, LLMs are no different). However, of course, some people will be using LLMs for programming. I do not and am against it, but people do what people do, however there is a difference between using it a tool and claiming that it will replace human programming. Neither is particularly amazing in my book, but almost everybody would agree that the latter is much, much worse, for jobs, the maintainability of the codebase, and user experience (broken apps are bad!)
Epic Games are claiming to “vibe code” all their software and repeating the rhetoric of the likes of OpenAI and Anthropic. Turns out, everybody does not like it when companies ship broken software that doesn’t work at all, and boasting about it will hurt your image.
I would imagine most people would prefer supporting someone who doesn’t donate to racists! LLMs, especially corporate ones, are typically associated with the far-right because it’s literally giving a handful of companies supreme power over your information, your code, your independent thinking, and most importantly in this case your creativity (a fascist’s wet dream).
Also, you seem to have ignored my main points about Valve. Their monopoly is bad, we should support other third-party stores like itch.io and GOG more! You have also ignored Valve’s contribution to Linux gaming and their genuine innovation in hardware (see Steam Deck, Frame VR headset, Steam Controller 2, etc.)
I respect that opinion & understand it. The reality is though that 90%+ of all programmers are using LLM to -at the very least- assist them with their work. Probably we are talking about well over 90%.
Yeah the difference is about 6 months of model releases.
Seriously, you would have been totally correct 6 months ago (or if you are still using models from 6 months ago). Now… not so much.
Remember when we all laughed when Anthropic tried to vibe code a C compiler and it was laughably broken? Yeah now they can vibe code an entire OS.
(They should really run that C compiler test again just to show how much progress they’ve made.)
Again, you have dodged all my other points. Plus, using it as a tool (checking for bugs, running tests, that sort of stuff) is very different from telling your staff to only vibe code software!
And it’s still not working, the Linux client for Epic Games failed spectacularly with lots of issues. The recent vibe coded replicas of Photoshop and friends have lots of bugs and crashes too (morally grey area in my book due to Adobe being an awful company, less people paying their exorbitant subscriptions is always better, but undeniably the “clones” are broken)
LightCraft was garbage 2 weeks ago, now it’s approaching something actually usable.