I will say, for anybody reading and finds it in any way uplifting, I have been a Windows user for 30 years, been a .net developer for 5 years at one point, groaned at how bad the 'Linux desktop' always was, but this year I finally switched to using Linux instead of Windows and I think it's because the inflexion point is starting to hit more of the masses.
This actually. AI is dramatically better at helping users deal with perennial issues that Linux gets, and using the command line to fix them, compared to Microsoft's fresh new bug introduced last Tuesday or navigating anti-user GUI.
I've been amazed at AI's usefulness for Linux sysadmin tasks.
Last week I installed KDE on a new computer with two identical SSDs. One I wanted to wipe for the OS and the other I wanted to keep the data on.
Unfortunately I couldn't tell which one to install the OS on! ChatGPT helped me open a terminal and run the correct udevadm and lsblk commands to see what files existed on each disk.
Good use case and an example of brokenness of UX on Linux in many small and subtle ways. It frustrates me to no end that GNOME, Nautilus, etc. will truncate or hide information in many places. I don't think this has been fixed but the launcher will truncate names with ellipsis. So analyzer will appear as "anal..."
You’re absolutely correct what used to take me hours reading docs and googling now takes minutes to find an answer to. Just this week my new beelink I couldnt get hardware acceleration working for transcoding. A couple prompts lead me to the fact the gpu was new and I needed a newer Linux kernel. I probably would have spent hours and hours fiddling with device drivers and configs in the past.
>AI is dramatically better at helping users deal with perennial issues that Linux gets
This was my first thought. All Linux troubleshooting with AI resulted in hallucinatory commands that don't exist or work on this version of the OS, and managed to help me trash one system installation.
That sounds... quite bad, to be honest. I never met one single Windows issue that can't be answered by simple google search for years. A desktop expecting the user to ask AI(!) for solutions that require command line(!!) is definitely not for average users.
High user count will make much more information available, but once in a while you will hit one bad issue that is niche enough, and your only hope will be to reach MS.
At that point I'd rather deal with an LLM that know the insides of my open source OS than MS support. Have you seen their community answers on the web? I'd rather deal with Arch wiki and the command line even without an LLM.
- AI is what gave Nvidia reason to finally work on GPU drivers (Linux is premier environment for AI/ML workloads)
- LLMs are more useful on Linux than on Windows (although manuals are still unbeatable)
- Windows users are sick of Copilot bullshit
...yeah, I almost agree. the best thing that happened to Linux is Valve consistently pouring a ton of resources into open-source Linux projects, but AI bubble firmly holds the second place position.
Unfortunately Nvidia support mostly depends on Nvidia, they've done their best to keep info about their cards away from the Open Source driver projects.
Ditto, but .net dev for ~20 years, now fully Linux for personal compute. Workplace is making a beeline for mac / linux full stack, and completely ditching Windows.
Alas, while all new development is on .net core, the main money maker is a .net 4.8 website, so no way to move off windows for me, for now (no, I don't want to dual boot).
I personally went with https://endeavouros.com/ in the end, so far working really great, I appreciate Linux distributions will have different audiences which is great, I think regardless of distribution though it seems like a lot more software (and hardware) is 'just working' and it's actually delightful to use the computer again, kudos to many hard working people making free software
This has to be setting off some alarm bells internally, a well written postmortem on an occasional issue, great, but when your postmortem talks about learnings and improvements yet major outages keep happening, it becomes meaningless..
I think the point they are making us that the intermediate representation the model works with is parametric and then converted to step for use with other tooling, I could be wrong, but I understand the argument both ways of their solution enables editing of that parametric version before conversion out.
I think actually Linux has come a long way and recently I actually dual booted fedora with windows and fedora was easily my main choice unless gaming.. unfortunately when updating from 41 to 42 there was clearly an issue with the GPU not having drivers for acceleration or cuda, updating the drivers bricked the OS immediately and while I could recover, I spent hours and hours on this and could never get the GPU drivers installed again without bricking it.. ultimately I realised how at mercy of drivers Linux is. I hope though that in the next few years things improve as windows is dismal to work on these days
I just had a problem with Windows and Nvidia drivers/CUDA not working properly on a two year old Windows 11 install. I had to reinstall the operating system after days of troubleshooting and attempting different things to get it operational again. It can happen on there as well.
This was my view once, but practically the languages have their own ecosystem strengths, now I think most of the major languages have similar concepts, async await, etc.. instead we now teach transfer your skills, and tbh I think it makes for a more interesting culture that's adoptive and excited by tech, and also learns better
It’s called being a software developer instead of being a JavaScript developer. If someone knows how to engineer software, their skills are transferable. Stitching libraries together is usually not software engineering.
software engineering is mostly a kludge of interacting with stuff people have made and will make, rather than physical limits of different materials and structures
stitching libraries together is the essense of soctware engineering
Yeah, I was concerned about that too. It looks like it has a small thin edge connector on the body of the Switch 2, sort of like a USB-C port but without the protective shield around it. If it's not designed well, we could see it snapping off in kid's hands and requiring expensive repairs.
I think the most exciting aspect here is this might actually push JavaScript to add some missing features that can make a lot of difference, the pipeline operator for example can really improve the design of JavaScript code and has been proven already in several languages, C#, swift, kotlin to name a few I know
