Zed Editor and AI Models
Вставка
- Опубліковано 9 чер 2024
- Exploring Zed Editor and AI Models locally installed
00:00:00 - Zed
00:04:31 - Superhuman Email
00:07:59 - Workspaces
00:08:45 - Building Zed
00:10:25 - Using AUR for Zed Build
00:24:49 - Initial Launch
00:30:08 - Extensions
00:34:02 - VIM Bindings
00:40:51 - Meta Llama
00:45:20 - Bazzite Update
00:54:04 - Starship Prompts
01:17:46 - Adding AI to VSCode
01:31:25 - Adding Ollama for Local AI
02:10:18 - Broke Bashrc
02:35:47 - Wrapping up.
►► Digital Downloads ➜ christitus.com/downloads
►► Reddit ➜ / christitustech
►► Chris Titus Tech ➜ / @christitustech
►► Twitch ➜ / christitustech
►► UA-cam Early Access ➜ / @titustechtalk - Фільми й анімація
48:20 he did, and he did. He managed to make the system unbootable even though the system was warning him repeatedly that what he was doing was a terrible idea.
1:02:47 Really nice configuration!
A little late to the show... a quick tip about ollama: the models it uses are quantized (they reduce their precision to make them smaller - it's an optimization thing). That 40G Llama3 70b model is quantized to 4bit per parameter. The original in FP16 format is about 130G in size. Same goes for the 8b version. Quantized to 4 bits is about 4G in size, while the original is 15G. Ollama also provides different quantization levels (see the tags for each model on their website).
Man.. CentOS going EOL. That takes me back to my early days starting off with RHEL3 and then CentOS4.2. Heck, I loved it so much, it became my daily driver.
I have used helix for 3month and i think it is solid text editor but not yet fully ready
I like your off-channel videos because you're more "unhinged"
one question, could you show one time your mic setup and audio interface you are using which works on linux, because it sounds so nice how did you made that sound so good ?
Linux audio isn't great and I used to pull it in via analog connection. If I were to do it in Linux, I'd get a sound card with a S/PDIF input and grab the audio from the audio interface to it.
However, I do not use Linux anymore for audio due to the limitations and lack of options. I moved to a PreSonus audio interface that uses USB 3 and it puts independent inputs in to a Windows System. This allows me to do multichanneling in OBS with Mics on Track 1, PCs on Track 2, and BG Music on Track 3. This allows me to remix the streams after they end and fix any audio issues as well. Another thing I do that most people omit is Audio Ducking. When I speak I make all the other tracks quieter so my voice will always punch through.
For Hardware I'm not a fan of the most popular mic (SM7B) as it sound a bit muddy, but has excellent background rejection. I prefer for a mic that grabs voice better the Heil PR40 with a Soyuz Lifter. Gives a cleaner audio track, but does pickup more background noise.
@@TitusTechTalk Why would I want to get an extra sound card for this ? I have a class compliant audio interface which works without any problems!
But thanks for the answer
Tried to watch your Arch Linux gaming video but it and several others are set to private. Is there a way I can watch it?
I private those after the stream finishes and edit them to remove all the dead air and reupload after edits are done about 3-7 days later.
@@TitusTechTalk OK cool. Thanks for the reply and letting me know. I appreciate it. Really dig your videos.
@@TitusTechTalk can you do an updated review of lapce
Zeditor!
"The speed of thought". For most of us, that's not very fast.
The whole premise of these things is a fallacy: most developers are actually capable of writing code faster than they "think", which is precisely the problem with software quality. It's not the coding part that hinders us, but the thinking part.
The vast majority of develops use Macs, especially in web dev, so they're developing for their largest platform first. No one is pretending anything; in fact, the least used platform when it comes to dev work is Windows when you take into account all ecosystems. But it's not just web development either. For example, the Go Developer Survey put macOS just behind Linux as the platform they use. The Zed team is also building a lot of things from the ground up (i.e. text rendering, GPU acceleration) because they're using Rust (GUI applications in Rust don't have that mature of an ecosystem) and thus it's actually a lot harder to create a cross-platform application out of the gate. With all of that said, Zed has actually made an astronomical amount of progress towards a Linux client since open sourcing it - must faster than I expected at least.
Interesting, I was wondering about why MacOS had such a strong focus. The project is very cool.
Stack Overflow Developer Survey 2023 shows 46% (Win), 33% (Mac), 26% (Ubuntu), 15% (WSL).
As a Linux user, these are sad statistics to hear. 🤣
VS Code uses GPU acceleration. So does Visual Studio itself, and has done for years. It’s an irrelevant feature.
Wrong. Windows is 45% development environment share and if you combine wsl at 15%, you have total 51% market share in development
First 🎉🎉🎉
Why you give yourself all of this trouble,
you can simply install vs code with the vim plugin it will emulate vim key bindings
with a much easier code editor to work with
For CONTENT! Also, I like trying new things and seeing alternatives.
Your memory hogging vs code 😂
@@weiSane
Completely true
But does we even have a choice 😊
@@ahmedthebestsupport zed it will be better than vs code in a few years once it matures
First.
It’s because Linux and Windows SUCKS! Period!
I once heard that the only thing Windows and Linux users can agree on is that MAC sucks. 😉
@@tellsackett4572 no lol macOS is 10x better than windows for everything except gaming.
@@happygofishing but Linux is the king!