The way you present screencapture and the command line (like at 2:10) is really creative. Visualizing programming on video in an engaging way is so difficult, and I feel like you’re onto something with your style. Definitely inspiring as someone trying to do something similar with my channel, so I respect the extra editing effort that went into this. Awesome. Subscribed!
Thanks buddy! Checkout your channel looks amazing! Yep, trying to spice things up with the terminal and how it's presented. Got a lot of inspiration from Takuya at @devaslife check it out if you haven't yet :)
I really suffer with channels with poor video quality. Good content and good video quality must walk together in order to deliver knowledge in a pleasure way. Thank you both for that.
@@allsunday1485 Got it a few times in the past and trying to play with things creatively while not making things worse for those who are easily distracted. Thanks for the feedback! I’ll take this into account in future videos
The resources consumed by ollama is not trackable with procs (or any ps tools like htop btop etc) because the model is on the GPU whenever you use it. You will need to use the specific tool for your GPU in order to know how much resource it's eating, nvidia-smi for nvidia cards, and radeontop for AMD cards
@@devopstoolbox on emacs there's something called ellama very similar to gen , same idea , altho this is amazing and i can see a bit of a future , it doesn't stick you know i don't see myself using it i don't know why , feels like faster to go and google some example or something like that
Hey, as you are DevOps Engineer you can make an video on "Good practices for DevOps Engineers" or any tips and tricks which can make things easier for them. *Like this one.
This is such a broad range to cover that I try to focus on these in my blog + newsletter + podcast, you can check them all! That said, I will try to incorporate more specific videos. omerxx.com topeaks.buzzsprout.com
IIUC this would require the model to keep running in the background, i.e. the container lives throughout. This means you'd either have to snapshot it when restarting /periodically / use some other system to make sure the context is chaced. That said, you may want to check with gen.nvim's authos / ollama docs to make sure.
There is a new project for nvim and ollama ogpt.nvim (fork of the chatgpt.nvim) with more advanced configuration and predefined roles than gen.nvim. Unfortunately, I have found a few bugs in it.
This is awesome. Thanks to the ollama devs and devops toolbox for shading light into it. Just one question, have you tried the starcoder model? it suppousedly outperformed other open code models last time I read about it but havent confirmed it
great video and channel! subscribed! love the keyboard! given you are a coder and neovim user, any chance you can do a video on your layout? thinking of buying one but worried about rebuilding muscle memory ...
Yep! Working on it but I want the layout to be dialed in perfectly before I do and that takes time. Muscle memory rebuilding is an integral part of the process but I must say it’s not that long. I reached same typing speed after ~3 weeks. The layout keeps changing as I learn / find new stuff. So I’m guessing only if another 3 weeks go by without me changing anything I’ll know it’s ready for review 😅
Thanks! Haha still playing around with getting used to the moon lander. It feels to me it has too many keys in a way that sometime is hard for me to understand "where I'm at" just by touching it. I'm used to knowing where the boundaries are when I reach out to the shift key for example or g on the other end. When shift is just another small key and there's more underneath it it's hard for me to find my way around without looking and I couldn't get used to it so I deicded to try and run with out it. Besides it doesn't make 100% sense to me to reach all the way down there so I removed it altogether :)
Planning to! It’s very much a mess rn so I’ll fix it up and share. You may have seen that I’m still playing with removing keys altogether and checking how it works
Great video all round, however, I had a slight issue with getting mistral to work... For some reason when I try to use the :Gen command, after entering a prompt (if applicable) I am presented with a blank box that does nothing. I have no idea what this means?
Hmmm try to debug: if you run the model through the CLI, can you get an answer from mistral? If you can’t, see if the mode is running or there’s a local issue. As for gen - check the config and see what mode it’s running
Of course! Always open at dotfiles.omerxx.com Might not have the Ollama config yet as I haven't finalized it but if this is what you're looking for I'll expedite things and push. LMK!
@@devopstoolbox Amazing, thank you very much. I am learning NVIM and entirely new so will have a play around with your config :) Appreciate it. Subbed as this is very useful content!
Oh, so the moonlander was at the end. Happy for your wrists. Now get into the rabbit hole of one shot layers, sticky modifiers and combos for maximum procrastination😄
AI tools suck without a front-end simply because terminal cannot format markdown. Sure this might come off as a skill issue, but I want to easily distinguish between explanation and code examples.
Firstly, there are great options, such as github.com/charmbracelet/glow But regardless, you can design the prompt to provide the output structure relevant. Especially if coding and you need either autocompletion / example with comments instead of MD etc
He should be able to rest his fingers and wrists by speaking words to the computer instead of typing on keyboards so much. He will save a lot of time, too.
Join my newsletter!
newsletter.omerxx.com
Thank you so much for featuring my plugin ❤
Thank YOU for creating it 🙏🏽
what AWESOME editing. These videos are looking amazing!
Thank you!! 🙏🏽
omg what a crossover
The way you present screencapture and the command line (like at 2:10) is really creative. Visualizing programming on video in an engaging way is so difficult, and I feel like you’re onto something with your style. Definitely inspiring as someone trying to do something similar with my channel, so I respect the extra editing effort that went into this. Awesome. Subscribed!
Thanks buddy! Checkout your channel looks amazing!
Yep, trying to spice things up with the terminal and how it's presented. Got a lot of inspiration from Takuya at @devaslife check it out if you haven't yet :)
@@devopstoolbox Oh cool, I’ll have to check them out too. Thanks for the rec
I really suffer with channels with poor video quality.
Good content and good video quality must walk together in order to deliver knowledge in a pleasure way.
Thank you both for that.
I find it distracting tbh
@@allsunday1485 Got it a few times in the past and trying to play with things creatively while not making things worse for those who are easily distracted.
Thanks for the feedback! I’ll take this into account in future videos
The resources consumed by ollama is not trackable with procs (or any ps tools like htop btop etc) because the model is on the GPU whenever you use it. You will need to use the specific tool for your GPU in order to know how much resource it's eating, nvidia-smi for nvidia cards, and radeontop for AMD cards
Yes!! Thank you! I was baffled for a while and was definitely waiting for this comment !
Aside from another awesome video produced by you, what an amazing view from your desk...!
😊 thank you!!! 🙏🏽
Just what I needed, stay fully in the terminal… thanks for sharing
Strange that there's no resource consumption issue on your side, somehow on my end it's chewing 7GB+ of Memory... wondering what im missing
Thanks for the channel.
I'm on my first steps on Odin Project.
This will be useful soon.
do just the foundations then skip the rest and do the full curriculum at fullstackopen
This is gold! Can’t wait to try it out! 🎉❤
Let me know how it goes!
i think ollama has been written by ex-docker employees also!
Honestly the number of amazing projects built by ex docker people is astounding.
@@devopstoolbox on emacs there's something called ellama very similar to gen , same idea , altho this is amazing and i can see a bit of a future , it doesn't stick you know i don't see myself using it i don't know why , feels like faster to go and google some example or something like that
So what do you prefer now? Codeium, Copilot, or local llama?
Codeium for everyday code completion, GPT when I’m writing text (blogs newsletters etc)
Hey, as you are DevOps Engineer you can make an video on "Good practices for DevOps Engineers" or any tips and tricks which can make things easier for them. *Like this one.
This is such a broad range to cover that I try to focus on these in my blog + newsletter + podcast, you can check them all! That said, I will try to incorporate more specific videos.
omerxx.com
topeaks.buzzsprout.com
Ollama literally uses docker. They are docker images, that's why it feels like docker 😊
Would it be possible to to expand on how to use this with local files as a callback for keeping the model learning?
IIUC this would require the model to keep running in the background, i.e. the container lives throughout. This means you'd either have to snapshot it when restarting /periodically / use some other system to make sure the context is chaced.
That said, you may want to check with gen.nvim's authos / ollama docs to make sure.
There is a new project for nvim and ollama ogpt.nvim (fork of the chatgpt.nvim) with more advanced configuration and predefined roles than gen.nvim.
Unfortunately, I have found a few bugs in it.
This is awesome. Thanks to the ollama devs and devops toolbox for shading light into it. Just one question, have you tried the starcoder model? it suppousedly outperformed other open code models last time I read about it but havent confirmed it
I did not! But now that you mention it I'll give it a go
Niceeee you got a moonlander!! Best piece of gear EVER
Yes! Replaced all the switches and still need to pad the inside so the typing won’t be as “metallic” but it’s amazing
@@devopstoolbox I got the kalith deep sea blues and they are **chefs_kiss
great video and channel! subscribed!
love the keyboard!
given you are a coder and neovim user, any chance you can do a video on your layout?
thinking of buying one but worried about rebuilding muscle memory ...
Yep! Working on it but I want the layout to be dialed in perfectly before I do and that takes time. Muscle memory rebuilding is an integral part of the process but I must say it’s not that long. I reached same typing speed after ~3 weeks. The layout keeps changing as I learn / find new stuff. So I’m guessing only if another 3 weeks go by without me changing anything I’ll know it’s ready for review 😅
@@devopstoolbox looking forward to it! although i ll probably buy one before that :P
@@NicholasAngelidis1 classic move Id make 😅
Great, great editing. Please share the color theme you're using in Neovim. Thanks!
Of course! Cattpuccin!
@@devopstoolbox ... And your split keyboard exact model ❤❤❤
@@elatedbento ZSH Moonlander!
how is the moonlander?
The moon lander is amazing! I must say the HHKB was a superior typing experience but nothing beats a good split kb with thumb clusters 🤩
Loved this ! Also why is your keyboard missing some keys xd ?
Thanks! Haha still playing around with getting used to the moon lander. It feels to me it has too many keys in a way that sometime is hard for me to understand "where I'm at" just by touching it. I'm used to knowing where the boundaries are when I reach out to the shift key for example or g on the other end. When shift is just another small key and there's more underneath it it's hard for me to find my way around without looking and I couldn't get used to it so I deicded to try and run with out it. Besides it doesn't make 100% sense to me to reach all the way down there so I removed it altogether :)
@@devopstoolbox Ahh I see it now, also that view look soo good to relax your eyes after long hours of sitting, I really loved it !
This is so cool, thank you!
🫶
Can you share your moonlander keyboard layout?
Planning to! It’s very much a mess rn so I’ll fix it up and share. You may have seen that I’m still playing with removing keys altogether and checking how it works
hey, super cool video. could you tell me what nvim plugin you used to achieve that sleek terminal interface from neovim?
Thanks! Here it is
github.com/numToStr/FTerm.nvim
Awesome content, thank you!
which wrist rests are you usign with your moonlander?
It comes with wrist rests that do a pretty good job. Used to have a wooden one for the HHKB that felt softer but no complaints
@@devopstoolbox thanks Omar and thanks for the awesome videos!
Great video all round, however, I had a slight issue with getting mistral to work... For some reason when I try to use the :Gen command, after entering a prompt (if applicable) I am presented with a blank box that does nothing. I have no idea what this means?
Hmmm try to debug: if you run the model through the CLI, can you get an answer from mistral? If you can’t, see if the mode is running or there’s a local issue.
As for gen - check the config and see what mode it’s running
@@devopstoolbox I can run the model fine in the CLI, no problems there, I'll check the config for the mode now.
@@devopstoolbox Update: Even with specifying mistral:instruct in the gen config, it still fails to work correctly. Very strange.
@@billbennings8990 do other models work okay? If you switch to codellama for example does it work?
I'm installing it now, good idea.
Is it possible to share your nvim config setup, I really like how your setup? Thanks :)
Of course! Always open at dotfiles.omerxx.com
Might not have the Ollama config yet as I haven't finalized it but if this is what you're looking for I'll expedite things and push. LMK!
@@devopstoolbox Amazing, thank you very much. I am learning NVIM and entirely new so will have a play around with your config :) Appreciate it. Subbed as this is very useful content!
Oh, so the moonlander was at the end. Happy for your wrists. Now get into the rabbit hole of one shot layers, sticky modifiers and combos for maximum procrastination😄
Trying hard 😅
man the keyboard view is sick.........
give me keyboard link ....
ZSH Moonlander!
The moon lander is amazing! I must say the HHKB was a superior typing experience but nothing beats a good split kb with thumb clusters 🤩
@@devopstoolbox I gues su meant ZSA 😂 .......zsh is overtaking
@@devopstoolbox yeah seems pretty awesome..... Gonna check it out soon....
AI tools suck without a front-end simply because terminal cannot format markdown.
Sure this might come off as a skill issue, but I want to easily distinguish between explanation and code examples.
Firstly, there are great options, such as github.com/charmbracelet/glow
But regardless, you can design the prompt to provide the output structure relevant. Especially if coding and you need either autocompletion / example with comments instead of MD etc
Bro said he wants to bump a minor 😮😅
Haha finally! The one who noticed!
He should be able to rest his fingers and wrists by speaking words to the computer instead of typing on keyboards so much. He will save a lot of time, too.
noice
11 minute video for a 30 seconds of content, yikes. and then doesn't even explain anything? dafugg. very proud of his keyboard, i guess. weird video
first
😅🫶
noice