Minor correction: "large"/"small" refers to the size of the model, not the amount of data it was trained on. Larger models will require more processing power. Presumably these are neural network models, so they should be far faster on a GPU, assuming the software is configured to use it right.
I like that you're trolling DT :). I like a lot Linux content on UA-cam, but I only follow 4 channels with Linux content and yours is one of them and DT's is one of the others
It does use the GPU if you activate it in the settings. AFAIK, it only works on nvidia cards. Note that you need like 12 GB of vram to run the large model. So, even with NVIDIA, you need one of the top cards to actually use the largest model.
Hey Matt, you need to move your „leave a thumbs up“ beyond the 3 minutes mark. Apparently, YT doesn’t count any like within the first three minutes. I always liked my regular channels straight away and I always wondered why most of them not even got past the 10% views-to-likes ratio. Might be the reason. Now, I got used to like further in the video or at the end.
I dont think this is a note taking app... I dont see any font editing/syntax highlighting or anything. It's more like a general speach and languages app. So while it's cool and I'm happy you've shared it with us, I think you might be framing it wrong! (Maybe it's SpeachNote, as in a musical note?)
I find it to be the best. Many offer similar or even slightly better features but none offer the full package - PC app, phone app, self-hosted docker container for reliable sync, E2E encryption, note sharing with other users and public note publishing if you want to share something outside the Joplin ecosystem. All without any 3rd party being involved with your data, all in two clicks. And the PC app theming capabilities are awesome, one CSS file and your notes look exactly the way you want.
01:42 text to speech works probably faster than typing a text. Maybe in the future, you don't need to type fast. Is there a markdown editor, which supports text to speech?
Want to help support the channel? Check out our merch shop! shop.thelinuxcast.org
lol "screw you DT" :D
i've been hunting for an app like this, would be really nice to see a option for the user to use their own local LLM to drive it
Excellent review, thanks!
Minor correction: "large"/"small" refers to the size of the model, not the amount of data it was trained on. Larger models will require more processing power. Presumably these are neural network models, so they should be far faster on a GPU, assuming the software is configured to use it right.
I like that you're trolling DT :). I like a lot Linux content on UA-cam, but I only follow 4 channels with Linux content and yours is one of them and DT's is one of the others
"screw you DT" hhhahahahaahahahahahahahhahaahaahahhaaahahahahhaahhaha
I just started using Joplin. It saves notes just fine
And yes, I finally fixed the dumb light reflecting in the monitor behind me.
Those neural network models should use your GPU, it would be much faster and much cheaper computationally.
It does use the GPU if you activate it in the settings. AFAIK, it only works on nvidia cards. Note that you need like 12 GB of vram to run the large model. So, even with NVIDIA, you need one of the top cards to actually use the largest model.
I don't think I'd use the speech recognition but the TTS models are so much better than espeak...
Definitely a cool app, that I will definitely install. Thanks for the video!
Actually one more feature would be great, passing an audio file. Then we can transcrive podcasts and read through it faster than listening to them.
You can do that. In the main tap, just choose that option from the file menu.
Well, at least it has good ASMR voice
the TTS is a game changer
Hey Matt, you need to move your „leave a thumbs up“ beyond the 3 minutes mark.
Apparently, YT doesn’t count any like within the first three minutes.
I always liked my regular channels straight away and I always wondered why most of them not even got past the 10% views-to-likes ratio. Might be the reason.
Now, I got used to like further in the video or at the end.
I really like what you do, but I don't understand why the dig at DT?
It was supposed to be funny. DT is a good dude.
Read the thumb nail and thought everything ever made was cooler the Notes, (Lotus that is).
I dont think this is a note taking app... I dont see any font editing/syntax highlighting or anything.
It's more like a general speach and languages app. So while it's cool and I'm happy you've shared it with us, I think you might be framing it wrong!
(Maybe it's SpeachNote, as in a musical note?)
It calls itself a note taking app, so that one isn't on me. I just read what's on the tin
11:38 For me, this is a deal breaker. I will not use this software. Thank you for your review.
Have you looked at Joplin yet?
I find it to be the best. Many offer similar or even slightly better features but none offer the full package - PC app, phone app, self-hosted docker container for reliable sync, E2E encryption, note sharing with other users and public note publishing if you want to share something outside the Joplin ecosystem. All without any 3rd party being involved with your data, all in two clicks. And the PC app theming capabilities are awesome, one CSS file and your notes look exactly the way you want.
It would be really nice if they added a plugin to obsidian. I would like this but I already committed to using obsidian for my notetaking.
there is a copy text button down to the right
You use Tumbleweed ? What are your experiences with prolonged use ?
this is neat, but kind of a gui frontend to whisper? Not much of a note taking app. Nonetheless seems fairly alright. Is available as Flatpak.
speech note 4.4.0 has save text file
Nice wallpaper. Where can i get the same?
Pen and paper.
1:57 SpeechNote almost sounds like a Russian accent
14:25 it's an open source application, you should report it as a bug and preferably even try to come up with a fix yourself.
I made a custom model for piper tts and bundled it into an appimage alongside a gui. Its great! but the default models are not as good.
01:42 text to speech works probably faster than typing a text. Maybe in the future, you don't need to type fast. Is there a markdown editor, which supports text to speech?
You mean speech to text?
@@kuwandak Ah, yes, you are right. My bad, sorry. Thank you very much for pointing this out.
03:08 #DTbashing
obsidian >
I was gonna ask for Org...
Hey Matt, what VPN do you use with opeSUSE?
Mullvad with wireguard
@@TheLinuxCast Thank you, I appreciate it.
Great question, I'm glad you asked!
Dayum, beef with dt?
Lets goo linux drama
I was just being funny. Or trying to be
@@TheLinuxCast yeah sorry, that was my attempt at making a joke/comedy phrase