Reply
  • Oct 5
    ·
    1 reply

    I was inspired by how ye "produces" things by just barking orders at some white kid on a computer
    he ultimately is doing what this program will allow you to do

  • Oct 5

    you can provide a personal idea or just have it build from scratch

    it parses your drum/sounds by bpm and style

    has a pretty good dictionary for what certain sounds are described as but can be expanded and manipulated to individual user

  • Oct 5
    ·
    edited
    ·
    1 reply
    Merged Grok

    yup exactly

    has a shared UI so you and the agent are seeing the same layers on the timeline

    you can also take manual control if you want

    I'm not fully tapped in, but couldn't this already be achieved via Computer Use mode on Astra? Since you can talk to it and let it interact with your computer/software in real time

    Even if the answer is yes, I can still see the value in creating a standalone, music-focused platform for the idea.

    I honestly think some version of this will make up a large portion of the music creation of the future, as opposed to Suno-style, fully generated tracks being the norm (though that will absolutely still be a massive chunk of it)

    Don't know how I feel about it from an artistic perspective. It's leagues more artistic than Suno/fully generated music to me, because you're acting more like a composer than a prompt writer. It still removes the blood, sweat and tears that make art so important though.

    Not saying that as a moral judgement on you or anything. Much, if not most popular music already outsources 95% of the creation process to faceless label drones, I guess this just allows you to do it yourself. Just thinking out loud about where music as an art is headed.

  • Oct 5
    Merged Grok

    I was inspired by how ye "produces" things by just barking orders at some white kid on a computer
    he ultimately is doing what this program will allow you to do

    Yea I see what ur getting at, kinda what I'm saying with the "composer" comparison

  • Oct 5
    ·
    1 reply
    Orangutan

    I'm not fully tapped in, but couldn't this already be achieved via Computer Use mode on Astra? Since you can talk to it and let it interact with your computer/software in real time

    Even if the answer is yes, I can still see the value in creating a standalone, music-focused platform for the idea.

    I honestly think some version of this will make up a large portion of the music creation of the future, as opposed to Suno-style, fully generated tracks being the norm (though that will absolutely still be a massive chunk of it)

    Don't know how I feel about it from an artistic perspective. It's leagues more artistic than Suno/fully generated music to me, because you're acting more like a composer than a prompt writer. It still removes the blood, sweat and tears that make art so important though.

    Not saying that as a moral judgement on you or anything. Much, if not most popular music already outsources 95% of the creation process to faceless label drones, I guess this just allows you to do it yourself. Just thinking out loud about where music as an art is headed.

    we need you on the team for soul preservation purposes

  • Oct 5
    ·
    2 replies
    Lowtierwhitegirl

    Currently building a model from scratch for spelling correction

    The detection of misspelled words is rly good but the correction is a pain in the ass

    theres this thing called spellcheck that exists on nearly every electronic device you might find useful

  • Oct 5
    Merged Grok

    we need you on the team for soul preservation purposes

    My first act as CWO (Chief Wokeness Officer) would be to position the tool as an alternative to the soulless Suno

    It only assembles and edits pre-existing parts (instruments, loops, samples, etc.), does mixing and mastering on the fly and lets you hear side by side comparison, can apply and tweak plugins/eq and let you hear the side by side

    Maybe even tell it what you're looking for, and it goes out and scours marketplaces to retrieve options. So say you want a synth that sounds a certain way. You describe it, it goes and finds options and retrieves their previews so you can listen to them, you pick one, and it either downloads it if it's free, or prompts you to purchase if it's not

  • Oct 5
    Sir Real

    theres this thing called spellcheck that exists on nearly every electronic device you might find useful

    we need AI spell check

  • Sir Real

    theres this thing called spellcheck that exists on nearly every electronic device you might find useful

    The language used is from a specific technical domain i dont think spellcheck would work

    Also spellcheck isnt as good at reading context as llms so it wouldnt necessarily be that accurate

  • Oct 5
    ·
    edited
    ·
    1 reply
    Merged Grok

    im developing a program that allows you to upload your stems of a half made beat along with the drum/sound packs you used and then you can dictate changes to it

    inspired by ye

    !https://www.youtube.com/watch?v=DYbER5mjttI

    How are u building this?

    Just curious

    U dont have to go into details but like what software or platform etc

  • Oct 5
    ·
    1 reply
    Lowtierwhitegirl
    · edited

    How are u building this?

    Just curious

    U dont have to go into details but like what software or platform etc

    what are you asking specifically?

  • Oct 5
    ·
    1 reply
    Merged Grok

    what are you asking specifically?

    Like what tools/software do u use?

  • Oct 5
    ·
    1 reply

    What i build with Claude is between me and me

  • Oct 5
    Gandalf The Black

    What i build with Claude is between me and me

    are you building cool s***?

  • Oct 5
    ·
    1 reply
    Lowtierwhitegirl

    Like what tools/software do u use?

    a Python audio engine (offline render, so there's no DAW to learn and it runs anywhere), a web console for the session board layer stack, waveforms, mixer, timeline, per-layer solo and an agent running the session in the background

  • Oct 5
    ·
    1 reply

    I don’t think it’ll permanently live in-browser
    but it does actively

  • Merged Grok

    a Python audio engine (offline render, so there's no DAW to learn and it runs anywhere), a web console for the session board layer stack, waveforms, mixer, timeline, per-layer solo and an agent running the session in the background

    Word sounds cool are u training the model urself or using something pretrained

  • Oct 5
    ·
    1 reply

    Honestly know nothing about agentic ai i build transformer models for science s***

  • Oct 5
    ·
    1 reply
    Merged Grok

    I don’t think it’ll permanently live in-browser
    but it does actively

    you can use tauri to make it a desktop app

  • Oct 5
    ·
    1 reply
    Lowtierwhitegirl

    Honestly know nothing about agentic ai i build transformer models for science s***

    Which field ?

  • Oct 5
    ·
    1 reply
    Classique

    you can use tauri to make it a desktop app

    Not aware of this but I will look into it thank you

  • Merged Grok

    Not aware of this but I will look into it thank you

    it's like a more modern electron type thing and faster than webview i believe

  • Oct 5
    ·
    1 reply
    Merged Grok

    Which field ?

    Biology

  • Oct 5
    Lowtierwhitegirl

    Biology

    Nice thats super interesting
    that is an application that im not super knowledgeable on but would love to learn more
    any papers you recommend?

  • Oct 5
    ·
    1 reply

    /agents on claude is an overkill for anything im building. I know theres way more advanced and sophisticated stuff but claude does anything