Gemini and ChatGPT and sound design? Are they better than Us?

In this video I'm asking Bard/Gemini and ChatGPT to give me precise settings of the described sound. I am then Manually setting up Ableton Live synthesizers to see if the sound is anything similar to what I described. Did it work? Is AI better than Humans?
---------------------------------------
Follow me on Instagram:
/ busiless.music
Also if you want to support my channel so I can make more and more awesome content you can buy me a virtual coffee ☕️
www.buymeacoff...
Keep making awesome music🎵
And see you in the next video! CHEERS 😎🤘

[@Gopalissocoollike]

Doooooood, sick video but for future reference the bell wavetable is under the instruments tab on wavetable (killed me that you searched for an hour)

[@Busiless]

Haha lol. Well not an hour. That was just to dramatise 😂 thanks. I’ll have a look again 😅

[@Busiless]

Freakin’ hell. It just… is there 🤦I seriously didn’t find it then 😂 anyways it wouldn’t change much 😅 I tested a bit ChatGPT 4 and it is a bit better. Maybe I’ll do another video 🤷‍♂️☺️

[@morizanova]

it possible that we can extract ableton instruments or synth preset data into text or csv format ? It would be fun editing from text format since we can doing intentional randomization

[@Busiless]

Hi. Didn’t try that, well that would be a good idea. The files aren’t really huge or anything. I’ll have a look a bit later. For now I’ve tried to convince Gpt to create a file that c I can download but no luck.

[@Busiless]

Ok. I've dug a bit and managed to take the drift preset in ADV format. It seems to be a zip file with a different extension. So I've changed the extension and uncompressed it and inside... There is a lovely XML file :D just without an extension again. But It's xml. I did try to force GPT to learn from that file and create another type of preset(then change extensions and compress again as it was) but Ableton doesn't want to open it. I'm sure there could be a way to do that though, just a better prompt or a different workflow. Let me know what you think that can be done here :D

[@funkahontas]

Great vid.
May I ask what you used for the dictation ?

[@Busiless]

Thanks. Just a Mac dictation option (sonoma) but works the same on iPhone and quite good in Gboard (Google keyboard) 🙂 I think they’ve put some kind of AI as I see it goes back to correct a word or two when gets the meaning of the sentence. All the best 🤘😎

[@basetpk]

you can also just tell it to generate the preset file. copy paste it and save it as a preset.

[@Busiless]

That’s a super cool idea. I’ll test it out for sure. Thank you! 🤘😎

[@Busiless]

Hi - I've just tried the ChatGPT (3.5 and 4) to give me a preset file for any ableton synth but no luck. Have you had any success with it?

[@basetpk]

​@@Busiless gpt3.5 wont do it. but I think 4 might. just play with it, try some prompts that start with "I need you to be, pretend to be, act " and activate neurons based on your needs. so a random reference to software/hardware knowledge is important. A good way to determine this is to identify your needs from a prompt , then identify areas of expertise in each part of the prompt. then mention those field names as part of your initial prompt. The beauty of LLMS is that you do not need proper grammar for the prompt to work.

[@Busiless]

@basetpk great thanks I’ll give it a go 🤘😎

[@mineralflex]

Is TH-cam really recommending good channels for me from newbies? I definitely like this video

[@Busiless]

Thank you! 🤘😎 check the rest of my videos. Lots of cool stuff for beginners ☺️ what music do you create?