SCOTT

SCOTT

@scottinallca.ps

tech/music/ai :: grimes, ktt, createsafe/triniti, lowkey. I like making cool stuff.

350 followers110 following20 posts

SCOTT
@scottinallca.ps

i introduced him to chatgpt a couple months back and he was blown away today showed me he uses it all the time for stockroom stuff and everything. people really aren’t ready for this tech. “we’re so early” etc

SCOTT
@scottinallca.ps

what’s the fastest way to extract audio of a target speaker from 100+ of hours of audio? the speed with pyannote on my m3 max is way too slow compared to the same code on an nvidia gpu, to the point that it would take days atp i’m just wondering if there are any easy proprietary solutions for this

SCOTT
@scottinallca.ps

concept: an infinite sound library for people who make music everything is generated on the fly with LLMs using implicit prompting from navigation. ai music tools should be more like this — augmenting workflows, not replacing artists note: this is just a clunky proof of concept with no audio

SCOTT reposted

victor
@ifvictr.com

Replying to @generativist.com

i use them to: - sand down my raw thoughts - revisit past social interactions to surface any nuances/patterns - run simulations of novel situations so i know how to navigate them when they actually happen. it’s a sandbox for decision-making and strategizing in the form of a text adventure game

You are up to date.