AI Music Revolution

Not Cheating. Directing.

• Josh • Episode 31

Share episode

Use Left/Right to seek, Home/End to jump to start or end. Hold shift to jump forward or backward.

0:00 | 10:41

Send us Fan Mail

This week's recap covers two posts that are really one idea. Why the voice telling you AI music "doesn't count" is aimed at the wrong target, and the single decision, made before you ever open the platform, that separates a Director from a Vending Machine Operator. If you read the posts, this goes deeper. If you didn't, you're getting the whole thing here.

Mentioned in this episode:

  • Blog: You're Not Cheating, a Case for AI as an Instrument, From Someone Who Paid Their Dues
  • Blog: Director vs. Vending Machine Operator
  • The Red Lab Library
  • The AI Music Release Lab (founding cohort, status in show notes)

The Unlock System is JG BeatsLab's methodology for serious musicians working with AI tools. Lane 2 work: human-authored, AI-assisted music creation.

Visit JG BeatsLab: https://www.jgbeatslab.com

Every Thursday morning, I send out a newsletter. One email a week. It's where I share what I'm working on in the studio, the research running in the lab, and the methodology I'm refining in real time. New books, new Blueprints, new findings from the Red Lab Protocol research. If that sounds useful, you can sign up at jgbeatslab.com/newsletter. Thanks for listening, and I'll see you next Friday.

AI Music Revolution is produced by JG BeatsLab LLC, an AI music education company building the methodology, research, and community for serious creators working in Lane 2.

Get more from JG BeatsLab LLC:

Connect:

Contact: josh@jgbeatslab.com

Stop gambling. Start directing.

SPEAKER_00

Hello and welcome to the AI Music Revolution. I am your host, Josh Gillalian, the founder of JG Beast Lab. Before we get into this week's recap, something I haven't talked about publicly yet, and I'll be honest, it feels a little strange bringing it up because it's my own numbers, and that always risks sounding like bragging. But I think it's worth sharing, not as a brag, but more as proof that means more than just me telling you that I know what I'm doing. So one of my own AI-assisted bands, it's a band called Anchor Down. It has a track called Wired Nerves. And it's in the middle of a real breakout on Pandora right now. It's over half a million streams total over a couple months. It went from about 64,000 streams in March to six figures a month for three straight months after that, April, May, and June. Pandora's own dashboard has the band kit currently sitting at around 70,000 monthly listeners. So I'm not telling you this to say that I've cracked the code. I know the magic. I'm not promising the same thing happens to your music. Because some of what drove this, I'm still figuring out myself. Once this picture settles, I'm going to do a full, honest breakdown of what actually happened here, what's repeatable, and what was the probability, just the algorithm finding it. For now, just know that when I talk about mastering, about finishing, about directing instead of gambling, it's coming from someone doing this every week on my own catalog, too. Not someone who, you know, stopped after doing a little bit of research. All right, let's get into this week's recap. So this is a new format that we've started rolling out, uh, where every Friday for the podcast, from here on out, the show will recap whatever went up on the blog Monday and Wednesday of this week. So if you if you read those posts, this is a slightly deeper cut. If you didn't, you're getting kind of the whole thing anyway. So this week, one post was about the voice in your head that says your AI music doesn't count. And one about the single decision that truly separates a catalog that sells from a folder of forgettable clips. Alright, let's get into it. You're not cheating. Let's name the feeling. You spent decades earning your musicianship the hard way. Calluses, theory, vantors, four-hour load-ins for 40-minute sets. And now you type a paragraph into a website, a finished song comes back, and there's a voice under the amazement that says, this doesn't count. I hear from a lot of you that this voice is the whole reason you haven't told anyone about your AI music yet. Sit with this instead. This argument is over a century old, and it has never once aged well. The player piano was going to kill musicianship. Recording was canned music, and union musicians fought it for decades. We all know synthesizers weren't real instruments. Come on, drum machines, they're going to put drummers out of work. Sampling, sampling, nothing but a theft. Auto-tune, hmm. Auto-tune ended singing. Every generation's cheating becomes the next generation's instrument, without exception. There's a real concern buried in there. And I'm not going to pretend it isn't there. Type sad rock song, hit a button, upload whatever falls out, and you haven't done much. And it, you know, honestly, it sounds like it. But that's not what most of you listening to this are doing. You're writing your own lyrics, you're making the arrangement calls, you're fixing the mix. There's a word for someone who makes every decision on a record without necessarily playing every part. That word is producer. Nobody ever accused Quincy Jones of cheating. And the part that should actually excite you, for most of us, this isn't a shortcut in the music. It's a door back into it. The songs that sat in your head since 1998, because finishing them costs too much time and money and other people's schedules, that cost just collapsed. You paid the dues already. This is exactly what they were for. Alright, this week's second post gets at the mechanism behind all of that. And it comes down to two people using the exact same tool on the exact same day. Two people sit down at Suno on the same afternoon. Same subscription, same model, same credits. One ends the day with a track worth releasing. The other ends up with 11 half-listenable clips and a vague sense of disappointment. That gap isn't talent and it isn't luck. It's one decision made before either of them typed in a single word. The vending machine operator types a vibe, hits create, and waits to see what falls out of the slot. Keep it if you like it, pull the lever again if you don't. The uncomfortable part for that person, the critics are basically right. Whatever came out, the machine made it. That's how you get a folder of anonymous tracks that sound like everybody else's. Because they are just like everybody else's. The director does something different, and it starts before the machine ever runs. Decide who's making this track first. The artist's identity writes the prompt, not the other way around. Generate a batch, and instead of shopping for a finished song, listen for the one take with just something in it you wouldn't have written on purpose. That's the golden seed. Then run it through one question. Strip the production off. Is the song still interesting? Keep only what survives that. The polish hides weak writing. Strip it down, and a lot of lines that sounded great inside the generation don't hold up. I took this to an extreme recently, building an industrial metal track about as far from Suno's comfort zone as you can get. The model scored maybe three and a half out of ten on actually following the direction, added vocals to an instrumental brief twice. A machine operator gives up right there. A director takes a different question. Is there a song hiding in this generation worth rebuilding by hand? I locked the structure, pulled the stems, and rebuilt every part myself. The generation was the sketch. You already know which one you are. The operator isn't listening to a podcast about the psychology of directing an AI. You are. Same choice both times. Bring the vision or wait to see what falls out of the machine. Full posts are linked in the show notes. In the complete system behind both of them, all seven books, the research, fader, is the Red Lab Library. $97, one payment. One more thing. I'm putting together a small founding cohort called the AI Music Release Lab, where I personally review your work every week and take one song from idea to finished and released in a month. Check the show notes for where things stand with it. Stop gambling. Start directing. And thank you for listening.