How AI Will Fail Like The Music Industry — Transcript

The reason I'm talking to you today is really because of a piece of hardware and software that came out in 1999.

I've mentioned this before, but it's called the Digi 001 by Digidesign.

It's one of the first digital audio workstations and interfaces made for a computer, which happens to be this G3 right here, which I got back in the year 2000.

As you can see here, it's a little unit that is an interface, then you have a cable that goes to a PCI card, which you can see here, and that went into the back of this computer, into the PCI slot.

Once it's connected to the computer, you can just plug a microphone or instrument cable into the back of the unit and you can start recording digitally.

You see, recording studios back then cost a lot of money to build; you had to have a building, a big building, because you'd usually have three rooms, you'd have a recording console in each room.

An SSL console like this one cost about $750,000.Many times you would have an Eve console for tracking in another room, that's $350,000 to $500,000, and then usually in the third room you'd have a mixing console.

These things, they would take out loans on, and they would be incredibly expensive.

Back then it cost about $2,000 a day, and when the studios bought the Pro Tools gear back in 1999, the bigger version of this Digi 001, which was called a Mix Plus system.

Those things, included the computers, cost about $30,000, so when they got those and they stopped using tape, they started charging people like myself an extra fee for Pro Tools, which would be about $500 a day, not only that, but they charged you another $200 on top of it for a Pro Tools certified engineer, even though I knew how to run Pro Tools myself because I was using it at home.

Why is this important, because the recording studios of that time are the data centers of today, incredibly expensive to build and maintain, the stuff inside is incredibly expensive, and they're spending hundreds of billions of dollars on these things.

The disruption that happened in the music industry 25 years ago, first because of Napster, which killed the profit model, and then because computers got faster and faster, and it made it easier and easier to run these Pro Tools systems in your own home studios, which is the way that is done pretty much in the music business by everyone in every style of music now.

Recording studios are few and far between, the same thing has happened in film and TV, I'm recording this video right now with a camera that I got at Best Buy, a Sony camera that pretty much all YouTubers use, and I'm going to edit it on my computer that's sitting right here, that's running the newer version of Pro Tools, that same computer is so fast, and computers have gotten so fast that they can run AI models on them.

Let me show you how.

Okay, so how do I get one of these LLMs or these AI programs on my computer, I'm going to go to this thing called LM Studio, LM Studio is a desktop application that provides a user-friendly graphic interface for searching, downloading, and running large language models locally on your own hardware.

This is going to be run on your computer that's not connected to the internet.

You can put all your personal data on and no one else can see it.

Click download for Mac.

I'm running a Mac.

I get it.

Okay, now I have to get an AI program or a large language model, so I'm going to go to this website called Hugging Face that has 2.7 million different free LLMs on it of all different sizes.

I'm going to download one that's decent size, this is Qwen 3.5, this is one that has 36 billion parameters, so it's a big one, but it's not that big, I can run it on my computer that I bought at the Apple Store.

Okay, I download that.

Boom.

Okay, I have it on the computer, now I'm opening it, open LM Studio, here we go, I'm going to put Qwen 3.5 is on here.

Okay.

Then I'm going to do new chat.

Now let me put in a prompt that somebody probably that's watching this would put in, so the prompt is this, I have these a list of these ingredients: eggs, milk, tomatoes, potatoes, flour, chicken, give me a recipe out of these that I can cook tonight, you don't have to use all of these, here we go, here's a hearty and comforting recipe that uses almost all of the ingredients to create a one-pan dinner, it combines the crispiness of potatoes with a creamy tomato chicken sauce, topped with eggs for richness.

Let's do another one.

This is something somebody might use, rewrite a rough draft into a polite, professional email to my boss asking for next Friday off, hey, I need to take Friday off because I have a dentist appointment and it's far away, can I go?

Thanks.

Let's see what it says.

Okay, so here it is, option number one, standard professional, recommended, hi boss, or hi Rick, let's say it's one of my guys, I'm writing to request next Friday off for a dentist appointment, since the location is outside of town, I will need the full day to accommodate travel time, I plan to complete my current priorities before then and ensure that all projects are up to date, please let me know if this works for you, best regards, Tom.

Seriously, Tom, can't you get a dentist in town?

Okay, so this is something that people will use AI for, this is actually a common thing.

Okay, let me do one more.

Plan a one-day trip to London on a budget, total of $100, list three top free attractions and give me an estimated price for lunch.

London is an expensive city, so 100 bucks might cost you 100 bucks to do stuff.

Now, remember, this is just running on my computer.

I'm not connected to the internet with this, this is doing it from this LLM that's downloaded to my computer offline.

My computer is not connected to the internet.

Okay, here we go, here's a one-day London itinerary designed to fit comfortably within a $100 budget, approximately 75 to 80 pounds, this plan focuses on walking, public transportation efficiency, and iconic free experiences, uh, morning, British Museum and Covent Garden, number two, lunch here in Covent Garden Market or Soho, afternoon, National Gallery and Trafalgar Square, evening, St. James Park and Hyde Park, it shows lunch estimate here for $20 USD or 16 pounds, and then it's got the budget breakdown for $100, transport $15, lunch $20, snacks and water $10, souvenirs $35, attractions free, total $80, money-saving tips for London, so on and so forth.

This is something that you use, now the great part about this is that people don't know you're going to London, ChatGPT does not know your whereabouts, you don't want these companies knowing everything about you, why are you giving them your data about your personal life and things like that?

Okay, so why is this important?

What happened to all these recording studios after there was no profit model that the labels relied on to make records with, they all went out of business, they couldn't afford to make the payments on those $750,000 SSL consoles or their Neve consoles or anything, they couldn't afford the electric bills because people were recording at their homes.

And this is what I think is going to happen with these AI companies, the data centers, they're going to be sitting there unused, many of them will not be built when people start using AI locally, meaning on their computer, and the same thing that happened to the music business and recording is going to happen to these AI companies.

If a 64-year-old guy like me can figure this out last night and show you today, how hard can this stuff be, people are using these things, Suno and UDO and everything for music creation, people are are taking their any of their business things, their financial information, they're feeding it into ChatGPT, people are taking their any of their business things, their financial information, they're feeding it into these things, and they're training off your data that you own yourself, your own personal information, your own intellectual properties, and why is that, because you and me are the product.

So who are the winners and losers of this, well, the companies that make hardware, like Apple, Google, Nvidia, any company that makes computers, computer chips, things like that, those are the ones that are going to be the winners, unfortunately, the ones that are going to be the losers, in my opinion, are the companies that make these AI programs, for most people's needs, individuals and businesses, even good-sized businesses, you can do these things on your own computers offline.

Love to know your thoughts, hit subscribe, leave a comment, thanks for watching.