HOW TO SURVIVE TO AI AS A MUSIC ENTREPRENEUR

HOW TO SURVIVE THE AI ARMAGEDDON

Well, I have no idea. Sorry. I swear that I would wholeheartedly share my robot-proof plan if I had one, but this is not a good excuse to holding my punches in the AI-bashing arena. Also, I’m an independent sample library developer and sound designer, which is another way to say I’m a small artisan in his small boutique making a living out of my craft. Hence, if this whole algorithmic armageddon has start to start somewhere, you can bet your money that’ll be my @$$.

So why on earth am I wasting my time with making another library? (a glimpse here).

Shouldn’t I just pack it and quietly hand my resume again to my old employer?Oh, I’m sure they’ll really love to hear how I’ve spent these last five years, unless there’s a robot filling that same position right now btw.

In case you’re wondering: yes, I’m worried too. Like some of you, few years ago I jumped from a somewhat normal job (more about this later) to the music tech entrepreneurial highroad. I’m immensely proud and happy with the choice. And it’s because of how happy I’m with my life that lately I’ve been spending a considerable amount of time into thinking about the future. How my career can stand up to such disruptive technological advancements? How are my libraries/plug-ins going to be competitive against an immensely powerful digital demiurge who virtually blows me out of the water at pretty much anything?

You know, worrying never solved any problem by itself. But worrying + analytical thinking has a good scoring of doing so. And if you add what we creatives are good at, which is lateral thinking, perhaps we might end up being onto something. Said this, let me share with you a few important things I came up with my AI-related late night rumination.

Spoiler: the picture is not as bleak as those bazillions of clickbait-y tech websites (probably with AI generate articles) want you to believe.

Prepare for a serious perspective refresh and a much needed morale boost (oh, and if you want a jump-out-of-the-roof morale boost, go here).

AI IN THE PERSPECTIVE OF A SAMPLE LIBRARY DEVELOPER.

If there’s a straightforward way to explain to a composer what is AI, there it is: AI is sampling. You might have heard about the complex neural-based machine learning algorithms, but those would be worth nothing if they weren’t fed gargantuan amounts of data. Exactly as Kontakt would be virtually useless without audio content imported and mapped, AI simply relies on input data. No samples, no music. No data, no AI.

There’s a profound difference though: once that those algorithm ingest said immense amount of data, they’re able to synthesize things by themselves.

Hence, if there’s a field of music tech where the implementation of AI is inevitable, in my humble opinion that’s exactly sampling.

Traditional music sampling is based on audio that has been recorded and organized in a way that the final instrument offers a convincing realism (if that’s the goal). Conversely, machine learning algorithms need data that have been meticulously pre-organized, categorized and labeled.

Messy audio sample-sets with poor categorization make unusable libraries.Random, incoherent data fed to neural networks make an inefficient and not-so-smart artificial intelligence.

Bear in mind, as this is a central concept: AI needs categories and standards. Categories and standards. It’s for this reason that I believe that “standard” libraries like orchestral strings libraries, solo violin libraries etc will sooner or later fall into obsolescence. Probably not for you and me, but young composers in the year 2050 will start their journey with a very different kind of instruments, possibly AI powered instruments. No more GB and GB of samples, but perhaps instruments capable of synthesizing real time realistic orchestral sounds and articulations, based on a core algorithm that has been trained with loads of orchestral string samples as input data.

Would be great isn’t it? If there was such an instrument capable of generating fresh realistic sounds/articulations on the spot, rather than relying on “static” chunks of pre-recorded audio, I don’t see who wouldn’t benefit from such technology.

AI IN MUSIC TECH – WHERE ARE WE NOW?

But are where there? Well, not exactly as far as I know. At least, not yet. Yes, I’ve seen some companies advertising their AI-powered drum machines and synths. However, if you ask me, I really don’t see who has to gain from a software capable of generating thousands of 808 kicks with a click, when with the same mouse-click I can get 10x more of the same kind of samples with a quick Google research. Or I might learn how to make said sounds in pretty much any already available soft synths, with just a few clicks more. By no means I’m meaning to bash anyone, but I don’t think you can win the abundance game in a context of already present overabundance.

What I’m trying to tell is: we don’t have an AI-powered Kontakt/Omnisphere. Or an AI-powered Massive/Serum etc. Or an AI-powered orchestral strings software with appreciable advantages over the traditional ones. In other words, we don’t have a new game-changing industry standard instrument as the aforementioned ones. Again, not yet. We still make music the same way, I don’t see the creative tech revolution, I don’t see nothing close to what happened in music when the first analog synths were commercialized in the 60s, or when the first soft samplers popped up in the market 30 years ago.

On the top of that, the only examples of AI application in music I hear are by random dudes deepfaking famous rappers or singers and make a sensation for one week before something else distracts us once again, and again, and again.

My first question is , and I want you to imagine me, kneeling in a stormy desert and yelling at an impenetrable sky: WHERE IS THE CREATIVITY? WHERE IS ART? Where are the AI generated “sounds we’ve not heard yet” that a lot of folks were raving about on Reddit last year? Where is this new AI-powered synthesis technique we couldn’t even dream of until 12 months ago? Don’t know you, but I’d dare to say: crickets.

And do you know why? Because AI is based on categories and standards. It doesn’t matter if it’s audio, music, images or videos,as long as there’s sufficient material that falls into a widely recognizable and categorizable standard, AI can cannibalize it. It can cannibalize a set of samples or a whole musical style and churn out thousands of sounds or entire songs in that style. Everything outside of those strict categories either doesn’t offer sufficient available data or simply doesn’t raise a financial interest for those who develop these technologies.

Again, it’s an abundance game. Which leads to the part that I suspect interests you most.

AI IN THE PERSPECTIVE OF A MUSIC MAKER

I happen from time to time to stumble into a relevant voice in the music business expressing his or her view about the future of the industry with AI as a new term in the equation. 

Such opinions on the matter seem to be quite polarized. On one hand, wealthy guys with prosperous careers generally have a quite laidback approach. They consider AI nothing more than another music tech tool. They frequently mention the first drum machines commercialized in the early 80s as a term of comparison. As we know that no drummer ever lost his job to a Lynn drum computer, let’s just chill everyone and “embrace” AI , and let’s go seeking greener creative pastures by taking advantage of it.

On the other front, some other guys decide to wear the Nostr-AI-damus hat and seem absolutely sure that every single one of us has 100% chances of  being turned into an IKEA low-priced commodity at best, or an obsolete piece of garbage at worst.

Frankly I don’t know where to stand. But I think that both positions suffer the bias of an hyper-focused view on  AI in the music industry, rather than envisioning it as a global epochal shifting agent,  which affects the way we interact with reality.  Music and music tech are in my opinion a mere consequence of said shift.

Generally, I think most artists will concur on one very specific thing: we don’t want to outsource the most precious, meaningful and significant aspects of our creative process to an algorithm just because of the “do more do faster” marketing mantra for both AI as non-AI instruments. In my personal case, not only I don’t want to outsource it, I need it to be difficult. I need to work hard to get, eventually, as close as possible to my initial vision. You heard me right, I want to suffer. Before you think I’m a sadomasochist psycho, let me explain my point and let’s for a moment adopt a thinking strategy I really like, which is absurd brain storming.

Imagine this: it’s year 2084, they did AGI (Artificial General Intelligence) and there’s hardly a single aspect of human life that is not touched by AI. AI can do literally everything on our behalf.

It can translate a language we don’t speak, so what’s the point of learning it. It can crank out hundreds of lines of code, so what’s the point of software engineering. It can summarise in 300 spoken words The Myth of Sisyphus by Albert Camus (phenomenal book btw)  and read it to us with a voice that has been custom-synthesised according to our taste. Hence, what’s the point of reading anymore. It can make financial decisions on our behalf according with our spending habits, so who cares about financial education. And of course, why learning music production if anyone and their grandma can snap their finger and have their banger in any imaginable style?

Basically , what’s the point of attempting anything that requires even the smallest effort, if there’s a digital genius that just can’t wait to serve us. In 2084, we’re on the way of outsourcing all our skills that took thousands of years to be developed and transferred from generation to generation. The AI leviathan gets bigger, and smarter and more dominant, so become the companies who develop these technologies.

If this is not absurd enough, let’s jump to year 2184 and what I see is a sort of futuristic Dark Age where the power is concentrated in the hands of few (I put a bare bones bet on the aforementioned companies) , whereas the vast majority of the population, despite having virtual access to anything, it lives in a condition of deep intellectual darkness where all those hard earned capabilities are lost. Gone, nada, ciao ciao. Creation becomes a prerogative of AI, and AI only, our evolutionary pattern took a permanent U-turn.

However, there’s something in the absurdity: yes, the Dark Ages were very, errrr, dark. Yes, the power was held by kings and common people was relegated to a condition of intellectual darkness with no chances of emancipation whatsoever. However, there was literature. There was music. There was maths. There was philosophy. There was art. In other words, there was a powerful intellectual elite. And where  said elite  had its leverage? They possessed knowledge and skills that the vast majority of the population didn’t have. 

Maybe my point is a bit clearer now: those who will stubbornly commit to nurture their skills and knowledge will have a remarkable leverage over those who will choose to outsource theirs. 

As a music maker, ask yourself: which competence/skill makes me unique and I would never ever give away? Am I available to devote my efforts into learning, into developing ideas no matter how challenging they can be? Am I bold enough to follow my creative instinct no matter the Instagram, Youtube, TikTok goddamn algorithm?

Think this: whatever the point you’re currently in life, a lot of it depends on things that you managed to learn. It wasn’t easy, it wasn’t all fun, but eventually that paid back. Read it again. We are what we learn.

And this, my comrades, is the spirit of lifelong learning. I got into it when I was a postdoc researcher in ship science at the University of Southampton, UK. It helped me to periodically assess my worth as a professional and my well-rounded-ness as a human being. 

And I’m confident it makes me more gaslighting-proof to the aforementioned mantra, according to which my worth can be only be weighted on the scale of me being a customer or not. 

A lifelong learner won’t consider AI as the absolute evil, but will possess the cultural means to use it in a virtuous way. A lifelong learner can discern between the structural but not artistically relevant tasks that can be outsourced to AI and those where AI is an assistant at the best. In the end, I strongly believe it’s a matter of attitude, devotion to one’s art and, eventually, hope in humankind.

Which leads me to the following point.

UNCOMFORTABLE YET NECESSARY EXISTENTIAL QUESTIONS

Ok, let me say that I’m afraid this might ruffle some feathers. So, huge disclaimer: I DON’T MEAN TO JUDGE ANYONE. Read my lips again : I DON’T MEAN TO JUDGE ANYONE. Really, nothing I’m going to say here has any intention of moral/virtue signaling.

We cool? Great.

Try a random music-related search on Youtube right now. Chances are you’re going to bump into a myriad of videos with titles like “How to make a (insert name of famous producer/composer)-style track in 30 seconds”. Or, “How to create a (other famous producer composer)-ish string arrangement”.

Please don’t take my word for granted, do it. Some of these videos have way more views that any of my library walkthroughs will ever get in one million years.

Please don’t mind the bold question: aren’t we already been playing at this replica game since years, AI or not? Aren’t we already piloted by the algorithms of this or that platform into said game? Aren’t we bending a little too much to what the damn algorithm wants in order to make our content (I hate that word) more visible? Aren’t we feeding the abundance game ourselves, as humans?

If , like me, you hear a remote voice in your heart feebly uttering “yes”, then that’s what we really need to talk about. We have somewhat accepted to be thrown in this abundance game that we can’t win alone, as independent composers, sound designers, creatives and the likes. Machines have always been created with the same purpose, and that purpose is , again, smashing the abundance game at minimum cost and maximum profit. Ironically, we are providing a very generous help in this endeavour.

Ok, I hear you, one gotta do what he/she gotta do. We don’t control the platform we rely so much on, the history of music makers is an history of adapting to whatever the circumstance in a given historical frame. If one needs to play that game, so be the game, fair. But I beg you to reflect for a second on this thing: by blindly playing the overabundance game, WE have created AI. Before Sam Altman, before Nvidia, before Google.
What I mean is that we, with our habits, with our daily actions, have prepared the cultural playground for such kind of technology to flourish.

Think typing. Do you know how we spend on average 3 hours everyday? There you go: typing. What is there to be surprised of that using AI mostly rely on some sort of typing input? Technology is nothing more than a snapshot of the human species at a given time.

Again, as a company sometimes I’m tempted as well to join the social platform AI-algorithm chase and make choices accordingly. Striking a balance between making stuff that I genuinely like and make it work financially is so challenging and if there’s someone who isn’t 100% immune to said mechanism, that’s me.

However, I think there’s never been a most critical moment  to interrogate ourselves about which kind of music community we want to build for our future. As utopian as this may sound, I believe we have a choice. On one side it’s where we’re now, with compulsive content (again, I can’t stand the word) posting, last minute trend-surfing and basically let the AI driven platforms decide which music we have to make, how to promote it, how to introduce it to people and perhaps attempting to replace us. On the other one, my dream is a world where technology keeps abiding to its service for humanity, by providing tools and solutions that serve us as humans and artists. Tools that leverage our creative pursuits rather than competing with them, tools that do not get in the way of being authentic and genuine with the people we want to connect with.

In other words , a human friendly technology. 

You know, AI is developed by engineers. I have a PhD in Mechanical Engineering and I’ve been working as a research fellow for a long time, I know people from the STEM world very well. I had the chance of working with some phenomenally talented guys, really cool people who are insanely passionate about what they do. However, I find that many of said guys suffer of what I call the almightiness bias. Given that STEM subjects provide the intellectual and practical tools for a very deep understanding of how things work, it’s very easy to assume the attitude that those very same tools can be used to understand everything in life.

I see many academics and researchers suffering this fallacy. The problem with the development of AI is that those teams hardly include people coming from the world of arts, music and humanities and if they do, it’s more a technical consultancy about a specific aspect. For example, some AI companies hire composers to provide custom music to be fed as data to generative algorithms.

Nevertheless, I think those teams need to  hire musicians, writers, poets, philosophers and artists to have some long and thoughtful brainstorming sessions about why the implementation of AI in music seriously risks to go in a wrong direction. 

As I said, I happened to navigate both the two worlds and I think that people need to talk to each other and artists need to get their voice heard.

CONCLUSIONS

I’m re-reading what I’ve written so far and I realise it’s a little crazy. You know what? Fair.

I hope my words can spark some reflection about being an artist and being human. There’s a lot to be proud of belonging to both categories, so if I managed to contribute to a little yet collective chin raising, then I’m done for today.

I’ll leave you by distorting a famous quote by Pablo Picasso.

When inspiration will come, it will find me painting” (P.Picasso)

When AI will come, it will find me making dope virtual instruments (and yelling at Kontakt at the same time)“(G.Caiazzo).

See ya,

Giuseppe

Giuseppe Caiazzo

Sound designer, producer, programmer, nerd. Founder of Silence+Other Sounds.