Testing Suno: Nhii on AI's Collision Course with the Music Industry
This is not an interview that happened overnight. It all started months ago, a casual dinner conversation between good friends: myself, Nhii–A prolific German-born, Brooklyn-based DJ, producer, sound and mastering engineer, and as some describe him, an institution of sound–and our spouses, when we went to Brooklyn's biggest food hall, which is a very American way to hang out, if you ask me.
Long story short, and a couple months later, we finally got to the official part of the interview: the Q&A, which you are about to read.
You have shared with us previously your take on this subject at UNFILTERED talks but to help us refresh memory, would you mind starting us off by sharing your experience of being a Beta tester for Suno? What attracted you to the idea of knowing how it works?
When I was approached by an employee of Suno, the platform was at Version 3.5, probably around late 2024. After diving into the background of Suno and learning the company had raised billions in funding, I was really not understanding how it could ever generate this amount of money. So I needed to look under the hood. The employee transferred me a ridiculous amount of tokens and I could start prompting. Shortly after Suno 4.5 dropped and I was experimenting my way through the platform.
It was really hard to get it to do specific musical tasks. To this day, it can't generate a song with specific musical instructions like: E harmonic minor scale, BPM, time measurements, etc. That was quite frustrating. My conclusion is that it's evident the platform is very much trained on and tailored towards the commercial market. You can generate a radio pop song à la Harry Styles, Ed Sheeran or Keinemusik within 20 seconds-you only have to type in: "Make a happy radio song with lyrics about love, make sure everyone will like it." And there you go. Final product, without intention and purpose, no artistry involved, no skill involved. The result is 80,000 uploaded AI songs with extremely similar sounds and arrangements. This leads to a whole community of Suno users that only have one thing in mind: scamming platforms like Spotify and Co for royalties. Imagine generating 100 songs a day and uploading them to these platforms to make $50-$1,000 per month on royalties. That's enough incentive to justify this action for many people in third-world countries, and I don't blame them.
I've been observing the Suno Subreddit for a long time now, and the mindset of these people is really incredible. Under the umbrella of self-expression, everything is allowed and the fact that Suno was trained on stolen data gets ignored-after all, everything is a copy, they argue.
I don't think self-expression can work with one click. Part of self-expression is to sacrifice time and labor and to be willing to learn a skill, then use that skill to turn emotions into something creative. And this is where we meet a space that really requires deeper thinking-a place where those Suno users stop and don't go further in their arguments.
So many psychologists and philosophers have been talking about this in extreme detail and the information is all there. Creating without meaning leaves you as a human being empty. Psychologists call this Self-Efficacy. When you invest time mastering a craft, the resulting expression carries psychological weight because you developed the capacity to produce it. It becomes proof of who you are. A one-click output can't serve that function because there's no bridge between inner state and outer result that you built. Mihaly Csikszentmihalyi called this core idea "flow"-a place where skill meets challenge and you lose yourself in the act of creation. His research showed that the psychological reward of creative work comes precisely from the struggle, the learning curve, the friction. Remove that, and you hollow out the experience for you and the listener.
Even Sartre and Camus both argued that meaning comes from committed action in the face of difficulty. If creation costs you nothing, it signifies nothing. The philosopher Albert Borgmann-who died in 2023-had a useful distinction here. He separated focal practices from device paradigms. A focal practice engages your body, demands attention, and connects you to a tradition of skill. A device paradigm reduces everything to input/output and hides the machinery. His argument was that device paradigms deliver convenience but erode meaning. Suno is the perfect example.
This is all possible because of the lawlessness of the AI world. Governments struggle to see where the line needs to be drawn. However, it would be so easy to draw it by focusing on one simple question: Does this technology empower the artist or does it replace the artist?
Are you using any AI tools in your current productions? And if so, which ones? And, ethically speaking, what do you think could be an acceptable way of using these tools, if there is any such one?
I use an EQ called Gullfoss with great joy, and it runs an algorithm that is modeled to please the human ear. To simplify: the human ear has some frequencies it really likes or kind of doesn't. It's no secret to audio engineers. The Fletcher-Munson curves show that. The EQ empowers me to get rid of detailed EQing I've done for over 20 years on specific channels inside a production. It has a simple interface, but you can still mess up the signal with too much of it, so on a basic level, it still requires a functioning ear to dial it in.
Many producers use sample libraries for their productions, which is in my opinion already sad, as you start hearing the same vocals and drum loops on different tracks-but at least you still have to make the choice to select a sample and mix the final song in a decent way. One tool I just found is called Illugen by Waves Audio. It has been trained by ethical standards with material Waves Audio owned. If you want to use AI for creating stems or single tracks inside your production, this is the way forward.
Why do you think it is important right now for us to care more about Nature?
I'm unfortunately seeing a trend in our industry where compassion is on the way back. It's cool now to fly private jets, eat meat, and play songs with disrespectful messages and gunshot samples. It seems like common sense has left the room. Everybody is so busy hustling and trying to make it work if you're not in the top 1%, but we seem to forget that if nature is not healthy, we will not be. And I believe this trend is happening because our leaders don't care and lack compassion on all levels. In my opinion, a big reason why we're fighting each other like there's no tomorrow is the way we treat animals. If you have no compassion for defenseless animals-pigs, turkeys, etc.-you can always downgrade the other side of humanity to animals and treat them the same way.
Albert Einstein famously said: "Nothing will benefit human health and increase the chances for survival of life on Earth as much as the evolution to a vegetarian diet." And now with another weight on nature's shoulders from AI, I think we do have to consider living a more sustainable life. Being vegan is the easiest way to do so.
In your personal experience, would you say AI has helped you with your workflow? And if yes, would you mind sharing a few examples of its use that you find to be helpful?
I use Claude for staying organized and planning my day. With running my label, Patreon channel, gigs and travel-live sets, vinyl sets, DJ sets-building speakers, learning Japanese, working on slapback.io and more projects, it really helped me to dial in a strong routine with some free time. It's basically a personal assistant. More specifically, I use it to create designs for building wooden speakers, run some audio wave math, and as a Japanese tutor.
What are some common misconceptions or opposing factors to all the good that comes with such tools? Are we sometimes over-relying on them?
Gullfoss, an EQ for any DAW, is a perfect example to think through here. It's a brilliant tool. But if you reach for it on every mix, you're outsourcing the development of your ear-the ability to hear a problematic frequency, identify it, and make a judgment about how to address it. That's a skill built through thousands of hours of getting it wrong first. Every time Gullfoss handles that for you, you get a better-sounding result in the moment, but you're skipping practice. Over months and years, that compounds.
So you can make the case here that without your 10,000 hours of audio engineering, it's hard to get the best out of Gullfoss, because you might not know how much processing the signal needs.
There's solid research on this from aviation psychology. They call it skill atrophy through automation bias. When pilots got autopilot and automated landing systems-helpful tools-their manual flying skills measurably degraded over time. Not because the tools were bad, but because the pilots stopped practicing the underlying skill. The FAA eventually had to mandate manual flying hours to counteract it.
When it comes to learning, would you say it is still important to learn tools in more analog ways, and later be introduced to tools that can simplify our work? Or is it perhaps an extra, almost outdated step you think?
I respect artists that mastered their craft-let it be pottery, painting, music, or any other art form. If I listen to a DJ, I want the DJ to be an absolute pro. Because I know, only if you mastered your craft, can you freely express yourself; otherwise, you are limited by technology and lack of expertise. I'm not interested in a party DJ that just hammers one hit after another. That, in my opinion, AI can do all day long, and probably even better.
A DJ needs to listen to music first and foremost. Nowadays we have content creators producing songs without actually having knowledge in music and music history. What can this person teach me when I listen to a set? At this point, you're even gaslighting music fans by blowing up your numbers and promoters fall for it. The result is simpler: the listener and ticket buyer doesn't get the best possible music but more of the same, because that's what's being copied-there's no inspiration from deep listening sessions and actual digging.
We don't need AI to create music. Everything we need is out there-go to your record shop and find an obscure sample from a folk track, sample it, twist it, have fun with it, and combine it with your skills. There's endless inspiration for new music right away. You don't need to use Suno for inspiration; it's a lazy excuse I hear often.
So personally, I think that the foundation you lay by getting into the nitty-gritty of your hobby strengthens your thinking process, confidence, and ultimately makes you a better human by failing and growing.
If you were able to change one thing about AI, what would that be?
AI is here to stay, if we like it or not. So I would definitely change the legislation around it yesterday and put some limits on what's actually allowed, especially in the creative area. Nobody asked for Suno or image creation. We have good people doing incredible things in creative fields. And meanwhile, there is no one-click solution for doing my taxes.
What are some of the human-robot collaborations that we should be expecting from you in the near future?
Yes, I am working on speaker builds where I use AI to calculate the inside proportions and explore some odd shapes. This will take some time, but I'm confident to have a first prototype within a year from now.
Another one would be Slapback, an idea started by Atish. It's a platform that acts as a mixing engineer and gives feedback on your mix-downs and unfinished tracks. We are using AI to analyze the track and then combine it with a self-built LLM that contains the proper academic audio engineering knowledge.
And from there we took it to Waverly Street.
In downtown Brooklyn's Waverly Street, there's a whole building full of music studios. One of these studios remembers the creation of "Feathers," "Secret Drums of Bwiti," and more recently, "The Color of Sound," all of which belong to Nhii.
Last week, when local deli storefronts, alongside the entire city, were bathed in gold, I took a short walk from my Downtown apartment to his studio. There I had a chance to visit Nhii in his usual habitat.
When you arrive, industrial-looking stairs deceptively guide you up, but only a few steps after you enter the building you find yourself heading downstairs towards the basement, an environment where the underground scene often finds itself the most comfortable.
Barely five minutes passed as we rolled the cameras and started recording what was initially thought to be no more than a short conversation, with a couple of topics discussed in the interview prior to recording the follow-up. The entire raw recording ended up being over an hour; for your convenience, we have trimmed it down to a more digestible 38 minutes and 54 seconds.
Full recording will appear on this page tomorrow, Friday, August 20th.
Signup to be notified
U.S. Copyright Office proposed the registration fees’≈43%
increase. In return, they received 164 comments from some of the nation's most influential creative and news organizations; the response suggests widespread concern about the accessibility of copyright protection.