✘ This newsletter is optimistic

And: A return to the spirit of punk; MMM can boost digital audio spend; Four kinds of night; The year of the music licensing legal wars; Three AI use cases

I spent some time this week talking about AI in academia. In the past year, I’ve myself noticed an increase in students handing in work clearly generated with ChatGPT:

• the em-dash

• too many adjectives

• a lot of ‘it’s not this, it’s that’ sentences

• hallucinated references

Is this just lazy? Why aren’t these students showing the level of critical thinking we ask of them throughout their studies? Are we, perhaps, not teaching them the right thing here? Are we, perhaps, not assessing them on the right metrics?

There’s a danger of paralysis here. I say students learn the kind of critical thinking needed to assess the output of an LLM. You say that by using LLMs they don’t develop these critical skills anymore. This is exactly where we get to the question of the human element in and through AI.

Music to my ears

This conversation about academia resonates with talk in music about AI. Music is in a place where the systemic failures rule the whole ecosystem. KKR, or private equity more generally, takes a hold of the festival landscape. People complain and artists pull out of festivals - nothing changes. These actions don’t impact the systemic disconnects. So when Deerhoof, or even Taylor Swift, withdraws their music from Spotify there is applause. But it doesn’t change anything. The underlying structures require work to actually affect change. There’s a few companies who dominate music, just like there’s a few companies who dominate the Internet. A lot of this has the kind of funding structures that ask for high returns and short-term gains.

Music specifically doesn’t sit well with this kind of double-digit growth desire. Margins are usually thin and the question of ‘who pays?’ usually ends up with the answer of either the artist or the fan. They can only be squeezed so much. This is why so much comes down to scale. Just look at distro.

This leaves open questions around how the various music industries or ecosystems can work through and adapt to new technologies. This isn’t just about AI, but also about mixed reality, avatars, wearables, and much more. The open questions, then, require new funding models as answers.

• Exit to community

• Releasing work together

• Prefunding and patronage

• Valuing return on culture

• Running experiments, lab style, to create new markets

This is just a number of ideas, there’s many more. I’ll write more about some things that I’m personally working on in the next months.

Find your creativity

A lot of this is about storytelling. Can we weave a narrative where we take the power away from the major companies building all this tech? We can even use their tools against them and ask the friendly LLM of your choice what to do to make this happen. And it is easier to build things. And it is cheaper to build things. It isn’t, however, easier or cheaper to be creative or have an idea. Sari Azout, who’s building the brilliant tool Sublime, has been putting it like this recently:

“Browsing past dozens of landing pages of new AI tools promising to automate, optimize, accelerate, and 10x what we already do, I can't shake the feeling that we're missing the point. That the real challenge is to rethink what we optimize for, and loosen our grip on the idea that the things we value can be neatly quantified.”

It’s this rethinking that’s exciting me, right now. Creativity demands of us to explore this.

Anyone can prompt anything into being rights now. The latest versions of LLMs will offer to build the app you didn’t even know you needed. But the thing, and this is what Sari is also after, that really matters is knowing what you want. If you don’t know what you want, you won’t get what you need. Now, there might be some cool serendipity happening that means you ends up with something useful. However, to really use the tools at hand to good effect, we should have our principles in place.

We’ve got the numbers

Indie’s have been growing in terms of revenue share and listener share on streaming. This is a good thing. These indie’s have the creativity, now they need the knowledge and funding to experiment and create new models. That streaming model is facing a serious slop issue. If this means we’re pulling into a new territory of music consumption the laidback listeners will have the most power again. It’s in these situations that listeners might not care if 80% of the music served to them is bang average and created by AI. If the other 20% gets their attention - that’s a good hour spent listening while doing something else (driving, cleaning, working out, etc.).

We see this happening, as MusicBusinessWorldwide has explained through their recent coverage of fake artists and the AI music problem on streaming services. MBW reporter Daniel Tencer asks a good question:

“Does writing the lyrics and having AI generate everything else – as seems to be the case with Aventhis – count as “sufficient” human input? That’s not entirely clear.”

And here’s the issue: we currently don’t have a definitive answer. Legal cases are ongoing, precedent is being set and contested. Policy is running behind. All the while, the tech keeps developing and this gives everyone a great opportunity to try and work magic on their creativity. At the same time, it also gives space to bad actors. Fraud has been a longstanding issue, but it’s being compounded by generative AI tools. Tackling this is a system-wide issue for music in general, not just for indies. Here, the majors play a big role - they stand to lose a lot, too. All of this preventitive or punative work will eventually lead to new legislation. In the meantime, there’s as much opportunity to make this work for good as it is to make it work for evil. That’s not to say that you should start to flood the zone with music. Instead, it’s about learning to build your own models and working creatively with them. This will help you untie yourself from the web of systems errors that hamper the industry.

Does this sound optimistic?

It’s a strange time. The line between “we’ve been here before, don’t worry” and existential dread is quite thin right now. Sure, we’re going through a tech cycle like any other. And, yes, it’s also different and things will change. If we hone in on that change, it might just help us get to those shifting business model opportunities. This means critically and creatively engaging with the tech at hand. And that means being able to ask the right questions of it. It’s a time of rethinking.

---

LINKS

🥀 Why we might be witnessing a return to the spirit of Punk (Hanna Chalmers)

“As mainstream culture is imploding in on itself and can feel like a stitch up, new ways of being are bubbling up at the edges. It's more off-line - or at the very least in private online spaces. It's smaller, DIY in sensibility, and it's a safer space for its participants. Fundamentally it's unsurveilled.”

✘ Love this take on the impact of our fully overindexed and mass-surveillance and very online world. Punk is definitely where it’s at.

🎚️ SiriusXM says MMM could boost digital audio spend – as long as brands get the right data (Anthony Vargas)

“But, if digital audio is going to grow its share of brand budgets, Collins said, then platforms need to pass advertisers data that can demonstrate the channel’s incremental reach and aid attribution for performance campaigns. But, she said, there’s very little standardization in terms of what audio ad impression data brands have access to.”

✘ Ah, the wonderful world of marketing mix modelling. There is a lot of ad dollars floating around in the world. Audio is definitely not deep enough in the overall pie, with video a very important player. But, the pie is so big that any growth for audio will be a substantial piece. Ads are always about scale, so individual or indie actors aren’t very well-placed. Good time for a coop to emerge!

🌃 Four kinds of night (Mikayla Emerson)

“Because you’ve been flown through the branches on a white lifestring and you’re hanging over the powerlines missing the ground. Because your knees buckle in the summer. Blisters and bruises and knots. Because nobody walks you back anymore. Nobody leaves the light on in the basement. Because 4AM means 7AM. Because this is the end of the track. Because the Q takes you to the ocean if you let it. If you miss it. If you allow yourself to be missed. I’ll remember you just like this: flittering through the air like fumes, like wind, like light. I don't have to promise. It always turns out this way.”

✘ Something a little bit different, but I really enjoyed this meditation on remembering nights out. Take a step back, close your eyes, and remember your own nights out from five years ago (or longer ago). You’ll see what you remember and ask yourself what’s actually true.

🪖 The year of the music licensing legal wars (Sarah Jeong)

“There was an implicit assumption on all sides that the iPod was legal, that the iPod was legitimate, that the iPod was worth protecting. The justices fretted that letting the file-sharing services win would destroy the music industry; but on the other hand, if they let the MPAA and RIAA win, it would destroy the iPod.”

✘ I like to say that while the industry won against p2p file sharers, they should have won this case against the mp3 players. There will always be demand if the supply is asking for it. This is a great reminder of what happened roughly 20 years ago and how that will impact how we deal with genAI this time round.

⚙️ The 3 AI use cases: Gods, interns, and cogs (Drew Breunig)

“Today, Interns are delivering the lion’s share of the realized value from AI. Engineering copilots alone are delivering massive returns, increasing the output and capabilities of expert programmers. I’ve heard of a few large companies who have slowed down hiring due to their effects and have witnessed many tiny teams ship products years beyond what they could have delivered even 3 years ago, thanks to Github Copilot and similar tools.”

✘ This is basically still where we’re at, even though conversation has moved more and more towards agentic AI, most of these agents are still intended as interns.

---

MUSIC

I’ve been listening to Nick León’s debut record A Tropical Entropy daily since its release last Friday. It captures a vibe somewhere in between breakbeat and latin grooves. It feels like the music comes through in a similar way as looking at a sunset through your eyelashes after an intense day. It’s both relaxing and at the same takes a toll on the listener.