Some quick takes on AI-based music composition startups

I just responded to two questions about AI-composition startups from a student and thought I’d share them here as well. This is a placeholder for deeper thoughts.

The questions:

  1. What are your thoughts on the uses of Amper music specifically as an AI generative software? A user can create their own personal track by determining the durations, instrumentation, descriptive moods and function of music. The user must pay a one-off fee, and in return acquires a global perpetual royalty free license (although all copyrights are reserved by Amper).  (https://www.ampermusic.com)  
  1. (https://www.theverge.com/2019/7/23/20707371/tiktok-jukedeck-ai-music-startup-acquisition) The tech start-up company Jukedeck has been around since 2012 and provides AI generated music to users in a similar way to Amper. However, in 2019, the software was bought by social media platform Tiktok (so that the software can generate its own music-more profit to be made, as it would eliminate the need to pay royalties to record label.) I am curious to hear your views about the ethical tensions that this situation may raise for the music industry?

Thanks for your interest [student emailer].  You should really talk with Eric Drott at Texas who is studying these things in depth.


Honestly, my main thought on AI-assisted music composition is: who wants it, and to what end?  We live in a world where each year more new music made by people is released than the last, much of it available for free or cheap online. The only people I can imagine who really want an AI composition service are people who are currently paying musicians who don’t want to pay them.   Amper is working in the “disruption” model where they are trying to take money from licensing firms who sell stock music.  So they are simply trying to elbow their way into an existing industry, Uber or AirBNB style.  JukeDeck seems similar.
“Ethical” implies that the business model could be tweaked to be fair.  It’s not; it’s political in that it’s designed to take something from one group to the unfair advantage of another.  It’s scorched-earth capitalism. The business model seems fundamentally at odds with any desire to support music or musicians, instead diverting money that could have gone into supporting the arts into the pockets of investors or tech firms.  Most institutional efforts at AI ethics actually serve the industry’s desire to avoid regulation.


That said, stock music firms are also not always on the up-and-up with musicians, but at least there is the possibility that they would in some way provide institutional support to music.  AI-based composition firms offer no such possibility.


Artists experimenting with AI-based or assisted composition or collaboration is a whole other thing, and something I wholeheartedly endorse.

EDIT: the very idea of a “music composition startup” is head-explodingly weird.

2020: “Hit record”

My New Year’s resolution is “hit record.”

Very simply: it means that I will record a lot of my musical improvisations over the course of the year. I’m aiming for twice a week give or take, but I’m not going to be precious about it. I have ideas for what might happen after that, but that part will take its course. Some of might become ideas for Volte songs. Some of it might turn into new electronic work. Some of it might feed into another project. Recording might or might not make me more focused in my improvisational practice. We’ll see. The point is to make it simple: hit record, see what happens.

Perhaps you’ve heard about Prince’s vault. He made a gazillion recordings and only released a relative few because that’s how the music business worked: you risked flooding the market if you released too much music.

Things are changing now, at least on sites like Bandcamp, where prolific musicians can release as much music as they want, and subscribers support them. Often, though not always, this is linked to a more improvisation-based musical practice. I’ve been in regular touch with a couple people who do this kind of thing–Steve Lawson and Markus Reuter–and have benefitted from lessons from each of them in 2019. I been struck by how different their approach to creative work is. It’s neither precious nor perfectionist, which describes my approach to writing but less so to music.

In 2015, my new year’s resolution was “play a stringed instrument 15 minutes a day.” It was the only resolution I’ve ever really kept and keeping it changed my life. I am now in two bands and music is a huge part of my life again. I am happier for it. It worked because it was simple and let the implications come.

After my successful November of writing I considered signing up for “Jamuary” where artists record and release a track each day. But I neither want nor need that speed of output. I don’t actually write like that (the writing does not appear immediately after composition, except here). I’m looking to develop a long-term habit here, so it’s more about setting up and using recording templates in my software, and having my music space “ready to cook” just like my writing space is (they are also now the same space, which is really nice).

Health update / McGill Bullshits on Divestment

Health update: I went back to the oncologist today. My tumour marker is back down, which means last time was a blip, or variance, or measurement error. Which is what I expected. The BP drugs are still not working so we are trying yet another one. Which is good because I’m getting monster headaches from one that I’m on right now. No other interesting symptoms to report at this time.

In other news, McGill recently announced they were “reducing the carbon footprint of their investment portfolio” which is a bullshit-PR way of saying that they are either too afraid to divest from fossil fuels or that they actually don’t believe we are in a world-historical climate crisis. I don’t know which it is, but at some point one has to ask why divestment is possible for plenty of other Canadian universities, but not McGill. As of yet, the Board of Governors has not offered a convincing explanation.

NaNoWriMo / #academicwriting month: a review

Bottom line: November is a terrible month for academics to make a push on writing. I recommend picking a different month. But it worked for me. I benefitted from publicly announcing a schedule and from the support of friends who knew what I was doing. Word counts ended up being useless to me: in most cases I far exceeded them. As I said at the beginning, producing text isn’t my problem. It was all about sealing edges and creating continuity.

If I were to do it again, I would:

  • probably choose a different month
  • again publicly declare my project and deadlines
  • build in a little more flex time for things to come up and go wrong
  • find some more ways to create goals related to finishing things up rather than sheer word count
  • again start with space to ignore reviewing and letters (mostly) for 4 weeks (there’s a big pile to get out this next week)
  • if it’s during a term, it needs to be a term with a light service load and a compact schedule

When I turned in my manuscript last Tuesday, it was immensely satisfying, though I was immediately struck by how tired I felt after pressing “send.” Finishing in November wasn’t easy, especially toward the end. If I hadn’t set a deadline for myself, I would have taken a few more days (or half-days) off towards the end, and possibly finished in December. As it was, even though I budgeted for 1 bad day a week (defined as “feel too bad to work” as opposed to “not feeling perfect”), there were a couple weeks with more than that, which meant I wound up pushing the deadline and even cutting a couple corners I wanted to square before submission. During the last week of writing there were definitely a couple times where I really had to “push through” to make it. That may show in the writing, or it may not. Whatever, I get to revise.

But sometimes an artificial deadline is good. My original deadline for the manuscript was 31 May 2019. That wasn’t even close with my medical adventures, and the surprising difficulty of writing what because two voice chapters. Then I set 31 August as my deadline. I whiffed soon that one, too. So then I put the manuscript on my grad syllabus, and that was enough. Even so, without the push in November, I wouldn’t have made it. Or maybe I would have if I hadn’t been sick all of September. I don’t know.

In some grand cosmic way, not getting it done on the 26th wouldn’t have mattered, except that I was at a bit of a traffic jam in my work: two book manuscripts at the same exact stage is a bad place for an academic to be. Now I’m not in that place. I’ve got one book manuscript under construction and one under review. That’s much better.

A few other things made this possible. My teaching schedule this term is compact: seminars that meet once a week W and Th, with some new preps but not a lot in November; I also don’t have a lot of students in my classes, which is also unusual. Because of the drugs I decided not to travel after Minneapolis this term. One of the things that’s clear is the amount of time and energy travel eats up during a teaching term, especially now that my recovery times are longer. There is simply no way I could have done what I did if I had a “normal” (for me) travel schedule. On letters of rec, I have moved to Interfolio for mass mail-outs, which means less weekly time spent doing clerical work. I still have to write them and send them in some cases, but it’s a lot lighter this year. Also more jobs are not asking for letters up front. Thank you search committees. And I have been careful with reviewing, doing stuff, not not excessively (much of my reviewing has happened in doctors’ waiting rooms this term).

Finally, I had to draw some boundaries. I had one situation where I kept getting a manuscript sent back with additional requests for checking small things, editing, etc. At a certain point, I simply said I didn’t have more time to work on in November. I made myself unavailable for some meetings people wanted to do on writing days. Writing advice always say to do this but as for a lot of academics, I have a hard time doing it. I think in some perverse way, the fatigue makes saying no easier, because I feel the cost of saying yes.

So that’s my review.

I thought a little bit about privilege as well. Obviously I benefit from institutional security and the personal security that comes with it, as well as a relatively light teaching load compared to most professors. At the same time, I would wager that my body is less cooperative than that of many academic writers at this particular moment. And while I have the benefit of research assistants and the like, RAs don’t have a lot to do with it once I am at the point of actually solo writing (they certainly transform the research process, and increasingly, also come on as coauthors).

I am coming around to the idea that the writing thing is really about time and mindset more than everything else. I don’t mean that in the neoliberal way: those aren’t necessarily things a person controls; mindset comes as much from the support you get from others as what you feel in yourself. But if we are talking about sitting at a computer for hours issuing forth mountains of wordage for a month, I think most academics, regardless of career stage or institutional setting, can do it IF they are able to carve out the hours and pick the right month. Four large courses, or two small children, or lots of travel or service or conflict, or lack of a support network will obviously get in the way. But all of those things are temporary conditions (even someone with a heavy course load has times of the year they are not teaching). There are real obstacles to writing and thinking in modern academe. But there is no point in waiting for ideal circumstances to write. Some conditions are necessary and sufficient, but many of the “it would be nice” things we elevate to necessities in order to write are, well, not necessary. This term I taught and wrote while feeling like shit. I wouldn’t have felt any better had I not taught or wrote.

*

I said that once I finished, I wanted to play with my synthesizer. I had one letter to send on Saturday, but Thursday I went to the doctor and celebrated thanksgiving. Friday and Saturday I mostly did play with a synthesizer (confession: I own more than one synthesizer). Now I’m blogging and watching football and will soon make chili for the week (we are mostly through our leftovers).

Tomorrow it’s back to work but I will be ready.

Medical update, 1 December

On Thursday, which was US thanksgiving, I went to the oncologist’s office. Things are basically fine and in a holding pattern. My tumour marker is up slightly, but so slightly that it might be within the margin of measurement error. I get measured so often, I am not at all worried. My blood pressure is still high so we are boosting the blood pressure meds.

Side effects wise, it seems I am less stiff now, but this week’s new side effect is periodic headaches. The usual drugs seem to take care of them. Everything else is also steady: I have good and bad days, and about once a week something happens to my body that takes time, but otherwise no surprises.

I definitely felt overwhelmingly tired upon submitting my manuscript, but that is for another post.

Writing Update: Made It!

About 30 minutes ago I finished a draft of the manuscript. It runs from beginning to end. There are still some notes to myself in the footnotes, but it’s readable. I sent it to my students as promised, and I also sent it in for review. Later this week I will start distributing it to people who said they want to read part or all of it (and provide comments–that’s the deal).

Academic Writing Month or NaNoWriMo or whatever it was called wasn’t perfect, and November is a stupid month for academics to binge write, but it worked for me in a lot of ways. I am not sure I will ever choose to do it again in November, but it was nice to have a goal and make space to reach it, even during a busy semester. The next couple weeks I’ll have to do all the things I put on hold (recommendation letters! reviews!) but that was the deal.

I am also planning to take more time off, especially over break. I did manage mot of my Sundays off, and I think I lost about a day most weeks to medical stuff — either doctoring or something going wrong that laid me out–but I kind of sort of budgeted for that. I came in about 2 days later and more crowded at the end than I thought, but I’ll take it.

Writing and Medical update #2

I am now officially working on a chapter about fatigue while fatigued, though I am not writing about my own fatigue (except for a small conceit at the very end). And as I said earlier, it’s not so bad–I’m a pretty happy person with a word processor window open and a pile of books on my desk (and a bunch of pdfs open in browser windows).

Still, last week was not as productive as I’d hoped. Monday and Tuesday were pretty good but Friday and Saturday were not, between medical appointments and sheer exhaustion. I had to bail on work Friday afternoon after seeing my endocrinologist–I’d planned to sit and write but instead went home to rest. In fact, I was so tired Saturday evening that I bailed on a party I really wanted to attend. And I tend to draw energy from other people, so you know it was pretty bad.

Today I was back at it in full force, but I suspect I’m about two days behind the schedule I’d set for myself, which is going to make it tight for the 26th. We’ll see. Today I far exceeded my words goal for the day and the chapter is taking form and morphing from the rough outline I had, which is good.

As for the fatigue, I feel pretty good today after resting Saturday night and all day yesterday. But toward the end of last week I had that horrible waking-up-tired-after-9-hours-of-sleep feeling. Today was really the first day since Friday it didn’t feel like that. Everything I’m reading on the history and theory of fatigue describes it as an absence of energy or affect but I sure feel like it’s something inside me, and not an absence at all. When it’s happening, it’s like a weight, like tiny stars exploding inside, like a sticky soup that coats the inside of my skin. When I wake up fatigued, I know it’s there right as I’m coming into consciousness.

Is it a combination of drugs? Is it the Lenvima doing something to me? is it just that everybody’s tired now that it’s November and getting darker and it’s that part of the semester? I guess I will find out….

In other news, I cut myself shaving on Monday last week and it kept bleeding. Then it would heal and bleed again. Friday I learned that that too is a side effect of Lenvima, such that I am supposed to go off the drug for at least 72 hours if I am going to have surgery.

#goodtoknow.