All posts by Lloyd Davis

Musical slop

At at the weekend, there was a session on AI music. I didn’t go, but a couple of friends who did mentioned it to me at lunch.

It seemed from the descriptions I heard that it was about the ease with which generative AI tools can make music that sounds like other music. The use case seemed to be “say you like Chopin’s Nocturnes. If you fancy something that sounds a lot like that, but are a bit bored with the canonical works, and Spotify’s recommendations don’t hit the spot for you, this tool could make you a playlist of stuff that satisfies that need.”

I choose those Nocturnes as an example because I’ve recently found that a) they’re really good for calming my otherwise feverish brain when overstimulated and b) that I do indeed find myself wishing there were a few more than however many opi actually exist and yeah Spotify hasn’t shown me anything yet that fits the bill. Debussy? No thanks.

Back to the AI. So I nodded along with this and asked questions in the way that you do at a barcamp. And thought, “that’s something shit that I hope never happens”. And then, yesterday, while perusing the Tube of You, I saw a recommendation for “Vintage Music For a Rainy Autumn Day” from a channel called “Vintage Relaxing”. I’m not going to link to their videos directly so that you can choose whether you want to give them any more view minutes, but as far as I can tell, they’re hour-long playlists of otherwise unidentified bits of music in a 1930s/40s generic big band style. I further surmise that these are machine generated – ie they are musical slop (“low-quality media—including writing and images—made using generative artificial intelligence technology”).

And, of course, because I did listen to some of this, there was more of the same, but different, in my recommendations this morning.

Screenshot 2024-11-28 at 09.38.52

It’s 3 hours long.

So much to think about here:

Some people (I’ve met them, believe me!) have always argued that spam is in the eye of the beholder (what an image!) ie that one person’s spam is another person’s nutritious pork-based product. And, by extension, this is slop if you don’t like it, but jolly old-time music fun if you do. I don’t buy this. There is a line and I think this stuff crosses it.

I’m not in the “all AI bad because theft (past, present or future) from creators” camp. But I do think there’s probably a horrible easter egg surprise in this for people making music on the web – not least the prospect of getting copyright strikes on YouTube (and ultimately everywhere your music is available) because someone like this posted a shit-ton of generic stuff that you end up sounding like, but they got there first. When I saw this argument presented first, I was a bit “yeah I can see that with cinematic ambient stuff that all sounds the same to my geriatric ears anyway” but now I’m not so sure. It’s likely already there for your own best-beloved genre and you just haven’t seen it yet.

The “for you” style of algorithm is out of control. A little curiosity gets interpreted as eager attention. On the user side, we’re going to need *another* set of new filters to say “not that’s *not* for me, thank you” and while we’re waiting, I think the filter will be at the app/platform level (ie “I’m just not looking at Instagram reels for now”). But because it so fiercely and effectively leverages under-priced attention it isn’t going to go anywhere soon.

I’m really trying not to be a purist about the music itself – my daddy taught me to be open-minded in my listening while knowing your own preferences. But it is also shit. It’s dressed up muzak that goes nowhere, real-life bubble-gum for the ears whose main purpose is to rack up viewing hours from low cost content creation and thereby extract highly profitable advertising revenue from your ongoing attention.

It’s also just weird. If you do decide to listen carefully, you’ll ear the audio equivalents of how badly generative ai portrays fingers. Is that a guitar solo or piano? Is that trumpet or alto sax? It’s none of them, it’s an approximation of a bunch of sounds that is similar to something that was once organic. It has less soul than a replicant.

A warmer why

The situation: after 4½ years of pandemic-encouraged working from home, organisations are still trying to get people back into the office and it’s still not working.

tuttle2texas
#tuttlela – folk in Long Beach working together in a café – March 2010

You may or may not agree with that basic premise – maybe you’ve given up trying to get people back in, or maybe they’re all back in, or maybe WFH is working for the people who like being remote and the office is working for those who like being in the office. Also, people flow easily between both modes and any work that needs doing gets done, whether the people involved are in the office or on a beach in Madeira. Yeah right.

For most people I speak to, it’s a long-term conflict that has possibly reached stalemate. Progress towards resolving this conflict has been made, but people broadly fall in to those for whom working from home has been a liberating godsend and those for whome it was a nightmare that they couldn’t get away from quickly enough.

Note that I say “progress towards resolving this conflict” – not “progress towards getting things back to normal” or some such weighted phrase. I don’t know what’s best for you or your organisation, but I do think that the door that was opened by lockdown situations (at least here in the UK) showed us a glimpse of other ways of working that we’re still struggling to come to terms with.

In the meantime, positions are hardening around whether your physical location has a positive or negative effect on your productivity and the overall success of your organisation.

Long-term readers will remember that (consults notes, faints, recovers) for at least ten years I’ve been saying “We will work anywhere, but not necessarily the same ‘anywhere’ every day” (captured in this talk I gave in October 2014 – oh there’s also video of it)

So I’m biased towards individual freedom in this matter, but I also recognise that asking for total individual freedom within a business or organisation that has a main purpose other than the total freedom and happiness of its employees (sadly, most of us) is always going to come up against that purpose eventually.

So my evidence is very anecdotal, but broadly the arguments seem to be

“I’m more productive if I can choose whether and when I come to the office. I’ve found that working from home releases me from distractions that I can’t escape when I do go in. I realised that those distractions had been detrimental to my mental health.”

versus

“We believe that our productivity and general mental health is enhanced by having all of our people in the same place. Also head office (or whatever, some higher authority than the person making the pitch) have decided that we need to have people in for at least three days a week and so we’re already recording how much time people are here”.

Perhaps you’ve heard more.

I think we’re on a hiding to nothing by rooting this conversation only in terms of productivity and mental health. Mental health, because it’s hugely subjective and one size does not fit all. Some people have been driven crazy by being at home *and* some people have (possibly always) been driven crazy by being in an office – and some people drift between the two over time.

I also agree with Cal Newport that the trouble with looking at “productivity” is that we just don’t have good measures of what it is in knowledge-based work nor do we have solid facts about how to achieve it and so we end up with “proxy measures of productivity” like busyness, messaging overload and presence in the office. And although not everyone has the language to describe it, we all know that something is off with the way we work together.

So it seems to me that the people who are pushing (or are being pushed) to make sure the office is full of people again need to come up with a warmer version of the “why” – it can’t be just about “your contract” or “an edict has come from up top”. It has to be couched more in human terms about how its honestly affecting all of us so that we can put that distraction aside and get on with what we’re supposed to be doing – serving clients, helping people, making cool stuff that will help people even more and having fun (or at least not killing each other) in the process.

microbloggage 2024-11-26

a reverse-chronological list of things I’ve posted today to lloyddavis.micro.blog – replies aren’t included

10:21: After 20 years of being trained to use web-mail, I think I might be ready to slouch back to something desktop-based (mainly because multiple accounts with different providers). What are you using, cool kids?


This is an experiment in trying to pull together all the things that end up in all the places.

microbloggage 2024-11-25

a reverse-chronological list of things I’ve posted today to lloyddavis.micro.blog – replies aren’t included

08:50: I’m waiting for a (non-urgent) telephone appointment with my GP. It was scheduled for 19 minutes ago. I’m reminding myself that this is more convenient than walking over to the surgery and sitting in the waiting room, despite my increasing levels of impatience.


This is an experiment in trying to pull together all the things that end up in all the places.

microbloggage 2024-11-18

a reverse-chronological list of things I’ve posted today to lloyddavis.micro.blog – replies aren’t included

15:49: There’s more than one way to skin a cat, but I’m not allowed to say that in our house any more (ditto comments about how little room there is in my office).


This is an experiment in trying to pull together all the things that end up in all the places.

microbloggage 2024-11-16

a reverse-chronological list of things I’ve posted today to lloyddavis.micro.blog – replies aren’t included

14:26: I’m going to barcamplondon xiii next week. I went to number 3 in 2007 and number 10 in 2014. There’s an arithmetical gag in there somewhere.


This is an experiment in trying to pull together all the things that end up in all the places.

microbloggage 2024-11-15

a reverse-chronological list of things I’ve posted today to lloyddavis.micro.blog – replies aren’t included

15:52: Good lord! It’s Friday afternoon again.
12:58OK. I’m seeing many places where this needs saying.: Many things can be true at the same time.

Even things that you have always thought were mutually exclusive.

I think the work for us to do now involves making spaces (within and between ourselves) in which all the overlapping, layered narratives can live, rather than trying to find one unifying story for us all to get behind.

It’s work, it’s not easy, it’s not necessarily how we were raised, or how we’ve lived for most of our lives, or even how we’d like it to be.


This is an experiment in trying to pull together all the things that end up in all the places.

The Enemy Lingers On

I’m reminded that it’s now eight years since I wrote this song:

At that point, it was easy for some to say that we didn’t know what was coming. I think it’s less easy now. Even so, the temptation remains to think that these men don’t really mean to implement the things they’ve talked about in election campaigns, that to do so would be crazy. But we live in crazy times alongside crazy people and the line between crazy and sane isn’t as clear cut as we’d wish in any of us. Neither is it as straightforward to diagnose in others as seeing whether a person cast their ballot the way we agree with or not.

I updated this song a little when I released it as part of “Half A Pound Of Monkeybread” – in particular, I wanted to point at fascism where others saw bufoonery. If more bufoonery is all we get, I’ll be very glad to be proved wrong.

They seem to know me…

I’ll blog about this properly when I’ve got my head around it a bit more.

Basically, I took a bunch of Tuttle-related videos and documents and fed them to Google’s NotebookLM – one of the standard outputs from that is an “Audio Overview” examples of which you may have heard elsewhere – they take the form of a “deep dive” podcast between a man and a woman, all artificially generated.

I followed the instructions on this video to make avatars of each of the speakers and animate them with lip synching.

I added the following prompt when generating the original audio: “Imagine it’s 17 years after the founding of the Tuttle Club. Target an audience who knows of Lloyd, but isn’t aware of Tuttle or his contribution to it.”

There’s another 14 minutes of audio, but I only had the patience to do the avatars for the first 50 seconds.