Adam's Brain Dump

📚 Finished reading Upgrade by Blake Crouch.

I read and enjoyed “Recursion” by the same author earlier this year, so was enthusiastic to try this, his latest novel, out.

It’s set in a somewhat dystopian but very recognisable world of the presumably near future. We haven’t solved our environmental problems, in fact parts of Manhattan are unusably flooded amongst other such places. But technology has advanced a bit, in particular our ability to edit genes.

Although gene editing was outlawed following a misguided attempt by top scientist Miriam Ramsay to enhance the resistance of rice to a particular blight. Best of intentions maybe, but there were of course unforeseen consequences which led to a mass starvation, hundreds of millions of deaths, and a ban on genetic engineering. Miriam killed herself.

The ban is enforced by the Genetic Protection Agency, where we find out protagonist, Logan Ramsay, working. Logan is Miriam’s son, who seems to be working there more out of a sense of guilt for the impact his mother had on the world - he himself was involved enough to go to prison for a while - than a love for the job.

One day, a raid goes wrong and he’s exposed to an unknown virus. The symptoms are agonising at first, but he recovers his health soon enough. And more besides. Suddenly he feels stronger, more intelligent, more sensitive, with a better memory. He can even beat his daughter at chess. Until, imprisoned for genetic self-engineering, he no longer has the opportunity to.

Then a figure from his past life turns up, also stronger, fitter and cleverer than either of them had suspected. The problem is that they strongly disagree what they should do about it. The potential consequences of the decision could hardly be higher.

To get to the bottom of that requires resolving several deep ethical problems. What risks do we have the right to take in the name of a potentially better future? And even what does it mean to be human? Not that you’ll need an ethics PhD to understand the situation, honestly it’s mostly an action thriller, substantially less intellectually demanding to me than the last work of fiction I read was. But the conundrum is real, and adjacent to one that humanity is already facing.

I’m not yet sure how I felt about the end of the epilogue, but was fully engrossed throughout the main story.

Book cover for Upgrade

"The Extended Mind" teaches us how to improve our thinking in some counterintuitive ways

📚 Finished reading The Extended Mind by Annie Murphy Paul.

This was one of those books that are so fascinating that when I attempt to highlight the interesting bits I find myself highlighting the whole book. I’d heard good things about it and it certainly delivered.

It’s on a topic that’s of great import and interest to me - essentially how to think “better”. It also share a thesis that wasn’t entirely intuitive to me in a way that suggests some practical avenues one can easily change one’s behaviour in order to improve the output of your cognition. What more could be asked?

The general idea is that our society has fixated far too much on the idea that the individual brain is the sole locus of thinking, of cognition, of creativity and problem solving.

It’s the “brain as computer model”, as though we each have a self-contained isolated CPU in our heads that performs our thinking at whatever speed and quality nature saw fit to bestow upon us. That would suggest that aside from sitting at a desk an individually focusing on formal education and training courses, perhaps we could improve our brain’s performance by learning new ways to think about things within our brain, some lifehacks, engaging in brain training or supplementation, but that’s about it. Intelligence is a “fixed lump of something in our heads”, and we design our home, schools and workplaces around that idea

But the author contends that this is all wrong. Modern research has shown that the brain should be thought of more as a magpie. It creates its output out of the materials it finds around it. Thinking uses resources external to the brain, and the nature of the materials available affects the quality of the thinking. Intelligence, as properly considered, isn’t fixed, but rather a “shifting state” that depends on the level of access you have to resources outside of your brain and your ability to leverage them. This book gives you the knowledge to try and improve both of those for yourself and those around you.

And this is essential work! Modern day life involves absorbing a ton of information, often in abstract forms that we had no reason to evolve to be good at, so we aren’t. Many of the challenges and tasks that face us are extremely complex.

It isn’t for nothing that the amount of journal papers and patent applications that have a single author is dwindling over time. At some point much “interesting” work may have exceeded the natural capability of almost anyone’s individual brain, no matter how ensconced in an ivory tower it is.

The book is grouped into three main sections, each of which deals with a different sphere in which we can consider extending our mind.

The first is using our bodies, rather than our brains to help us think. We are more than brains on a stick. The second is adapting the environments we reside in to enable improved thinking. In some ways the trend in recent times has been to engineer ever worse environments for thinking as time goes on. And the last is to leverage our social natures, to “think” with our relationships to other people. Groups of people can be more than the sum of their individuals, particularly if proactively designed to work as such.

The argument is that at present we’d be much better served in spending time figuring out how to improve our capabilities in using these external resources than undergoing individual training to improve some personal skill.

Three sets of general principles arise. The first concerns the “habits of mind” that we should adopt to improve the output of our thinking.

  • Offload information from your brain whenever you can. This could mean anything from writing down your thoughts or “socially offloading” them to other people.
  • Transform information into artefacts, ideally physical ones. Interact with them, tweak them, show them to other people.
  • Be proactive in altering your inner state. Take some physical exercise before trying to learn something. Synchronise with others before you attempt group work. Spend time in nature if you have an upcoming creative task.

The second set of principles uses our understanding of what the brain evolved to do in order to grasp how mental extension works.

  • Aim to “re-embody” information. Allow your body’s interoceptive signals to influence your choices. Use physical movement to enact concepts. Focus on your gestures and those of others.
  • Re-spatialise information when possible. Our brain processes information via mental maps. Use memory palaces, concept maps and the like to leverage that evolved capacity.
  • Re-socialise information. We process it better when we involved others. Teach others what you know, learn from them, imitate, argue, debate and tell each other stories.

Lastly we have a set of principles based on “what kind of creatures we are”.

  • Deliberately create cognitive loops. Use your body to help you think, then spatialise the information, then run it through the brains of other people. Keep looping it through each realm, again and again.
  • Create “cognitively congenial situations”. Issuing orders to your brain is a strategy often destined to fail. Instead, create environments that draw out the desired result of your thinking. Explain things to your peers, share stories, create a meaningful private space, walk in nature. What you do should depend on your cognitive goal.
  • Embed extensions in your day-to-day environment. This can range from arranging “identity cues” in your workspace through to deliberately cultivating a transactive memory system with your colleagues.

I guess the main takeaway is that the best thinking does not generally occur when you are sat still on your own in a bland, neutrally-lit grey office of the sort often designed to aid thinking via being “distraction-free”.

But other people, other environments, other ways of being, your feelings and emotions are not always distractions to good thought. Far from it. Often thoughtfully seeking them out and leveraging them to extend our minds can produce far better thinking outputs, much more suited to the modern world, than the environments we’ve historically designed.

The title of the book seems to be a shout-out to Clark and Chalmers' paper “The Extended Mind”, which is referred to in the text. In that they argue in a somewhat similar vein for the “extended mind thesis” - namely that the mind isn’t limited to our brains, or even our bodies. Rather that external items in the physical world - a diary, a computer - can be considered as part of the cognitive process and as such as extension of the mind.

My full notes are here.

Book cover of The Extended Mind

🎶 Listening to GUTS by Olivia Rodrigo.

Everyone’s favourite childhood High School Musical character to present-day singing a wildly popular song about getting a driving license artist, Olivia Rodrigo, is back with a new(ish) album in 2023.

It’s at least as good as her first one. There’s a variety of styles on show, plenty of pop-punk , with some big piano ballads and a smattering of rap. Much of it riddled with the self-doubt, insecurity, anger and clever disses about idiot men that seem to be part of a lot of the newer music I chance upon these days. Sign of the times I guess.

Her first album was also pretty great but everyone’s already heard it so you probably already knew that.

Everyone was right, The Three-Body Problem is great

📚 Finished reading The Three-Body Problem by Cixin Liu. I remember this book got pretty rave reviews when it was published in English (and probably before - but I can only read English). Obama liked it. George R. R. Martin, Bill Gates and Mark Zuckerbergtoo. But don’t let that put you off, plenty of people with less wealth and, depending on your point of view, megalomaniac desires like it too. It also won a Hugo Award in 2015.

I left it some time, trying to wait until I felt like I’d the opportunity to really dedicate time and focus to appreciating it fully. Naturally that time never came, but thankfully I gave it a go anyway. And, predictably, also loved it.

The subject matter is pretty attuned to my interests. We start with 1960s Chinese politics, in the midst of the Cultural Revolution. Our protagonist is unfortunate enough to watch her father die in a struggle session. She herself is arrested for suspicion of not being a good enough Communist. The prospective penalties are severe.

By the end we’ve worked our way through alien civilisations, conspiracies, detectives, spies and, metaversey-stuff, with some nice doses of science and philosophy. Inexplicable phenomena abound; why are a bunch of scientists killing themselves? What condition is causing someone to see numbers wherever he looks? What actually is the purpose of the strange military installation? Or the VR game that seems to have appeared out of nowhere?

But what always fascinates me most of all are authors’ ideas about the weird ways that human society might react when very important, very unprecedented, situations occur. Some of the ones herein might seem unlikely at first, but then again very little can be more wild that the IRL emergence of the QAnon cult and its troubling downstream effects. Would Pizzagate sound realistic had it only appeared in a sci-fi novel?

We’re told that so far “the entire history of humanity has been fortunate” - hard as that is to believe in current times - but sure, perhaps there’s a way in which that’s true in comparison to what is on the agenda here.

The book is not difficult to read despite the depth of the intellectual topics. It’s full of big, awesome, imaginative, and occasionally explicitly philosophical ideas that consumed my brain for a while. I’m sure it’ll stick in my mind for a good long time.

It’s the first book of a trilogy. There’s no way I’m not going to read the others at some point after both the enjoyment this one brought and the way that it ended.

There’s also a TV show version of it expected to come to Netflix next year.

Book cover for The Three Body Problem

In 2019 OpenAI was too scared to open GPT-2 to the world

Remember the halcyon days of circa 2019, when OpenAI were too nervous to release GPT-2, yes, two, on an unsuspecting world?

…OpenAI said that it would only be publishing a “much smaller version” of the model due to concerns that it could be abused. The blog post fretted that it could be used to generate false news articles, impersonate people online, and generally flood the internet with spam and vitriol.

A lot changed in a few years.

2 generations of GPT later and now anyone who has $10 to spare can access the latest greatest GPT-4 - which has recently learned to ‘see, hear and speak’ in-between its only-increasing potential to ‘flood the internet with spam’ and ‘generate false news articles’, which it is indeed doing to at least some extent.

I’m not saying that’s necessarily a bad thing. Well, except the spam and fake news parts - these days there exist several extremely low effort websites that have leveraged it for exactly that purpose. After all, rich companies hoarding rare resources is rarely the most obviously optimal way to benefit humanity as a whole.

Some in the machine learning community have accused OpenAI of exaggerating the risks of its algorithm for media attention and depriving academics, who may not have the resources to build such a model themselves, the opportunity to conduct research with GPT-2.

Perhaps slightly Ironically, they seemed a lot more about putting the Open into OpenAI back in those days, philosophically at least. What as far as I know is their original manifesto is still there on their site.

OpenAI is a non-profit artificial intelligence research company. Our goal is to advance digital intelligence in the way that is most likely to benefit humanity as a whole, unconstrained by a need to generate financial return. Since our research is free from financial obligations, we can better focus on a positive human impact.

We’re hoping to grow OpenAI into such an institution. As a non-profit, our aim is to build value for everyone rather than shareholders. Researchers will be strongly encouraged to publish their work, whether as papers, blog posts, or code, and our patents (if any) will be shared with the world

Now we have proprietary secrets - ‘OpenAI’s GPT-4 Is Closed Source and Shrouded in Secrecy’ says Motherboard - as well as exclusive deals with anyone who has a few billion dollars to throw their way.

“The Data Warehouse Toolkit” teaches us dimensional modelling and the Kimball method of data warehouse design

📚 Finished reading The Data Warehouse Toolkit: The Definitive Guide to Dimensional Modeling by Ralph Kimball and Margy Ross.

Most everyone that’s been in charge of designing a large database has recommended this to me as a good summary of how to think about the structure a data warehouse.

It teaches the Kimball method - you might notice that’s one of the authors' surnames. So if you’ve come across terminology such as ‘type 2 slowly changing dimension’ or the facts and dimensions used in dimensional modelling et al. then it’s very possible that your interlocutor has read this tome.

Whilst I’m more an analyst than an engineer, I have dabbled in the latter and this book will be good guidance on how to think things through going forward.

A key aspect, and one I’m grateful for, is to focus on making the data easy to understand and analyse in downstream tools, even if it requires substantially more effort - both technical and diplomatic - up front.

It’s been helpful to me to understand why systems I use are set up like they are - e.g. a distinct absence of using null value where I’d initially thought it’d make sense to use them. It’s also a refresher on basic patterns one can use to enable common analysis requirements like ‘how many times has this value changed over time?’ or ‘what would the results look like if they were remapped to the historical structure of the organisation?’

It’s written rather prescriptively - follow this rule or regret it forever! I have been told by practitioners that sometimes it’s necessary, or at least preferable, to break the occasional rule in reality but that their recommendations are good ones for the majority of the time.

It’s also organised slightly confusingly in that most of it is divided up into chapters seemingly aimed at various specific business applications e.g. e-commerce or insurance. But they’re at pains to say that you shouldn’t just read the one that’s most similar to your organisation or the task you’re trying to accomplish as you’ll not understand it fully without all the context provided in the earlier chapters that from their title don’t seem relevant to you. So I’m not entirely certain why they structured it like that in the first place.

But that aside, it’s invaluable reading for anyone designing databases or using other people’s databases that are designed this way.

Book cover of The Data Warehouse Toolkit

The Conservative party appears to hate law and order

Remember when the Conservatives thought they were the party of law and order? Hilarious.

I’m not thinking here of the numerous illegal cringe-parties the government and its associates held during the Covid-lockdown. Although do watch the docu-drama “Partygate” if you want to increase your level of annoyance at their doings even beyond whatever it currently is. But here I’m thinking about rather more structural stuff.

Like how our former Home Secretary tried to coerce the police to ban certain peaceful protests on the topic of Palestine. The police said no, mainly on the basis that there is no law that would let them do so despite her suggestion that waving a national flag might be illegal - obviously Union Jacks would be exempt. She wasn’t at all embarrassed, but rather wrote a rant basically about how the police are too woke in a national newspaper which, thankfully, eventually, at least in part, got her fired.

It comes to quite something when we have to rely on the shattered remnants of the British police of all organisations - a large contingent of which have recently been officially determined to be “institutionally racist, misogynistic and homophobic” and seemingly pretty terrible at actually solving crime - to be the last desperate hope against an illegal trampling of our civil liberties, to be the “woke” ones. But credit where credit’s due, I applaud their actions in preventing a crime taking place in this case.

Now the courts ruled that the obviously illegal plan to ship off applicants for asylum to other countries was indeed illegal. Despite the fact that it seems very plausible that his latest Home Secretary described the plan as “batshit” in the past, Sunak once again isn’t embarrassed. Instead he’s trying to find some way to get through dubious new legislation that “allows” him to ignore the court’s decision, or pander to the worst of his party’s instincts to exit the ECHR, as well as presumably any other legislation that has a chance of protecting the outgroup, of binding the ingroup.

What some of his party say is up there in terms of incitement to the famously irresponsible “Enemies of the People” Daily Mail front page a while back. The police believe that Braverman’s comments on the protests were an important driving factor in the violent attacks that far-right protesters - note that this was a group that she wasn’t trying to ban - went on to perpetuate against them.

Further back of course there was all the Brexit stuff. Boris Johnson tried to pro-rogue Parliament to avoid mere democracy getting in the way of whatever his favourite idea of the time was at the time. This was deemed to have been illegal.

More recently, when the quagmire of the Brexit they’d “negotiated” continued to become ever more apparent, some of the Conservative party - including their leader - decided that they’d unilaterally invent some nice new legislation so that they could ignore the deal they’d already signed up to regarding trade with Northern Ireland. Even their own ex-leader, Theresa May, said the Northern Ireland Protocol Bill was illegal under international law.

There was a time when they also claimed to be the good at the economy. Haha.

Humans will go to considerable effort to get obviously useless information

In a finding that may not surprise anyone who endures reading this blog, a recent study found that “Humans will trade pain for useless information”.

The useless information in question was the implications of the results of a coin-flip. Participants would get a reward based on the result of a series of coin-flips. But by default they weren’t told what exactly the prizes associated with heads or tails were until after the fact. Importantly, it was set up to ensure that the participants knew that they would receive the exact same reward irrespective of whether or not they knew the values of each side in beforehand.

Participants were then offered the chance to know what the reward associated with each side of the coin was in advance, but at the cost of having pain inflicted on them - a “flash of heat” applied to their arm. When the pain was set to low, about 3/4 of participants requested to receive the pain and learn in advance about the rewards, information which has no obvious utility given the setup. Even when the pain was set to maximum participants would choose to undergo it nearly half the time.

One theory the researchers have as to why this occurs is that as humans we’re just generally deeply uncomfortable with uncertainty.

This willingness to endure pain in exchange for noninstrumental information may stem from a deep-seated aversion to uncertainty, Bode says—to the point that some people are willing to go through physical discomfort for a few scraps of solid information. “Not knowing is really painful,” he says.

As noted in the original article, pain isn’t the only adversity we’re willing to go through in order to learn about something we know can have no effect on or bearing to what happens to us. Previous studies have shown that people are prepared to pay money or put in unnecessary physical effort in order to learn in advance how a lottery would pay out in a way that could not possibly affect how much they’d win.

Facebook permanently disabled my account, no right of appeal

Never trust a social network you don’t own, lesson number 999.

After a period of disuse, Facebook has permanently locked me out of my account for something I definitely didn’t do.

Some time ago - I guess a long time ago - I deactivated my Facebook account. I haven’t really missed it all that much. Anyway, yesterday I decided to reactivate it so as to download all the photos I’ve ever uploaded to it for safe storage, along with any that I’m tagged in that I might like to revisit one day. The latter isn’t a native functionality as far as I’m aware but browser extensions like this one purport to enable it.

This should of course be easy. From their docs:

If you’d like to come back to Facebook after you’ve deactivated your account, you can reactivate your account at any time by logging back in to Facebook or by using your Facebook account to log in somewhere else

However, upon logging in to reactivate the account I was met with this message:

Facebook message indicating account has been disabled

My account has apparently been permanently disabled for contravening their spam policy. And because it had been over 180 days since they decided to do that - which I had seemingly no way of knowing had happened given my account was temporarily deactivated - there is no way to appeal the decision.

Obviously I haven’t been spamming anyone given I deactivated the account but there we go.

At the very least, given I’d explicitly suspended my account so wasn’t logging in, I’d have expected to have received some kind of robotically generated email to let me know of the supposed spam crime within the 6 months window that would have allowed me to appeal if I cared enough.

The good news is that it was still possible to export my content; that’s the 1 button that’s still available even with the block. Although, it comes to you in a single big intricate zip file that may take some time to dig through. But if I had a desire to start using my Facebook account again - something that temporarily deactivating your account is explicitly supposed to allow - then that opportunity is gone.

As is my social graph of course. The photos are in the zip somewhere, along with a pretty useless list of all the comments I ever made stripped of context - who knows what it was that I loved so much 7 years ago!? But the connections between me and my friends are gone. I’m not aware of any social network you could upload that kind of data to even if your friends all agreed to sign up to it. Mainstream social networks are silos.

Also naturally it only exports the content I uploaded. So photos by me; but not the photos of me that someone else uploaded, tagged me and shared that I’d probably have liked to keep, let alone photos of friends or family, old or new.

This isn’t all that terrible from my personal point of view. I’ve little interest in using Facebook and I probably have better-quality backups of most of the photos. But some people might fairly be quite annoyed or upset in this situation.

As always the lesson to take away is that if you ever post anything to a social network you might ever care about, always keep a copy of it safe somewhere you control. And if you value your social graph, well, I’m not sure that there is a solution other than being sure to have some other way to contact your nearest and dearest.

Similarly with your interactions; the to-and-from flow of any meaningful conversations you had or any threads that meant a lot to you. You’ll be able to export your ramblings, but without the context of what you were responding to. We all know how annoying hearing only half a telephone conversation is. Again given the private and proprietary nature of the service, it’s hard to know how you can personally mitigate this. Perhaps it’s time to act like it’s 1997 and start printing out everything you found fun on Facebook?

I’m obviously an outlier because I haven’t been using my account for a long time. But other people find themselves unable to access their Facebook accounts for all sorts of reasons in ways that are much more upsetting to them. There are endless discussions online about it, Reddit is a good place to start if you want to find examples.

Here’s this phenomenon making national news last year - I assume it’s not an April fool despite the publication date.

…despite not being an avid user, finding her account locked was still upsetting: “All of the images from my university years and family occasions are on Facebook

“I will no longer have access to 15-plus years of content, which is genuinely sad

“It is also quite stressful not knowing what the issue is, and having no recourse to resolve it. To be given no warning and then no way to access our own data is mindboggling.”

Of course no-one has a legal right to have a Facebook account. It’s a private company entitled to make its own decisions outside of basic legal requirements. Personally I think a lot of people might be better off without an account. But, believe it or not, Facebook truly is a positive experience for some people. And its sheer size and scope creates a kind of monopolisation such that as much as you’re able to “export your data” - which is certainly a good and important thing - there is often little of use that the average person can do with that data.

You can’t simply replicate everything you’d made the effort to create in Facebook at home or in a competitor’s site. Even if you could, you wouldn’t be able to force the people and organisations you want to talk to to move over to someplace new, which is what you’d have to do considering Facebook is not a federated social network.

Being locked out forever would be especially upsetting for those people for Facebook is where their community is, whether this is how their family prefers communicate, where the organisations they have to deal with post or where their support groups are. Innumerable organisations - everything from book clubs to government agencies to medical services to advocacy groups and beyond - seem to have chosen Facebook or one of its equally siloed competitors as a primary place to communicate to their members. It’s a risky thing to do when you do not have the final say as to who is able to participate.

Sam Altman, the famed head of OpenAI seems to have been fired.

Mr. Altman’s departure follows a deliberative review process by the board, which concluded that he was not consistently candid in his communications with the board, hindering its ability to exercise its responsibilities.

The reason given seems so generic and vague that it feels like it’s either concealing some unspeakable sin, or that the AI has already escaped and chatGPT 5 is outputting its concept of a bland corporate press release in order to cover up what it actually did to its creator.

The Sherlock & Co. podcast updates the Sherlock Holmes stories for the Facebook era

🎙️ Listening to Sherlock & Co. podcast.

Meet Dr. John Watson, an army medic who had to leave military life after suffering an unfortunate field injury. He ends up joining the seemingly infinite hordes of folk who are trying to scrape a living together via becoming podcasters. Specifically in his case a true crime podcast.

Sure enough, soon a mutual friend introduces him to Sherlock Holmes. Holmes is bit of an odd-bod in the social realm but has uncanny powers of observation. They end up renting a flat from Mariana, who works for Hudsons estate agent. And suddenly Watson has a lot of stuff to podcast about.

These are the tales of Sherlock Holmes, each one acted out in radio play format, retold as though from the modern era. Gone are the letters of inquiry from Dukes. Instead we hear of the Insta posts from cringey male-feminist influencer wannabes. A little less reading through the day’s newspapers, a little more trying to geolocate a Facebook video, Bellingcat style.

Very fun, if you can tolerate the adverts. And a good excuse to read the originals again. So far they’ve re-done The Adventure of the Illustrious Client and The Noble Bachelor.

Sherlock & Co. podcast logo

There's now a dangerous black market in anti-obesity drugs

Demand continues to dramatically outpace supply in terms of getting access to the recent second-generation GLP-1 anti-obesity medications.

Inevitably a problematic black market has formed. The UK health agency, MHRA, warns us of the existence of at least hundreds of vials of entirely fake medication.

Some of the syringes being sold as containing semaglutide (Ozempic) or liraglutide (Saxenda) in fact contain insulin. This has led to some users getting ill enough to require hospitalisation.

To the extent that Novo Nordisk et al. appear unable to manufacture anywhere close to the amount that would match the demand perhaps some of this is unavoidable, if still upsetting. The Faustian bargain that sure, these companies might become dazzlingly rich - which they are, Novo Nordisk is now Europe’s most valuable company with a market cap larger than the GDP of the country it’s situated in, Denmark, not all that surprising when each course of these treatments retail at thousands of dollars in some places - but that it’s the only way to ensure that everyone’s demand is being met has once again not really been working out well so far in this case.

But for any unmet need that’s attributable to explicit or implicit gatekeeping or denying access to this medication to patients who could benefit from it, this example reminds us of what’s at stake when we don’t enable people to safely access medications that could dramatically improve or even save their lives.

Crime and Punishment - the original psychological thriller

📚 Finished reading Crime and Punishment by Fyodor Dostoevsky.

This is the book that has been acclaimed by some as the one of the first ever psychological thrillers, but I feel has also gathered a somewhat off-putting reputation as being difficult to read. Which I don’t think is deserved at all, at least not the translation I read. If you have any interest in it, be bold, give it a go.

We follow the life of young Rodion Raskolnikov, who lives a life of seemingly at least somewhat self-chosen poverty in Moscow. He’s something of an over-thinker, a philosopher of some kind. And so we find him pondering on whether or not he would be justified in murdering a mean old pawn-broker who has been less than generous to him and others in times of need in order to steal her money.

It’s no spoiler to say that between deciding that her life is immoral and overdosing on the writings of Hegel et al to the extent of considering himself as a candidate Great Man of History he manages to convince himself that it’s a reasonable, and perhaps even moral, thing to do.

…all men are divided into ‘ordinary’ and ‘extraordinary.’ Ordinary men have to live in submission, have no right to transgress the law, because, don’t you see, they are ordinary. But extraordinary men have a right to commit any crime and to transgress the law in any way, just because they are extraordinary.

This all goes down near the start of the novel. It doesn’t go quite to plan of course.

In what follows we watch him trying to come to terms with what he did and to deal with what will come next. He spirals down through life-threatening maladies, both physical and mental. The police are poking around, his family do their best to care for him but sometimes in ways he cannot abide, ways that will put them at too much risk even whilst he cannot find a single nice word to say to them.

As the noose of suspicion tightens we see his mind whirl between the possibilities of confessing all - perhaps the mental torment would at least end? - versus fleeing, leaving everyone who loves him behind - but what right has he to be loved? - vs the satisfaction - or is it delusion? - of getting one over on whomsoever he sees as his enemy at the time, interlaced with some contemporary politics and philosophical thinking

The frantic, frenetic, paranoid mind of a man who deluded himself into doing something he has no idea how to get away from, or even if he really wants to get away from it, makes for a compulsive reading experience.

We Are Bellingcat: Bellingcat's origin story and how they do what they do

📚 Finished reading We Are Bellingcat by Eliot Higgins.

This is the story of how Elliot Higgins went from curious internet user, through to an increasingly renowned commentator on blog articles, to starting his own blog all the way through to running an increasingly large and influential organisation dedicated to open source intelligence investigations called Bellingcat.

Open source intelligence, or OSINT, is “the collection and analysis of data gathered from open sources…to produce actionable intelligence”.

Bellingcat, named after the fable “Belling The Cat” , has famously used this technique to investigate and produce damning evidence on many geopolitical episodes, including the Malaysian Airlines flight MH17 crash, the poisoning of Putin’s enemies Alexei Navalny and Sergei Skripal, the unmasking of neo-Nazis in Charlottesville’s rally as well as various unpleasant and illegal incidents occurring in recent Libyan and Syrian conflicts . This book contains details of many of their increasingly complex and impactful investigations.

They also run training courses to let other folk learn their very transferrable, very versatile skills. They’re open about how, why and what they did. Transparency is one of the key tenets of their organisation. Knowing that there’s no reason for you to trust them beyond any other voice on the internet, they make a big deal of recording and sharing exactly what they did, what their sources were, what they inferred vs validated and all the outstanding questions that they couldn’t answer.

…an online claim is nothing more than a hypothesis, one validated only with backing evidence that others should be able to corroborate themselves.

Their motto became “Identify, Verify, Amplify”.

The most fascinating aspect to me is around the tooling. We all splurge our data every day all over the internet. Most notably on social media of course but many of our daily transactions that we don’t explicitly share are in someone’s database somewhere - records of your phone usage, travels, identity, financial transactions, property ownership and a whole lot more. Sometimes this kind of data is publicly available legitimately for free or at a reasonable cost. Other times hacked versions are floating about the lesser known parts of the internet. On occasion the Bellingcat team seem to slightly deviate from the “open source” aspect of OSINT and hold their nose and resort to paying off employees of various organisations to share data with them.

Knowing what sources they’ve found useful in their work both enables you to conduct your own investigations as well as be a bit more aware of what you’re potentially unknowingly sharing. This is surely a recommended skill for anyone who’s been alive in at least the past decade or so. Some of the tools they’ve used include:

  • Several Google products: Earth, Maps, Translate, Youtube, reverse image search.
  • An app called SunCalc that lets you estimate the time of gday a picture was taken via the shadows in it.
  • Search engines less known to the average British or American internet user, such as Russia’s Yandex.
  • Online catalogue and databases of munitions, vehicles, property.
  • Various specialist message boards, military sites like Janes or sites like Uxoinfo that describe unexploded ordnance.
  • Hobbyist sites - e.g. plane-spotting or license plate websites.
  • Wikimapia
  • Dashcam videos that have been shared online
  • Pixifly , allowing searching of Instagram by location and time (seems like this is now shut down).
  • Panoramio for seeing geotagged photos users post.
  • Any social media site you’re likely to have have heard of, and others you may not have - VKontakte, Odnoklassniki.
  • Zello, a chat where users share audio clips.
  • Digital Globe and other commercial producers of satellite imagery.
  • Checkdesk - an app that lets people sign up to join an investigation.
  • Syrian Sentry - an app where volunteers recorded planes taking off from military airfields.
  • Europol’s “Trace an Object
  • Calling people up on the phone to get an audio sample that can be compared with other snippets.
  • Leaked customer databases
  • Open source phone databases. There are also apps that share phone numbers people have in their contacts list such as TrueCaller - key here is that people’s names appear as they do in individual’s contact books. And some people list agents working for secretive organisations with the organisation as part of their name!
  • The many messenger apps that let you see if a given phone number is currently online.
  • Leaked data from phone companies showing where a phone was at some point in time.
  • …and many more.

Some of these may no longer exist in the form they were originally used in. But Bellingcat keeps an up-to-date list of tools they find useful in their investigations here - their “Online Investigation Toolkit”. It’s quite eye-opening.

For anyone who’s nerdy enough to be able to cope with running scripts or compiling code, they also develop their own in-house tools which are available on their Github repos to all and sundry. As they note in the book, there is a tension between being entirely transparent and open when building killer new investigative tools and the fact that many of them could be used for nefarious purposes by folks with bad intentions, so, you know, please use them wisely.

The Bellingcat method has endless applications. What unifies our work is a drive for accountability. We take scattered facts online and try to turn them into justice.

My fuller notes on this book are here.

Cover of We Are Bellingcat book

Is Britain on the verge of introducing an age limit on baking?

Poster refusing sale of eggs or flour to people under 16 years old

Meta-owned WhatsApp confirms (once again) that adverts are coming to its app.

Code to enable this was already written at least 3 years ago but withdrawn at the last minute. So this isn’t the first time they planned to do this. I guess they just got less worried about annoying either its regulators or its users.

Arcade claw machines are even more scammy than I thought

Only recently did I learn that those arcade claw machines where the game is to precisely navigate a claw over some stuffed toys or a priceless watch such that when it descends it grabs it and pops it in a chute for you to take away as winnings are literally impossible, rather than just extraordinarily difficult, to win most of the time.

Photo of a claw machine

It turns out that a given claw has different strengths at different times. Full strength will grab and deliver the prize if you navigated it exactly right. But at all other times the claw isn’t even in theory strong enough to hold the prize, so you’ve no chance.

How strong the claw can be, and often the claw is strong, can typically be chosen or changed by the owner as this extract from a manual that Vox shared shows.

Extract from a claw machine manual Another extract from a claw machine manual

Some claws may have several parameters. For example on occasion it might be strong enough to pick up the prize, but then then deliberately weakens and drops the object of desire just before it get over the retrieval slot, making for those ultra-addictive seeming near misses.

So really claw machines are not so different to “pure” gambling machines like fruit machines. Except you have to be both lucky and somewhat skilled. And they’re less regulated in some jurisdictions.

Based on that manual extract it sounds like there might be a viable strategy involving secretly watching other people playing with the claw until it’s been a while since anyone won anything and swooping in in the hope that strong mode is about to trigger. If you can live with yourself and really want that big fluffy toy teddy anyway.

Also as the owner can set the parameters of the machine, it’d definitely be true that machines in some places might “pay out” a lot more than machines elsewhere.

The deleterious effects of bad business models, incentives and AI on music streaming services

When Spotify and its competitors launched it heralded a whole new way to listen to music. Suddenly everyone who could afford a relatively cheap monthly subscription and a device to use it on had access to a near infinite list of the world’s most loved music. Recommendation algorithms would expose them to new music they’re very likely to love. At the same time, it’d cut down on the pull that led some people to pirate their music, so that artists would get paid more fully. Smaller artists who didn’t have the means to promote or produce huge worldwide release could get an audience beyond their local pub. What’s not to love?

Inevitably, plenty. Naturally it’s not only extremely bad AI-written books that are attempting to flood our media channels, and occasionally charts, with dross. Music provides another example, for many of the same reasons.

There’s the AI side of things - both the generative AIs that can “create” music. But also the more traditional, less hyped, AIs that are used to engage the user with the apps concerned. And I suppose the AIs that listen to the music, more on that later.

But there’s more to be concerned about, including bad business models, bad incentives, and the exploitation that inevitably shows up once a few giant middle-men effectively monopolise the relationship between people that produce things of value and those that wish to consume them to the extent that they’re incentives are entirely distinct from either party.

I’m an increasing fan of Marshall McLuhan’s idea that “the medium is the message”. Sure, songs are still songs, but “song” is vague enough that there’s plenty of room for the medium of distribution to influence its content. Perhaps the delivery method is also the message to some extent. His colleague, John Culkin, opined that “we shape our tools, and thereafter they shape us”. Our music-creating, music-promoting and music-listening tools have certainly changed in recent times.

Torrents of new “music” flood the system each day

Apparently there are 120,000 new music tracks being uploaded to music streaming services such as Spotify each day. Each day. MBW calculates that at that rate this year we’ll have added 43 million new tracks to the archives - substantially more than even the estimated 34 million tracks uploaded last year.

(An ISRC is an International Standard Recording Code, used to uniquely identify sounds and music recordings.)

Unsurprisingly, not all of these are new Taylor Swift songs, Beatle re-releases or even exciting new work from up-and-coming new indie bands. Some of it is literally noise; white noise, fan noise, rain noise, other noises are available. More of it is a near infinite of variety of customised “Happy Birthday to X” songs with X replaced by the names of most people you might imagine sending well-wishes to.

And an increasing amount of it is AI generated. And not even 2023-level-AI generated:

On Universal Music Group’s Q1 earnings call, Sir Lucian Grainge criticized what he calls the “content oversupply” that currently sees around 100,000+ tracks distributed to music streaming services each day.

“Not many people realize that AI has already been a major contributor to this content oversupply,” said Grainge.

“Most of this AI content on DSPs comes from the prior generation of AI, a technology that is not trained on copyrighted IP and that produces very poor quality output with virtually no consumer appeal.”

The incentive structure encourages people to game the system

But, similar to the way Kindle Unlimited pays its authors, it’s not really the quality of the track that matters. It’s whether it was played a lot or not. It’s the word that we’ve perhaps all learned to dread a little: “engagement”. 1000 people playing a minute of white noise often produces the same revenue for the creator as 100 people listening to a 10-track album of Taylor’s finest.

As the CEO of Warner Music puts it:

It can’t be that an Ed Sheeran stream is worth exactly the same as a stream of rain falling on the roof.

I note he doesn’t specify whether an Ed Sheeran stream is worth more or less than a recording of the rain, but either way you can probably think of some track of music you personally feel confident has a great deal more value to you and perhaps even to the world at large than listening to one of a million varieties of rain, a good proportion of which might actually 2010-era auto-generated AI fake rain noise.

So where do we end up? With the same dynamics as we had for those terrible AI-written Kindle books.

….the inevitable coming together of the generative AI infocalypse, enshittified subscription service monopolies, misguided algorithmic recommendation systems and the incentives of modern-day capitalism

This of course leads “creators” to game the system. Which is why every time I try to listen to something popular like “Flowers” I have to be quick on the pause button if I don’t want to end up listening to an algorithmically generated rotation of various knockoff variations of the original. Or, to be honest, one reason why for the most part I relieve the pre-Spotify era and almost always listen to music in units of albums, rather songs or playlists. As a sidenote, if this too is your preference and you’re an Apple Music user then I heartily recommend the iOS Albums app which near-forces you into that mode of operation.

Not all of the system gaming is AI adjacent of course. Humans are more than capable of coming up with their own quantity-increasing quality-reducing schemes. Six years ago, Vulture wrote up some of those shenanigans for us to peruse.

There are the aforementioned people trying to trick you into listening to their knockoff version of whatever the current genuine-hit song is. It might be a cover, or it might be something totally different but with a similar sounding name.

…the coverbots and ripoff artists who vomit out inferior versions of popular songs every week, flooding the website with dreck that only succeeds when users are misled. No one would willingly listen to King Stitch’s “Sit Down, Be Humble,” a third-rate cover of Lamar’s original, but the track has been streamed more than 300,000 times thanks to Spotify’s broad search results and a clever title designed to confuse those who don’t know the song’s real name.

Then there’s entries that are not a malicious deception. They’re just not really the same category of thing as the classic idea of an album, even though they are displayed indistinguishably.

Some artists, a term used very loosely here, are providing people exactly what they want. It just so happens that what they want is ephemeral nonsense. Take, for example, the artist Happy Birthday Library, whose Spotify catalogue consists of hundreds of personalized versions of “Happy Birthday” streamed more than a million times.

The algorithms and incentives naturally have an effect on the actually talented and occasionally mega-famous folks too, reshaping what they produce and share. Vulture, and others, suggests that the reason Chris Brown released a 40 track album is because that’s the way to boost your streaming numbers the most. Remember, you get paid per stream of each track - not each album. This huge-track count trick can naturally be used by lesser-known folk too

As they conclude, albeit perhaps with rather rose-tinted spectacles on:

Never before has a song title or artist name been more important than the actual songs themselves. … Never before have so many songs existed just so an album can have a 20th, 30th, or 40th track

Overall they list 11 ways they’ve noticed artists adapting to Spotify’s incentives:

  • The Happy Birthday gimmick mentioned above. Who wouldn’t want to listen to a song entitled “Happy Birthday whatever-your-name-is”?
  • Writing songs about everything someone might conceivably search for; their example being “the song “Have You Been to Radnor Township?” on the 93-track album Pennsylvania Songs: PA Exiting”.
  • Releasing multiple albums with the same songs, often titled differently.
  • Release an absolutely silent “song”.
  • Convincing people to get their computers to fake-listen to the song.
  • “Filling the void” - releasing covers of songs from famous artists who aren’t on Spotify.
  • Trying to trick people by naming their songs very similarly to more famous ones. Someone searching for Imagine Dragon’s song “Demons” might not realise they accidentally clicked on the 1-hit wonder Imagine Demon’s song “Demons”.
  • Covering everything famous, all the time.
  • Adding unnessecary extra tracks to albums as noted above.
  • Putting your past popular singles in your new album.
  • Seeding popular Spotify playlists with fake artists - something Spotify has been accused of doing, more on this below.

OK, so is it all greedy users exploiting the opportunities the likes of Spotify inadvertently make available to them? No, it is not. It’s easy to go too far in absolving the platforms for their responsibility for this.

For one, it’s 2023. We know how these things go. Everything product decision is a choice. The big platforms decide where to prioritise their development. They decide what they accept and how. If some of the above isn’t what the platforms want to promote then surely if users can find it easily enough to repeatedly write articles about the issue then so could they, given their access to reams of confidential internal data.

In fact they do, sometimes.

Earlier this year Spotify removed thousands of AI generated songs from its catalogue. Many were those autogenerated by Boomy - a site that sells itself as enabling you to “Create original songs in seconds, even if you’ve never made music before” and " Submit your songs to streaming platforms and get paid when people listen".

Now as with many generative AI tools, it may not be a bad thing in principle. It might enable people who otherwise wouldn’t have the opportunity to create their own music be able to, in the same way as the image generators could allow expressions of novel creativity otherwise unavailable to most. But it’s also catnip for folk that care more for income than art.

Robots are creating music for robots to listen to

So why did Spotify go on a deleting rampage of these tracks? Not particularly because they weren’t a great experience for the average user. And there’s certainly no prohibition on using AI to write or help write one’s music.

Rather it was because the people listening to that music were fake. Remember that Spotify pays out based on how many people listen to your tracks. So “creators” built bots that “listened” to the tracks again and again. It’s essentially fraud, given the terms of service.

So in a somewhat dystopian-feeling loop, people are using AI to create music that is mostly listened to by other AIs, no humans involved.

The business model and recommendation algorithms of the mainstream music streaming platforms - quantity over quality, quantity over everything - makes them as vulnerable to clickfarms as any other type of site that provides some kind of reward for attracting lots of traffic of any kind. In this case Spotify has to pay for it directly, which is finally enough of a reason for them to act on it.

People are trying to measure these fake listeners. A French government organisation reported that 1-3% of all French streams in 2021 were fraudulent based on data from Spotify, Deezer and Qobuz. This is likely an underestimate due to the use of VPNs and the like.

Beatdapp, a company who sells a tool that looks for streaming fraud so is admittedly incentivised to claim that there’s a lot of it, wrote that at least 10% of all streaming activity in 2022 was fraudulent. That’s equivalent to $2 billion of revenue illegitimately extracted. They believe that it’ll be worth $7.5 billion by 2030 if it is not stopped.

The fact that more streams = more money has also been more publicly exploited in the past. Eternify was a service that lets users input the name of their favourite artists and it’d play their songs on repeat for exactly 30 seconds each on Spotify. 30 seconds was the threshold at which the artist gets paid, so this was a way you could optimise the income of your preferred musicians without the obstacle of having to actually listen to their output. The site would slowly count up the revenue your mouse-click had accrued for them.

It seems to be have been more of a protest statement from some understandably disenchanted musicians than something that aimed to enable fraudsters. And it probably wouldn’t have saved many starving artists without a truly mass takeup. Forbes calculated that using it for a solid day would generate perhaps $1.44 for the artist in question. In any case, it cost Spotify directly enough to make them go to the effort of shutting it down.

The “pay” most musicians receive from the service is incredibly low

The $1.44 a day figure reveals another not-so-hidden truth. These services are really not paying out a lot of money to the creators of what they’re selling access to. So much so that even a committee of British MPs - a group not particularly known for its sympathy or care for artists - are worried about it.

From all that revenue you’re providing to the streaming services, the folk involved in producing the music are receiving about 16% of it.

Spotify is thought to be paying out between £0.002 and £0.0038 every time you play a stream. Apple Music pays a bit more, a whopping £0.0059. Youtube a bit less, about £0.00052.

And “pays out” is doing a lot of work there, as even many of the reports about “Spotify only pays artists £x” seem to miss. That 0.2 pence has got to be divided up between anyone with interests or rights to the music, as well as each member of the band, the composer, the writers, and so on.

The BBC provides an infographic to help us visualise what that means.

A diagram showing where the revenue paid out from streaming goes

I don’t think that the recording industry was particularly renowned for being favourable to artists even before the streaming revolution. Anyway, all in all, the recording artist(s) often might expect to receive about 13% of that 0.2p. Under a third of a penny.

Thirds of pennies don’t add up fast. A survey by the Ivors Academy and Musician’s Union suggested that over 80% of musicians made less than £200 a year from streaming during 2019. 93% didn’t break the £1000 milestone.

Sure, the megastars might earn substantially more than enough to survive on, but even their entourage aren’t necessarily doing well. The writer of one of the songs on Kylie Minogue’s number-one charting album “Disco” reported receiving £100.

“Right now, hit songwriters are driving Ubers,” she told MPs. “It’s quite shameful.”

Spotify seems to be gaming its own platform

Back to the fake streams. As wrong as it is, let’s not feel too sorry for some of the platforms subject to this streaming fraud. Spotify’s seeming attempts to game it’s own system at the expense of other artists and their own users have been written about elsewhere.

There’s the time it was found to be burying songs from musicians because they’d also made deals with their rivals. Maybe that’s understandable, maybe not, but I’d guess most users don’t understand that to be what’s going on when they search for a song.

What’s even less obvious is that sometimes Spotify decides to pay producers a flat fee to create tracks such that Spotify owns the copyright. They then upload them to their own serivce under fake artist names, people that don’t exist. Why?

From Music Business Worldwide:

Why would Spotify be instructing producers to write and record tracks of this nature?

Bingo: to appear on some of its relaxing first-party playlists, which boast millions of followers between them.

Now they don’t seem to get anything like as many plays as the most popular songs on the platform so it’s not something that’s necessarily dominating the user experience. But each time users plays a track from one of these fake artists Spotify only has to pay itself the stream royalties. Over 50 fake artists of this kind have been found now, almost all of which have had millions of streams.

As ever, it’s not entirely a black and white issue. Spotify have explicitly denied some (but perhaps not all) of these allegations. And, in this case, behind the fake artists are real people playing real music. They’re just paid in a way much more favourable to Spotify at scale than other artists, and they seem to have an automatic route to being included in much-coveted very popular playlists.

Spotify curated playlists apparently account for around a third what’s listened to on the platform in terms of time. To get onto one of them, most musicians have to actively pitch their music for inclusion, one song at a time, before the song has been released. Most of the pitches, around 80%, are turned down without explanation.

Arguably, similar dynamics exist in other types of streaming services. Netflix does of course make its own shows which it heavily promotes on its own service. But as a user it does feel disingenuous to me at least. Why not label these tracks as having been produced by Spotify? Why not explain how they got onto the playlists?

Based on some patent applications, some folk seem to think that eventually Spotify might take the human out of the creation process altogether, potentially filling its playlists with music they generated themselves via AI. Presumably competing against those independent users uploading their own AI music, albeit with potentially a thumb on the scale.

Enshittification, yet again

Platforms such as Spotify appear to potentially provide great examples of Doctorow’s theory of enshittification. When they first launched they appeared provided incredible value for users, revolutionary in some ways. You could access literally millions of much-sought after music for a relatively low monthly cost - or even free on Spotify at least if you were happy to tolerate adverts - entirely on demand. A radio station perfectly attuned to your minute-by-minute preferences. A way to experience far more of the musical art than you could ever afford to if you were paying £12.99+ for an individual album like ye olde days. An experience much safer and easier, not to mention less criminal and more conducive to rewarding musicians than yesteryear’s internet underground piracy scene.

As The New Yorker wrote about Spotify 5 years ago:

The company is essentially insisting that freer and easier access to music is the only thing that matters; everything should be available to everyone, because freedom of choice is an essential freedom.

The platforms attracted so many listeners with this fabulous promise that they became attractive, perhaps essential, to artists. Even many of the very rich very famous names who have sworn off Spotify at some point in time came back, presumably for the audience and its impact on the charts. If you weren’t popular on Spotify then you’d not so much hope of ever topping the bill, attracting the most revenue, being the most famous.

Or maybe they’d even enable vast numbers of people to realise their artistic dreams in a way they’d otherwise never have been able to. As they wrote 4 years ago:

“the goal of our marketplace strategy is to harness Spotify’s ability to drive discovery to connect artists with fans on a scale that has never before existed with the goal of enabling 1 million artists to live off of their work.”

But now, in the final stage, the platforms claw back more and more of the value for themselves.

They stop gaining marketshare only by virtue of being able to sustain periods of loss. They change the user interface longer than their competitors. What’s promoted, when and how becomes something less aligned with the reasons that many users may have originally signed up. They try to make it inconvenient for either users or creators to leave, and costly for artists to make deals with competitors. They acquire potential future competitors or expand into adjacent-ish markets to increase the perceived lock-in. They raise prices - although some believe that the recent near-universal streaming price rises in conjunction with a cost-of-living crisis is resulting in an increase in piracy. They alter recommendation algorithms, alter incentives.

All very logical from the point of view of shareholder value et al, at least in the short term. It’s not necessarily driven by maliciousness; I’m sure many of Spotify’s employees love music, love musicians, embrace that vision of enabling a million artists to live off their work. It’s just that at the end of the day the company’s own incentive is all about coming up with the financial arrangements that most benefit themselves in the marketplace, whether or not it’s at the expense of either or both sides of the user-creator relationship that’s the fundamental reason that they can exist in the first place.

The tools shape us

It’s more amorphous to define and subjective to evaluate, but perhaps these services have significantly changed how we tend to consume music beyond the technical side of things. McLuhan and Culkin’s quotes apply on the demand side too. The medium is the message. We shape our tools, and thereafter they shape us.

In days of yore, listeners would usually have to invest financially in a specific number of physical instantiations of music; a cassette tape, a CD or an LP album amongst other formats. In the iPod era this moved onto digital downloads, but still at the level of purchasing individual songs rather than access to a smorgasbord of (a somewhat misleading perception of) everything everywhere, all at once, all the time.

For most people this put a limit on how much music they had access to. This seems prima facie bad at first glance, and perhaps it is. But maybe it also encouraged folk to spend more time and intention in choosing what they listened to, more time exploring any one individual work. I expect a fair proportion of people don’t even know what they recently listened to when they’ve opened their app at the start of the day and hit the “especially for you” playlist.

When the Guardian talked to a bunch of people who’d moved from streaming music back to more traditional forms of consumption they found some examples of this. Here’s Lethem:

[With streaming] there’s endless accessibility, but you’re not really listening to anything. At least that’s what it started feeling like to me. I’m experiencing so much music, but am I really listening to any of it?

And if you don’t like what the algo chose then you can just hit skip 5 seconds in and be provided with something else that tries to hook you in. And it’ll continue forever. There’s no real concept of sitting down to appreciate an album end-to-end when the album is a literally endless playlist.

Many of these services do let you listen by album of course, at least if you pay them for that ability. I’d be curious to know how many people still do that. It’s certainly not the style of listening I feel like I see most heavily encouraged by the apps. Similar to the Youtube and Netflix style next-episode auto-play, the last thing they want to do is make you think you “finished” something to the extent you close the app.

Fairly incredibly, Spotify even used to default to shuffling the tracks of the album you selected into a random order before Adele requested they stopped doing that because, of course, there is usually a reason why the tracks were put in that order in the first place. To quote her:

We don’t create albums with so much care and thought into our track listing for no reason. Our art tells a story and our stories should be listened to as we intended.

Some more conceptual albums would surely simply make very little sense out of order.

So hey, it’s not like Spotify never listens to feedback, at least from megastars. Although even now you do have to be a paying customer of Spotify in order to be given the privilege of a “play in the proper order” button.

More recently Amazon actually went in the opposite direction for its basic music streaming service. Subscribers to Amazon Prime, a service that costs around £9 a month and offers a variety of benefits to Amazon customers including music streaming, found that instead of the previous behaviour of being to choose what they listen to, when they requested music by a particular artist they would get a single track from that artist followed by a load of tracks from other people. Even their manually customised playlists started to shuffle around.

An unhappy customer reports his confusion:

Mr. McKenna, who lives near Albany, N.Y., spent years curating Amazon Music playlists….Recently, the app began playing a rock song he had never heard. When he got to the office, he tried to click on a Kenny Chesney song in his playlist, but it didn’t work. Songs he didn’t recognize kept playing.

For Bob, his chronologically-defined playlists not any more being chronological was enough to drive him back to CDs and Windows Media Player - “It’s like I’ve gone back to 2005”. The only way to keep hold of your Amazon Music collection and listen to what you want to when you want to now became to upgrade to the more expensive “Amazon Unlimited” service.

Shifts in people’s listening choices are certainly not all about these services removing features that were once there. Even where you can in theory still choose to listen to exactly what you want, here’s Nick from the previously-mentioned Guardian article confirming a shift in his behaviour he experienced soon after switching to streaming that I suspect is quite common:

I found myself selecting more and more just one-off songs from an artist, whereas before I’d been inclined to save a whole album

The obvious “discovery” alternative 20 years ago was to listen to the radio or a TV channel like MTV where for the most part you are subjected to other people’s musical selections. From beginning to end, with no skip button. This might let you discover something so new and wonderful that you - and your 2023 recommendation algorithm - would never have considered if left to your own devices. On the other hand you had to listen to a lot of dross you didn’t like, and would perhaps never happen to be listening at the right time to hear songs you were more likely to enjoy - so again, nothing here is all good or all bad. But it’s worth thinking about.

Finlay, in the same Guardian article, spoke to the issue above:

With streaming, he says: “If I didn’t gel with an album or an artist’s work at first, I tended not to go back to it.” But he realised that a lot of his all-time favourite albums were ones that grew on him over time. “Streaming was actually contributing to some degree of dismissal of new music.”

Jared speaks to both upsides and downsides:

My musical experiences definitely feel more dedicated and focused. It’s not as convenient. I’ll reluctantly admit that I listen to less music. Although on Spotify, I wasn’t necessarily listening to stuff. I was checking out the first 15 seconds and hitting skip. Now, I have to work for it and I like that. I can use the internet as a search tool but I’m not using it as a means to listen. I really have to seek things out and research.

Virginia likewise, exhibiting a pattern I found myself adhering to at first:

I will listen to one song 100 times in a row, but I won’t give the rest of the album a chance. Before I used streaming services, I would listen to the whole thing.

Even conceptually, the word “streaming” has connotations that may affect how we feel about what we’re listening to.

Before there was streaming music, what else was streaming? This idea that you can just turn on a faucet, and out comes music. It’s something that leaves everyone to take it for granted.

On the other hand, many of these ex-streamer folk talked about the extra work they now had to do to find what they like. It can comes at a financial cost too, which risks making some of what I wrote seem elitist and gatekeepery as well as Ludditey. After all why should only people with time and wealth to spare have access to reasonable quantities of music? What’s wrong with having convenient and cheap access to so much art? Nothing on the face of it, if we can find a business model that doesn’t incentivise its enshittification and abuse of the creators.

Streaming isn’t necessarily cheap, and you end up owning nothing

Although I would note that for the moderately well-off that it’s not clear to me that paying for a year of Spotify is necessarily cheaper than buying a few albums that mean something to you. Right now, Spotify paid edition would cost you £131.88 a year. That would also buy you a good amount of singles, albums, whatever you’re into. That might well suit people who prefer to or simply happen to end up listening to a certain subset of everything repeatedly. And it’s almost certainly better for the artists concerned.

It also means your music becomes “yours”. You can keep a CD or, let’s be realistic given it’s 2023, an MP3 or lossless music file whether or not you want to keep subscribing to any given service. I’ve been an Apple Music subscriber in my time, and I liked the service. But assuming one day you will choose to, or be forced to, stop paying the company the monthly fee then it’s all gone. If I want to switch to a different service then it’s a faff. Copying your favourite playlists, likes, subscribes, downloads et al to a new place isn’t necessarily easy - the new service may even be missing some of your tracks. And moving your carefully refined recommendation algorithms across services isn’t on anyone’s radar at all and probably won’t ever be; it’s each company’s secret sauce.

You also certainly can’t give your collection or pieces of it to anyone. You can’t give what you don’t own. I can’t find it now, but I remember reading a story about the joy and meaning someone got from inheriting the music collection of their sadly recently-deceased father; the feelings and memories that listening to what he’d collected over a lifetime evoked. But that’s something that likely becomes unavailable with a shift to streaming.

I couldn’t find any documentation about what happens to someone’s music streaming account when they die, outside of several posts on the Spotify forums that mainly end in “try asking customer services”. But it’s surely not something that someone inherits> If you don’t know the person’s password and can open to their email then access may be gone for good. If it’s a pay-for service then I imagine every trace of it simply vanishes as soon as the credit card fails to bill.

Talking about dying is admittedly a little dramatic. But there are other ways your favoured streaming services might be denied to you even whilst alive. Services close, even ones from hugely successful companies

Say goodbye to Google Play Music, the service will be going away for good in December 2020 and with it, all your music if you don’t transfer it over.

Are you old enough to remember Rdio? Beats Music? Yahoo Music Unlimited? Or any of the other 20-30 services currently listed on Wikipedia in its “Discontinued services” section, either entirely shut down or purchased by and replaced with the current handful of big players.

Or you may simply lose access to your account. If Apple, Amazon, Spotify, or whoever you’re with decide for some reason that they don’t want you to have an account then it’s all gone. The various chatboards of the internet are full of stories of people losing their accounts from the big tech firms. Oftentimes they’re more concerned about their emails or photos than their music. But it’s all the same. If Apple, Spotify or the rest of them decide you’ve been misbehaving, or can’t be convinced that you are who you say you are, then whatever effort you’ve put into maintaining your collection is lost. I’d imagine it’s probably rarer for this to happen with music streaming services than some categories - it’s harder to imagine being accused of “abusing” a music streaming account but as many Reddit posts show, it’s not impossible.

For the paid-only services, which is most of them, of course the same happens if your circumstances change such that you can no longer afford the service. And even if not, you’re bound to implicitly supporting any changes in the streaming company’s rules, strategies, personnel or ethics if you want to keep the music flowing. This might well be a price worth paying for many people, for most people. But it is a price.

If one of these eventualities does happen to you, then there are third-party services that try attempt to transfer your streaming collections and various metadata between one such service and another, although I haven’t tried any of them out enough to know how well they work. But not if you already lost access to the service you want to move from. And of course they won’t be able to transfer your playlists, likes et al to a service that doesn’t have the particular track involved.

This interactive visualisation is a great way of explaining or visualising what a given Cohen’s d effect size actually means in the real world.

Enter the d in question, or by clicking the cogwheel you can enter means and standard deviations, and the slightly-arcane-seeming resulting ’d' value is translated into an explanation rather easier to get one’s head around.

For example:

With a Cohen’s d of 0.80, 78.8% of the “treatment” group will be above the mean of the “control” group (Cohen’s U3), 68.9% of the two groups will overlap, and there is a 71.4% chance that a person picked at random from the “treatment” group will have a higher score than a person picked at random from the “control” group (probability of superiority).

Moreover, in order to have one more favorable outcome in the “treatment” group compared to the “control” group, we need to treat 3.5 people on average. This means that if there are 100 people in each group, and we assume that 20 people have favorable outcomes in the “control” group, then 20 + 28.3 people in the “treatment” group will have favorable outcomes.

Expect lots of fun new pun-ridden URLs to show up soon. Starting December 5th everyone can purchase fancy new domain names ending in .ing and .meme.

Braverman thinks homelessness is a "lifestyle choice"

Our Home Secretary, Suella Braverman, tell us that the homelessness crisis in Britain is in fact people making a “lifestyle choice”. Especially those pesky foreigners.

Even her fellow Conservative, Bob Blackman, who heads up a group that aims to tackle homelessness thinks this is a phrase too far and believes she should use “wiser words”.

Amnesty has it right - claiming that there’s queues of people just longing to live the decadent luxury of shelter-free lifestyle is nothing but a “very convenient cop-out” for a Government whose repeated failures do nothing but exacerbate the crisis.

In reality the unhoused “lifestyle” is not so glamorous. A group of interested charities put together a letter better explaining the experience:

People sleeping rough frequently experience violence and abuse. The impact on their physical and mental health is significant. The average age of death for people experiencing homelessness is just 45 for men and 43 for women. This is not a life people choose.

The homelessness issue, particularly the more extreme cases where people find themselves forced to sleep on the streets, is especially frustrating because it’s also entirely unnecessary. We know how to solve it if we really wanted to.

Because we already did, back in early Covid times when the UK chose to find a way to house the vast majority of rough sleepers. Yes, under a Conservative government! Some of those provided shelter even gained access to assistance with things like benefit applications or medical treatments.

But apparently all decent and humane things have to come to an end. In July 2020 the funding was withdrawn and governmental responsibility for helping these people in need shirked - even whilst Covid was of course still very much circulating around.

Back to contemporary times where not only are we supposed to believe that living on the streets is universally nothing but a lifestyle choice, but that the largest threat to the fabric of British society is…tents.

So much so that the Home Secretary is seemingly trying to slot in a new policy that will fine any charity that gives a tent to a person without shelter. This seems to be a step too far for even many of the current Conservative party, so tbd on whether it makes it into the Kings Speech or not.

Either way, with ideas like this a certain contingent of today’s Conservative party does seem to have developed some real “The Cruelty Is The Point” vibes.

Ooh, after the success of the Super Mario Bros Movie they’re making a live-action Legend of Zelda film.

TIL: The collective noun for a group of ferrets is a “business” of ferrets.

A black-footed ferret

So now you know the easiest way to become a business owner.

Jennifer Daniel shares what we can expect for official new emojis in 2025.

Just seven this time, apparently selected for their linguistic flexibility, amongst other things.

  • Face with bags under eyes
  • Fingerprint
  • Root vegetable
  • Leafless tree
  • Harp
  • Shovel
  • Splatter
New emojis for 2025

Leafless tree could apparently be used to represent a “state of barrenness and death” which feels very 2023 at least. Face with bags under eyes is surely another that contemporary times will inspire great use of. Unless of course by 2025 we either managed to fix the world or destroy it enough that all that remains of us is splatter .

The UK does not need to widen the definition of extremism even further

Via some leaked documents the Observer got their hands on, we learn that the government is thinking about widening the formal definition of ‘extremism’ such that it’ll read:

Extremism is the promotion or advancement of any ideology which aims to overturn or undermine the UK’s system of parliamentary democracy, its institutions and values.

You don’t even have to be doing ‘extremist things’ yourself. You might just not be disapproving of the folk that do hard enough.

Part of the definition that this Brave New World of people who spend their time sitting in other people’s cars talking about freedom and their friends will be considering, if the leak doesn’t kill it, also includes:

Sustained support for, or continued uncritical association with organisations or individuals who are exhibiting extremist behaviours

I’ve a slight worry that this blog, perhaps even this post, will become technically ripe for a referral to the anti-extremism authorities.

I mean, with the criteria of undermining ‘the UK’s system of parliamentary democracy’ featuring up top, might we be in theoretical trouble for having cast a vote in favour of changing the electoral system we vote under in the 2011 national referendum? What remains of the Lib Dems better watch out!

The various civil rights groups we have are rightly rather unhappy about this.

From the director of Liberty:

This proposed change would be a reckless and cynical move, threatening to significantly suppress freedom of expression

The editor of Index on Censorship:

This is an unwarranted attack on freedom of expression and would potentially criminalise every student radical and revolutionary dissident.

The racial justice director of Amnesty International:

The definition of extremism and its usage in counter-terrorism policies like [counter-terrorism strategy] Prevent is already being applied so broadly it seeks to effectively hinder people from organising and mobilising. The proposed definition takes this even further and could criminalise any dissent.

The leaked Govenrment document lists some organisations that they’d consider as being ‘captured’ by this proposed new definition. These include Muslim Council of Britain, Palestine Action and Muslim Engagement and Development.

It’s not like the current definition of extremism we have is particularly narrow.

…active opposition to fundamental British values, including democracy, the rule of law, individual liberty and the mutual respect and tolerance of different faiths and beliefs

But I suppose at least you have to be ‘actively opposed’ at present. And, judging by the fact that more of our politicians aren’t in extremism jail, it seems like at least the latter part of the definition isn’t particularly strongly enforced.

Not that it’s stopped at least 45 peaceful environmental activists being referred to the Government’s current anti-extremism program. That situation that provoked Liberty to speak out once again at the time:

This reinforces long-held concerns that the government’s staggeringly broad definition of extremism enables the police to characterise non-violent political activity as a threat, and monitor and control any community they wish.

As ever, even if for some absolutely inexplicable reason you trust the current government and other arms of the state not to abuse the extraordinarily wide definitions of these very hot-button, very political, very emotional topics, one must remember that things change. Some even worse folk could come along in the future. And it’d be nice if they didn’t automatically inherit the power to silence or punish anything that could be vaguely categorised as dissent against their preferences.