The Braindump Blog

Amazon Kindle provides a popular, easy-to-use, and (sometimes) cheap ecosystem of devices and books for anyone who enjoys reading eBooks. Of course they are also not the most ethical of companies which might make the proposition entirely unattractive to you - which is not a topic I’ll deal with here. However their monopolistic ways are such that many folk end up with a stash of Kindle ebooks they bought.

One problem though is that Kindle books are (usually) encrypted with DRM. This means that whilst they open smoothly on Kindle devices, Kindle apps and the Kindle Reader website they won’t work anywhere else. You can’t switch to a different ereader device, a different app or read them on an operating system that doesn’t have the app (except perhaps via the web).

Furthermore presumably if you lose access to your Amazon account then you’ve also lost all your books. That’s presumably unlikely to happen most of people, but there are plenty of stories on the internet of people this has apparently happened to.

And it’s not absolutely unheard of for Amazon to magically make a book you thought you purchased vanish. Famously, people woke up one day to find that the copy of George Orwell’s book ‘1984’ of all things - you know, the one about how the government retrospectively rewrites history by changing the digital records of what happened - had vanished, from the Amazon store and their own Kindles. They’re also able to change the content of the books after you purchased them - in theory for good, but each individual case might depend on your opinion. So all in all, whether you’re switching apps or devices or simply trying to make a safe backup, you’re going to end up with a bunch of unusable files if you try to open them in something not officially created by Amazon. I don’t love this, to say the least.

But it turns out that the wonderful internet people have figured out how to remove the DRM from Amazon eBooks. Once the DRM is gone you can open it on any device or app that supports the file format, or use a tool such as Calibre to convert them to other formats that your chosen reading setup can work with.

Obviously if removing DRM is illegal where you live - and to be honest it probably is, but hey, you never know - then you shouldn’t do it. But once you moved to a country where it is permitted then the below is - purely in theory of course - one of the simpler ways you might do it.

One method to do this is via the aforementioned Calibre ebook management software which is free, open source and available for Windows, MacOS and Linux operating systems and one of its plugins. Calibre is actually well worth a look if you’re into ebooks even if you aren’t stuck with a load of DRM-laden Kindle files.

Below I’m going to outline what I believe to have been the most reliable approach over time.

There are two downsides to this approach though:

Firstly you have to have a hardware Kindle eInk device currently registered to your Amazon account. It shouldn’t matter which model as long as its one of the eInk ones and not one of their Android ‘Fire’ style tablets. I have heard of people buying an old and cheap ereader from eBay just to connect it up to their Amazon account to enable the below, even if they have no intention of using the device itself.

It doesn’t actually matter if your Kindle is later broken or lost; the key is that it must currently be registered to your Amazon account.

Secondly you have to download your books from the Amazon website to your computer one at a time as far as I can tell. This will be painful if you have a lot of books.

If it’s unbearable then I know some people have had success using the Kindle apps for Windows and Macs to download books en masse. But the encryption Amazon uses in that case is different than that used when you are using a Kindle eInk eReader. It seems to change often enough that depending on exactly what and when you try it you might come up against problems with the actual DRM removal.

That said, if you want to try that way you’ll still need the plugin mentioned below - but also I’d suggest you search for another guide as there’s more to it than what I’m outlining here. The Calibre subreddit is perhaps one of the best places to find this stuff, but there are plenty of possibilities.

But if you can cope with that, then:

• First make sure you have Calibre installed.
• Then download the DeDRM_tools plugin. I think the newest version is from the noDRM repo; this one. Download the zip file at the bottom of the page ( DeDRM_tools_10.0.9.zip at the time of writing).
• Extract the files from the zip somewhere onto your computer.

Then load Calibre and do the following inside it:

• Preferences -> Advanced -> Plugins -> Load plugin from file.
• Choose to load the DeDRM_plugin.zip file you extracted from the zip above.
• Reboot Calibre (maybe).
• Preferences -> Advanced -> Plugins -> DeDrm -> Customize Plugin -> Kindle eInk ebooks
• Add the serial number of your Kindle, which is available either somewhere in your Kindle settings or on the Amazon website if you go to the page called “Manage your content and devices” and enter the devices section. Now you can see why you need an eInk Kindle to have been connected.

Now back the Amazon website page “Manage your content and devices”:

• Go to Content and then Books and find the book you want.
• Under “More actions” select to download and transfer via USB.
• Pick the device that you added the serial number of when it asks you which devices. This will download a file, currently ending in azw3.

Lastly, Inside Calibre, import the azw3 book as usual via Add Books menu. And there we go, the book should hopefully open just fine in Calibre and be freely available for reading, conversion and all the other good stuff Calibre does.

🎶 Listening to Kaiser Chief’s Easy Eighth Album by The Kaiser Chiefs.

Here’s another entry from the “sort of surprised they’re still going” hall of fame.

Unlike the last album I mentioned, 19(!) years after this bunch were famously Predicting A Riot, their latest album sees them taking a new and rather more calming path. Here they turn their hand to the pop-funk genre, with mere hints of their earlier style now and then.

📚 Finished reading The Physics of Sorrow by Georgi Gospodinov.

This is another wonderful, mind-bending, book from Georgi Gospodinov, who wrote the “Time Shelter”, a book I greatly enjoyed earlier in the year. This is somewhat peculiar story, but written in a way that feels very suited to my brain; somehow calming even in its melancholy.

Presumably it’s not only my brain that it’s suited to - it became an instant best-seller in its country of publication, Bulgaria, when it was released around 2011. So it was written much before Time Shelter, but has only been released in the UK relatively recently.

As the FT notes, The Economist once referred to Gospodinov’s homeland Bulgaria as a particularly sad place.

Although richer countries are clearly happier, the correlation is not perfect, which suggests that other, presumably cultural, factors are at work…the saddest place in the world, relative to its income per person, is Bulgaria.

It’s somewhat an autobiography - the main character has the author’s name and grows up in the country of the author. The progression through his life is clear. But it’s also fantastical, whimsically and seriously so, as we follow his life - and those of the other minds he inhabits - over time. See, as a child, the book tells us that this Georgi suffers from “obsessive empathetic-somatic syndrome” which essentially allows him to get trapped in and compelled to live out other people’s memories. When in one of these episodes, “I” in this case would not refer to Georgi, but rather his host, for want of a better word.

The story is fragmented, moving from one scene to another and back again, from his point of view to someone else’s, obsessing for a while of some old list he found and then ruminating on what it feels like to be something else, or retelling a snippet of someone’s life story. Time fast-forwards, time rewinds. It’s labyrthinic, and deliberately so. Very much like the human mind, of course.

Gaustine, of Time Shelter fame, gets a shoutout. References to Borges, the author of some of my favourite short stories and similarly a fan of narrative labyrinths are made.

As time passes, story-book Georgi grows out of the afore-mentioned childhood syndrome, and no longer inadvertently leaps into other people’s minds. Instead he becomes a collector of everything; every note, ticket, picture, menus, whatever - boxes of the stuff. But primarily a collector of stories. He’ll pay for a good one.

One of the themes he constantly returns to is that of abandonment, and those oft-related emotions of isolation and loneliness.

Interwoven through the experiences of “normal” people he comes back again and again to Greek mythology, specifically the story of the Minotaur. But in this retelling the minotaur is no vicious beast. He’s just a sad, lonely child who through no fault of his own got born with an alien face.

There is sorrow in him which no animal possesses.

Poor Minotaur’s father hated him. His mother abandoned him to a dark, depressing, lonely labyrinth. At least in this novel you do get to hear the Minotaur have his day in court at one point.

This is the lens which much of the presumably more autobiographical elements and aspects of social commentary are shown through. The claim is made that the “history of the world” can be described through a series of children being abandoned.

📺 Watched season 13 of Death In Paradise.

More (mostly) light-hearted murder in the sun. I guessed wrong about who Sunset Chaser was. But still, good to know that getting that a comment on otherwise very boring blog can indeed change your life.

Last season.

The results aren’t published yet, but it looks like we’re going to see that the anti-obesity GLP-1 drugs such as Zepbound are potentially useful for treating sleep apnea too.

I’m curious how much of that comes from the associated weight loss - there is a known relationship between e.g. sleep apnea and obesity, albeit potentially a fairly complicated one - vs if there’s a more direct pathway in addition.

📚 Finished reading Becoming a Data Head by Alex J Gutman and Jordan Goldmeier.

This book is a nice mix of layperson and a certain amount of technical content, all on the subject of using data. It’s primarily focused on the uses of data in the workplace.

It’s not going to teach you the intricacies of statistics - although it does introduce some useful concepts it’d be helpful to understand just to navigate one’s way through everyday life - nor how to construct cutting-edge AI models. But you probably will come away understanding more about how those things work and their limitations. As well as, probably more importantly for the target audience of this book - business people, executives, engineers or novice/wannabe data scientists - when and why you should consider using these kind of techniques, or, if you’re managing people or processes, when to permit someone use them on your behalf.

It’s thus frequently more about the mindset of data work than the technical details of it. When should you use data? What makes a project worth embarking on? When should you abandon it? And how should you react when folk present the results of their data projects to you?

The book is organised in to 4 sections, which give an idea of the trajectory readers will be following:

1. Thinking like a Data Head
2. Speaking like a Data Head
3. Understanding the Data Scientist’s toolbox
4. Ensuring success

The last one there is mostly referring to success of the data projects you or your colleagues may embark on.

So what then is a “Data Head”, that thing they want you to become?

…these are data skeptics, although the skepticism is based on employing their data critical-thinking skills rather than just to be annoying. They advocate for data where it’s useful, but question what ought to be questioned. Their skepticism comes from having technical knowledge and domain expertise. It’s delivered with empathy.

Their very reasonable claim is that using data for the sake of using data is not a great use of anyone’s resources, no matter what the hype says. Rather:

To be an effective Data Head you must use data to drive change.

And some things simply aren’t possible no matter how much you want them to be. Let’s never forget the famous phrase:

Garbage in, garbage out

This, according to the book, is especially the case with some of the fancier data techniques - if you are not a company with the unbelievable amounts of data, skilled personnel and other resources that say Google has, be more than a little cautious about believing that you can produce what they produce.

But almost certainly there are ways you and your organisation can productively use data. After all, the book proclaims that data may be the most important part of your job, whether or not you want it to be. You just need to distinguish productive avenues of inquiry from red herrings.

The book is impressively concise, and a pretty easy read if you have any familiarity with the topic. Along the way there are plenty of checklists and lists of good questions to ask at all phases of any data project, from conception to conclusion.

I can see these being useful for more knowledgeable data workers in addition to the target audience - the sort of people who already know their field well in terms of technical details, but also in some cases may risk getting carried away with their urge to spend time and money having fun with the latest and greatest algorithms on a project that never really stood a chance of working in the first place.

As an example, before starting on a data project the authors implore you to ask:

• Why is this problem important? If your answer focuses on the methodology or the deliverables involved them you’re probably missing the point of the question.
• Who does this problem affect? Can they actually use the results you’ll get? How will their job change?
• What if we don’t have the right data?
• When is the project over?
• What if we don’t like the results? Meaning that the project went well, but produced an answer the stakeholders didn’t want.

The last point is one I have often wished more folk consider than apparently do

Overall, the book helps you ask better questions, both of yourself and of others. There’s also a handy list of pitfalls later on in the book that are a good reference for checking your own work and (politely) challenging that of others.

A couple of weeks ago Apple started allowing emulators onto its app store. Not a scene I’ve dabbled in before but once the free “Delta” mostly-Nintendo emulator starting taking over the charts I had to give it a go. It works great. Supposedly it’ll work with games from any of:

• NES
• SNES
• Game Boy
• Game Boy Color
• Game Boy Advance
• Nintendo DS
• Nintendo 64
• Sega Genesis / Megadrive

Me, a saint, has obviously been magically backing up my purchased Nintendo game cartridges since I was a baby (to be fair devices exist that can help do that sort of thing). But for anyone who has temporarily lost their local backups, searching for “roms” reveals a whole new world of archived games I didn’t really know existed.

Finally I can experience the joy of Mr Nutz once more, this time in mobile form.

I’m two years later to the party, but I believe this is the source of the famous “AI experts believe there’s a 10% chance that it will destroy humanity” statistic that was the subject of a certain amount of understandable-on-the-surface media frenzy a while back (e.g. here or here).

The exact question asked was:

“What probability do you put on human inability to control future advanced AI systems causing human extinction or similarly permanent and severe disempowerment of the human species?”

And the median answer was 10%.

The responses are from some subset of 738 questionnaire responses that came from a group consisting of researchers that published something at a couple machine learning related conferences. This represents a 17% response rate from the 4271 people solicited. It sounds like this particular question was given to a smaller subset of the 738 because recipients could get different questionnaires depending on whether they’d already completed another assessment in the past.

Perhaps not unrelated, 69% of respondents think that AI safety research should be given a higher priority than it was at the time.

I wonder what goes through the head of someone doing work that they think stands a substantial chance of destroying everyone and everything that they love. Particularly if they’re in a group which mostly thinks that insufficient attention is being given to safety. I’m curious to see if there’s any clues in the detailed data.

Some ideas that come to mind before actually bothering to take a look include:

1. Maybe those with high estimates actually work in AI safety, or organisations concerned with mitigating the impact of AI. Doing AI research doesn’t mean you’re promoting AI. After all, virology researchers don’t usually want to infect everyone with a virus. But I don’t know if there’s enough (prestigious) AI safety jobs to go around for that to be the case every time.
2. Perhaps they believe that the upsides of AI are so great that a 10% risk of total destruction is worth the risk.
3. Perhaps they think that they’re more moral, careful and sensible than everyone else working in the field, and thus, given that someone is going to produce these systems, it’s best that it’s them. After all, we (nearly) all think we’re above average drivers.
4. In some part similar to the above, maybe it’s more that yes, some day there’ll be an AI that destroys everything, but it’s not going to be my silly little model. Maybe it’ll be 1000 years in the future and come from an entirely different line of research. We’re not surprised that people who build space rockets do what they do just because nuclear missiles look and work vaguely similarly.
5. Or perhaps the respondents who place high probabilities on absolute destruction don’t actually viscerally believe that number and its implications, or at least not all the time. Maybe they never thought about the possibility until they were asked. It is possible for a person to hold two different and incompatible thoughts in one’s head at different times or even simultaneously, particularly when one concerns your livelihood here and now and the other is a potentially distant future theoretical risk. We’re not constantly shocked at how people from oil companies don’t quit even though there’s a lot more evidence that their work is directly contributing to the destruction of the planet.
6. Let’s never forget the Lizardman’s constant. There’s a certain amount of respondents who will say they agree with almost anything, no matter how outlandish, on any survey. Even when they either don’t actually agree or never really thought about whether they do or not.
7. And then there’s the fact that most humans are not really very good forecasters, especially when it comes to rare or unprecedented events. Being good at making AI is a different skillset to being good at making forecasts about it. In particular we tend to be “disproportionately swayed by improbable but extreme eventualities, such as terrorism, that come to mind easily”, to quote Lieder et al, which the total destruction of everything would probably count as.
8. Because of the fact that some researchers were given different questionnaires to others it is also possible that the 69% who feel like insufficient attention is being given to AI safety are not representative of the sample who provided answers to the question resulting in that 10% chance of total destruction.
9. There is of course, as ever, the possibility that there are other methodological issues with the survey which a detailed reading might reveal.

📧 Reading the Data Is Plural newsletter - “a weekly newsletter of useful/curious datasets”.

Lots to download here for anyone vaguely data-obsessed.

Homeless people should not be arrested just if they smell - minister

One of the increasingly rare Conservative MPs with a good opinion! Wokeness runs amok!

Sometimes one can tell a lot about the legitimacy of a policy by the ludicrousness of the headlines it generates. This one is in reference to legislation the Conservative government is trying to get through that had the galaxy-brained idea of fining any rough-sleeping people - who of course famously have an excess of wealth to spare - up to £2500 for causing a nuisance. Prison sentences also available. Nuisance is defined widely enough that being “excessively” smelly or simply looking a bit like you might intend to sleep rough would be eligible for punishment.

The existing legislation - which is naturally a law from exactly 200 years ago: the 1824 Vagrancy Act - already makes rough sleeping and begging a criminal offence. Thousands of people without homes have been arrested for such crimes as “vagrant being found in or upon enclosed premises” or “begging and wandering around” in recent years. And there’s a concern that these changes would lead to a even greater criminalisation in practice.

It’s a policy so stupid and cruel that even several of today’s Conservative party are likely to rebel if the government pushes this forward. Good on them, especially “rebels” who are trying to work across party lines to decriminalise rough sleeping entirely. We desperately need new legislation on the subject, but not this new law.

Homelessness is not a lifestyle choice, and we know how to solve it.

Trump forced to see mean memes about him shared by prospective jurors

(from Rolling Stone)

Donald Trump’s latest trial is making for some slightly comedic scenes in amongst the horror. But this does raise the question of how on earth you find a jury of American citizens who have no pre-existing bias either towards or - particularly around Manhattan where this trial is situated - against ex-President Trump?

Find me 18 folk who have never made a comment or forwarded a meme even tangentially related to Trump or his extremely controversial clickbait policies and that’ll be 18 people not at all representative of any Americans I know.

Anecdotal evidence is coming to light that suggests that folk on the revolutionary GLP-1 anti-obesity medications such as Ozempic, Wegovy et al. are more likely to get pregnant. Sometimes this is joyfully welcomed. Other times - especially in cases where the person concerned is taking birth control pills - it’s more of a problematic surprise.

There’s a number of theories in play as to why this might be.

From Jezebel:

Experts say it’s possible that people are getting pregnant on these drugs because weight loss can help treat insulin resistance, which can lead to more regulated ovulation and menstrual cycles.

But the medications work by slowing stomach emptying, and that can affect the absorption of oral medications when they’re taken at the same time.

The drugs can also cause vomiting and diarrhea, which could mean people aren’t absorbing enough of their birth control pill to prevent pregnancy.

Early days yet as to how much this is a big and real phenomenon. But for it seems as because not a great deal is yet known about any effects of these drugs on pregnant humans and their offspring, if you do get pregnant whilst on these meds your doctor is likely to advise you to stop taking them. And if you don’t want to get pregnant, you might want to avoid relying on oral contraceptives alone - something which providers should hopefully already have informed you of.

Spotify has changed its payment structure such that playing most tracks they host will result in a $0 payout to the artist. That’s because they’ll now require a song to have at least 1000 streams a year in order to be paid out on.

To be clear, the claim is not that Spotify will keep the money themselves (not that they’re against that concept in general), but rather it’ll just get redistributed into payments for more popular tracks. Nonetheless, whether you think this is a good or bad move it’s interesting to learn a bit most about the makeup of the Spotify catalogue.

From NME:

According to Spotify data, there are around 100 million songs on the service, yet only around 37.5million meet the new requirements to generate revenue.

…

Spotify said that 99.5 per cent of all streams on the platform “are of tracks that have above 1,000 streams.”

It’s a real “winner takes all” situation. And darkly amusing to learn that one of the non-malicious reasons they’re doing this is that, because they anyway pay so badly, artists with those kinds of listening numbers were in practical terms not receiving any payment even under the former regime.

On average a song that got under 1000 streams over the past year earned a whopping $0.03 per month. And not even that 3 cents made it to the place you probably thought it did given the minimum withdrawals and transaction fees that the non-Spotify institutions like to apply. And that’s even before we consider the fact that, as we covered before, most money that does eventually get out of the system goes to somewhere other than the artist’s bank account for reasons of varying legitimacy.

US imposes first-ever limits on levels of toxic PFAS in drinking water

(From The Guardian)

PFAS are a set of chemical compounds that have been traditionally used for making things like water, stain or heat resistant products.

The problem is that they have been repeatedly shown to be risk factors for many illnesses:

The chemicals are linked to cancer, birth defects, liver disease, thyroid problems, decreased immunity, hormone disruption and a range of other serious health issues

They also don’t break down naturally, so end up being dissipated all through our environment, including into our drinking water. Hundreds of millions of Americans are likely being exposed in this way. Along with people from the rest of the world - a recent study found places all over the world where the groundwater exceeded the recommended thresholds, even when we know the levels are likely underestimated.

…this study suggests that a large fraction of surface and groundwaters globally exceed PFAS international advisories and regulations and that future PFAS environmental burden is likely underestimated.

These laws don’t cover every type of PFAS and certainly aren’t a ban. And only a minority of the average person’s PFAS exposure comes from water - probably around 20%. Other sources include the food we consume and the dust we breathe in. But, limitations acknowledged, they should improve the safety of drinking water.

EPA scientists calculated that the new limits will result in thousands of fewer birth-weight related infant deaths, kidney cancer deaths, bladder cancer deaths and deaths from cardiovascular disease

Assuming there are least acceptable alternatives for any critical use-cases, perhaps we can also hope to see more companies following 3Ms example and no longer seeing it as desirable to bother producing PFAS in the first place. Their decision to stop was at least partially due to the trend in regulations like these.

January 2024 saw the probably the biggest known leak of personal data yet known. An astonishing 26 billion records from various sources was found by researchers to available via open resources.

“The dataset is extremely dangerous as threat actors could leverage the aggregated data for a wide range of attacks, including identity theft, sophisticated phishing schemes, targeted cyberattacks, and unauthorized access to personal and sensitive accounts,” the researchers said.

Rather than newly hacked data, it seems to at least mostly be a collection of historical data breaches all handily available in one download. Which makes sense when one realises who the leaker was, outlined below.

Cybernews reports on some of the sites whose data was in the haul - site name and number of records:

• Tencent (21.5B),
• Weibo (504M),
• MySpace (360M),
• Twitter (281M),
• Wattpad (271M),
• NetEase (261M)
• Deezer (258M),
• Linkedin (251M),
• AdultFriendFinder (220M),
• Zynga (217M) ,
• Luxottica (206M),
• Evite (179M),
• Adobe (153M),
• MyFitnessPal (151M,)
• Canva (143M),
• JD.com (142M),
• Badoo (127M),
• VK (101M),
• Youju (100M).
• Daily Motion (86M),
• Dropbox (69M),
• Telegram (41M),
• and many other companies and organizations.

They set up a site where you can see if your data has been leaked which includes the above collection.

It seems like it was actually a site that offers the service of checking whether your personal data got leaked that was the inadvertent offender in this case. Whoops.

Leak-Lookup, a data breach search engine, said it was the holder of the leaked dataset. The platform posted a message on X, saying the problem behind the leak was a “firewall misconfiguration,” which was fixed.

Started learning how to be a ‘Full-Stack Developer’ with Mimo.

Being a DuoLingo streak addict, as well as a not-entirely-proud owner of such fine web properties as this one (along with a heady mix of being desperate to learn things in a world where a 2 minute break is too much to ask for) it was hard to resist when a friend tipped me off that there’s Duo equivalents but for coding.

I’m trying Mimo. It’s got basically identical types of gamification as Duo - streams, leagues, gems (well, coins) and similar 2 minute fill in the blanks and/or finish the line type exercises. The path I’m on promises to tech me everything web from html through to backend databases. Some of which I kind of know at least substantially outdated versions of already But I learned, as many of us did back in the day, more from the View Source menu option of a browser than actual education. A further variety of languages are available to learn if this isn’t your cup of tea - Python, SQL, Swift, JavaScript, HTML and CSS.

Sololearn is a similar looking competitor if Mimo doesn’t do it for you. They have a few extra languages on top of the above including C, C++ and Java as well as some more general fundamentals courses like ‘Tech for everyone’ and a new offering on using generative AI.

It’s too early to tell so far whether 3 minutes a day is going to make me the world’s best web programmer or not. Mimo’s site does proudly suggest one can ‘get hired’ after working through their offerings. But given I still don’t think I could hold a conversation in Spanish even after literally 1000 days. I guess I’ll see what it feels like if and when I get into topics I’m not already sort of familiar with.

TIL that an astonishing (to me) 1 in every 21 adults in the UK is a landlord.

The article I learned that from has made me question my thinking a bit. Previously I’d assumed that the broken UK housing market is essentially a problem of supply and that we should basically get over our NIMBYism and build, build, build. That still feels like it would be a good idea, and actually attracts mainstream political support these days, at least to some extent.

But apparently the statistical evidence isn’t tremendously in favour of this being the only, or perhaps even primary, problem. The ratio homes to households has actually grown over the past 25 years. And it’s not an especially bad ratio compared to other countries, the UK having about the average homes per capita for an OECD country, 468 per 1000 people in 2019.

So perhaps a large part of it is also a problem of legislation. Houses are simply getting into the hands of the wrong people. If so, then the argument goes that what we really need is some far less friendly policies when it comes to (private) landlordism.

It’s not clear to me that people’s homes should really ever be investment vehicles. They should be high quality, decent and healthful to live in. That may not be compatible with unbridled capitalism in a country of great economic inequality.

In a recent paper “Beware of Botshit: How to manage the epistemic risks of generative chatbots” , Hannigan et al. introduce (I think) the concept of ‘botshit’.

It relates to a particular style of use of output from one of the currently-fashionable generative AI models - ChatGPT, Bard et al - and is defined as:

Chatbot generated content that is not grounded in truth and is then uncritically used by a human for communication and decision-making tasks.

And “not grounded in truth” is a common feature of today’s large language models, as at the end of the day:

…generative chatbots are not concerned with intelligent knowing but with prediction.

An obvious example of botshit in practice might be the hundreds of websites that NewsGuard identified as having been created by generative AI and then published without care or concern by humans

Botshit can be distinguished from “bullshit”, which necessitates the same style of applying potential nonsense to a task, but in this case the content is specified as having a human origin.

Human-generated content that has no regard for the truth which a human then applies for communication and decision-making tasks

And it’s distinguishable again from “hallucinations” which is what occurs when chatbots produce:

…coherent sounding but inaccurate or fabricated content

Because botshit requires not only that a hallucination may have occurred, but that a human actively used the results for a given task.

Once LLM content potentially containing a hallucination is used by a human, this application transforms it into botshit.

To date I’ve quite liked the “fluent bullshit” formulation for describing generative AI nonsense output, although botshit is both catchier and makes is easy to distinguish the specific act of using AI generated ramblings for a specific task from other adjacent concepts.

📧 Reading the Garbage Day newsletter.

This often hilarious, sometimes thought-provoking, always addictive newsletter is all about what one could call internet culture, drama, beefs, whatever.

The latest edition I read caused me to learn the incredible fact that YouTube megastar MrBeast is spending the same amount on some of his mind-splattering retention-edit-pilled videos as an episode from the first five seasons of Game of Thrones cost to make. And there’s not even a dragon in the couple I put myself through. Yes, some of Señor Beast’s content apparently costs him $5 million per ep.

Want to stay up to date on the latest TikTok spat or AI growth-hack monstrosity without having to lend the actual offenders your eyes? Read about it in this newsletter instead.

Garbage Day is for folks that remember growing up in the west wild of AIM and Kazaa and message boards and know that, even though it’s probably politically destabilized most of the planet, the internet can still good and fun.

📺 Watched season 21 of of Dragon’s Den.

BBC ‘Business Thursday’ hit our household again this year with the Thursday back-to-back showing of both this den and the (aforementioned) Apprentice.

Again it’s a program where little changes over the years. They shook things up a little with the Dragons with a couple of ‘special guest’ celebrity-style judges popping in to throw some money around. And there was an episode where they had to overlay a health warning having failed to call out that there was little medical evidence that ‘ear seeds’ cured myalgic encephalomyelitis (ME).

This was possibly less of a faux pas than the Apprentice of course, one of whose candidates mysteriously vanished from the after-show gatherings having been found to be spouting ‘vile and sexist’ stuff about feminism, Judaism and Hinduism. He’d previously set up the “University of Masculinity” to find Moroccan wives with the ideals he prefers for divorced Muslim men,

But overall, the den provided oddly reassuring old familiar stuff from an earlier era. Sure, we could lecture sternly about the ever-increasing inappropriate valorisation of business, money and personal status above the social good. But at least that’s an old, pre perma-crisis problem.

Today’s DuoLingo seems to be 100% sponsored by our dog.

📺 Watched The Power.

This is the TV adaptation of Naomi Alderman’s justly-famous book of the same name. Or at least part of it.

It’d been so long since I read the book that I’d forgotten everything about it except the main premise - that women everywhere start to develop the weird ability to create and transmit electricity from their bodies at a level far more than what’s necessary to fatally electrocute people - and that I enjoyed it a great deal.

The key point being made here is around the implications for a world in which the average woman is no longer fundamentally threatened without recourse by physically stronger males; a world in which women by default have the upper hand when it comes to power. What would such a world look like? And, always of interest to me in this type of story, how would the transition phase go, how would those who today hope the power react to their potential dethronement?

Which made the start of the show a little disappointing, it being less captivating than I remembered the book as being. But as time went on either it got better or I subconsciously learned to like it. By the end I was into it enough that it’s very obviously not-really-the-end made me hanker for a second series.

R. A. Fisher’s quote from his address to the 1938 First Indian Statistical Congress remains extraordinarily true, even if the guy himself may have had some rather problematic views in other domains.

To consult the statistician after an experiment is finished is often merely to ask him to conduct a post mortem examination. He can perhaps say what the experiment died of.

True enough to make for a good email signature if nothing else.

Immediate low-key stress response whenever I see a message like this from a product owned by Meta.

(It popped up on my WhatsApp recently.)

‘Refugee Astronaut III’, from a recent visit to the Wellcome Collection’s exhibition ‘Being Human’.

From their commentary:

…it is hard to see Yinka Shonibare’s ‘Refugee Astronaut’ as anything other than some sort of dire omen from an environmentally catastrophic future.