Webcurios 16/02/24

Reading Time: 35 minutes


(Presuming, of course, that you do in fact consider continued existence to be a general positive – I’m currently ambivalent)

It’s a lovely day here in London and I would quite like to spend at least some of it not staring into the digital abyss – so you’re not getting an intro this week, and instead we’ll pile straight on into the links and the pictures and the songs and the good stuff so YOU’RE WELCOME YOU INGRATES.

I am still Matt, this is still Web Curios, and you are beautiful, no matter what they say, words won’t bring you down.

By Kevin Horan



  • The Encyclopaedia Automatica: Have you noticed it yet? The gradual erosion of information integrity slowly beginning at the edges of our field of vision, the slow creep of words and images that look at first, cursory glance like they make perfect sense but which on closer inspection reveal themselves to be merely sense-shaped? WELCOME TO THE CONTENTPOCALYPSE! Oh, ok, fine, I am once again being a bit hyperbolic and a touch alarmist – equally, though, stuff like this seems to be happening more and more often and it doesn’t feel like a positive development. Still, let’s ignore the pollution of our information ecosystem and instead explore the world of experimental AI content generation, courtesy of the Encyclopaedia Automatica, an interesting project which looks to build a parallel Wikipedia as a joint man/machine effort – the idea being that it combines the direction and guidance of a human editor with the magical ability to spaff out infinite copy of an LLM. So anyone can create an article, spin it up with GPT and then use prompts to tweak it to ‘perfection’ – change the tone, add more depth, etc etc – and the whole thing is basically an art experiment project to scope the limits of utility of an LLM as a knowledgebase. As you’d expect, the articles in here (of which there are a surprising number) have the classic GPT-generated flavour of *feeling* like they are telling you something but leaving you post-reading feeling oddly empty and more stupid than you were before you started, which, one might argue, isn’t exactly the ideal reaction to a knowledge Wiki, and there’s part of me that worries slightly that this gets big enough that the…less-critically-aware just start using it as a Wikipedia alternative. Still, if you’ve ever wanted to spend a happy five minutes cajoling a machine to imagine the history of, I don’t know, the pessary, then you will enjoy this very much.
  • The Grannies: I am SO happy that this is now online – this is a documentary film (presented on a neat little website) which was presented as part of the Now Play This festival at Somerset House in London a few years back, and is all about the experience of a group of friends playing Red Dead Redemption multiplayer together and discovering the weird, liminal, broken spaces that exist at the edges of the game maps, where the normal rules of the digital environment stop applying and everything becomes surreal and slightly dreamy…Honestly, I appreciate that what I have describes sounds…less-than-compelling, but if you have any interest in games, the idea of digital ‘spaces’ as being meaningful in some way, or just of travelogues in general, this really is lovely – sit with it for 30 minutes, I promise that you will enjoy it more than you expect.
  • Fun AI Video Manipulation Stuff: If you are ANXIOUS TO LEARN about the new OpenAI text-to-video model Sora (what is WRONG with you?!) then you might want to skip to the longreads section – this is instead some experimental tech from TikTok, which looks like allowing for impressively-granular degrees of control when applying animation layers to static images. Click the link and see for yourselves (no really, do, it’s super-impressive) – basically though, this lets you select ‘areas’ of an image to animate, which then enables you to create specific, guided movements, in a way which is presumably easier to control and limit than simple prompting. I have been slightly astonished at the pace of development of this sort of tech, and in much the same way that ‘low-end logo monkey’ is sadly not really a job anymore I am pretty confident that ‘low-end video editor’ is equally going the way of the dodo in ~24m or so.
  • SecondSoul: I am not, it may surprise you to learn, a particularly committed theologian, but a decade or so of Catholic school, the whole ‘being half Italian and spending a lot of time in Rome’ thing and a potentially-ill-advised bargain made with the devil aged 17 means that I have a passing acquaintance with the concept of the ‘soul’ – and I am pretty sure that there’s no way in which it’s compatible with a company whose strapline is, and I quote, “Monetize your community with your AI clone”. Just take a moment to read that back again – MONETISE YOUR COMMUNITY WITH YOUR AI CLONE. God, it just SCREAMS soul, doesn’t it? Anyway, this is a terrible company with a terrible product that, God willing, will die on its ar$e – it’s also, I think, a precursor of Things To Come. The idea, as you may have been smart enough to work out for yourselves (well DONE!), is that using this platform you can create bots of yourself, which you can then deploy on Telegram, to ‘interact with your community’ – the idea being that people will pay actual cashmoney to chat with your AI avatar, while you kick back on a  beach somewhere raking it in from the passive income. Which, obviously, LOL, but also I can 100% see this sort of thing taking off amongst kids who need a new hustlehope to cling to now that the ‘creator economy’ bubble has finally been revealed to be a gigantic lie.
  • Open Souls: Seeing as we’re on ‘souls’, I can’t quite tell how sincere this is, but, well, it’s the Twitter account of “a group of insane devs who truly believe they can give AI souls”. There’s a Discord you can join, should you be so inclined – I only found this at about 643am this morning and as such haven’t had a chance to dig around, but if any of you know ANYTHING about this I would love to learn more.
  • Put Your Voice To Work: Following on from the ‘create a digital slave and make it toil for you!’ link above, this is another service along vaguely-similar lines – London-based ElevenLabs, which has managed to maintain its industry-leading position in AI voicecloning over the past year or so, is now offering you the opportunity to create and monetise your own AI v/o artist – create a clone of your dulcet tones and you can make the model available for others to use, getting a cut of the fee every time they do so. On the one hand, this is probably a no-brainer for anyone currently making a living from voiceover work – I mean, you may as well, right?; on the other, it does rather feel like a naked attempt to get a lot of pro-quality training material for the model (cynical, moi?), and I have…significant doubts about the likely demand/supply ratio for voice models and the resulting monetisation opportunities that will in fact result. Still, LET’S ALL MAKE DIGITAL SLAVES OF OURSELVES!
  • Pint Prices: It’s slightly terrifying how quickly Londoners have become inured to the now-insane prices in pubs – “ninetythree quid and a pint of plasma for a half of overcarbonated p1ss that tastes, inexplicably, of grapefruit? Make it two, barkeep!” – but should you wish to attempt to FIGHT BACK against the madness then this website, which maps the price of a pint across the capital, might be useful. I can’t vouch for how up-to-date this is, or how accurate, but it might be useful.
  • Lovely Interactions: This is SUCH a nice piece of webwork, and a really lovely calling card for the digital agency in question (Off Brand, apparently) – basically it’s a little game where you have to identify five different types of interaction that you, the user, can undertake with the site, and for each one you find you’re rewarded with a really satisfying little animation and sound effect; this is very simple but SO nicely-made, and communicates the joy of a nice piece of UX/UI really effectively.
  • Barasol: This feels a *bit* like a site that was hacked together as part of some sort of joke, a digital response to ‘what would happen if you created a website that offered the opportunity to explore some incredibly bargain-basement travel options?’, but, equally, I think there’s something potentially quite fun about where it might take you (or, alternatively, something potentially INCREDIBLY DANGEROUS) – Barasol does a very simple thing, namely running searches for flights on Ryanair AND the cheapest available Airbnbs in a given location, thereby offering you the cheapest possible travel options to any given destination. Which, on the one hand, GREAT! But on the other, if you have ever a) travelled Ryanair; or b) stayed in a cheap, sketchy Airbnb then you will know that the potential for this to go very sideways very quickly is…not-insignificant (I have stayed in two spectacular Airbnbs, fwiw – one which was, quite literally, a poorly-converted garage made of breeze blocks which had a hotplate, a tiny fridge and, inexplicably, a shiny black plastic toilet; the other was literally a windowless basement accessed via a floor hatch in the owner’s actual flat. Let me know if you want details of either!).
  • Frame: I’ve been amused to read the reports of early Apple Vision Pro adopters taking the devices back this week upon realising that they don’t actually need or want a massive, heavy, face-mounted home cinema rig – still, I remain convinced that some sort of augmented glasses-type revolution is coming in the not-too-distant future, and stuff like this rather reinforces that. Frame is, basically, a revamped Google Glass – the specs, available for pre-order and apparently shipping in April, look a lot like the Snap Specs from a few years back, and will feature a combination of different AI tech to enable image recognition, realtime translation, chat with an AI assistant, and all the other stuff that three years ago would have sounded impossibly scifi and which we’re now just all sort of like ‘wevs’ about (genuinely mad, that, by the way). I am skeptical about the user experience with these – and indeed for the company’s ability to make the featureset available without a not-insignificant additional subscription to the various AI services that you’re using (OpenAI, Whisper, etc) – that said, I think these are also open source-ish, meaning that enterprising enthusiasts will have rehacked them within minutes of receipt so that they are instead running a homebrew version of LLAMA trained on 4chan or something. I don’t think I can stress enough how deeply, deeply weird everything is going to get when everyone has their own AI stack embedded in their wearables – AND NOONE KNOWS WHAT THAT IS OR WHAT IT CAN DO OR WHAT IT IS TELLING THEM. Anyway, if you want to look like a colossal pr1ck come April and have everyone in your life treat you with immense suspicion every time you wear the damn things then, er, preorders are open!
  • Somnivexillology: I don’t dream. Or rather, I probably do but I don’t remember them – thanks to decades of weed abuse I basically fall into a coma, and whatever weird pirouettes my subconscious chooses to undertake while I’m passed out remain forever mysterious to me (apart from rare occasions when I have to go cold turkey, at which point my sleep becomes like some of the more overwrought bits of Fantasia). Which is by way of unwanted, unasked for and largely-uninteresting preamble to this subReddit which features people sharing examples of flags that they have seen in their dreams. Personally the idea that people not only dream with vivid enough specificity to conjure up recognisable flags, but also that they can remember them upon waking, boggles me entirely, but for all I know you’re all similarly occupied at night and your dreamscapes are neverending parades of ships’ masts fluttering colourful standards as far as the eye can see. Anyway, you can read a load of anecdotes about these dreams, along with seeing illustrations of the imagined flags in question – if you want a flavour for the vibe here, this is pretty illustrative: “I had a dream where there was a new Hotel Transylvania movie but where Vlad Dracula managed to get Vampire Transylvania as smth like Liechtenstein mixed with Kosovo and with Moldova and it was named Vampire Kingdom of Transylvania, it was a semi-constitutional monarchy with Vlad Dracula as the king” – SERIOUSLY THOUGH, WHO DREAMS LIKE THIS?!?! Is it…is it all of you? Am I missing out?
  • Have A Good Today: Everything is hard and – not to be a downer, but it’s important to be honest and I like to think of myself as a Realistic Friend – it’s not going to get any easier. Which is why I think it’s important that every single person reading this, all thirteen of you, click this link and make this site your new homepage, because I guarantee that, while it won’t actually make anything better, being confronted with a small, slightly-garish animated gif wishing you a good day every morning will slightly take the edge off The Fear. If noting else it will give you a GREAT collection of graphics for the next time you decide to cosplay as a grandmother on the internet.
  • Panorama: This is a nice idea – it’s a bit slow and a bit clunky, but at least one of you can definitely steal this idea as pitch-filler. Panorama is a Google Maps/text-to-image mashup, which basically lets you use AI to modify Google Streetview in interesting visual fashion. Reimagine the street on which you grew up as a dystopian hellhole (I grew up in Swindon, no reimagining necessary lolzzzzzzzzzz), see what the view from atop Christ the Redeemer would look like were Rio underwater, turn Manhattan into a Barbs’ paradise, etc etc etc – there are a million-and-one different ways you could use this for campaigns, so I won’t insult you by suggesting any (do your own work ffs).
  • Arvind Sanjeev: I came across the site of Arvind Sanjeev via his prototypical design for an AI synthesiser – which you can see here, it looks LOVELY and is (another) great example of the fun things that you can do when you combine generative AI with physical creation (WHY AREN’T MORE PEOPLE PLAYING WITH THIS STUFF WHAT IS WRONG WITH YOU ALL?!?) – and I am linking to it because a) they make loads of really interesting stuff and it’s worth having a poke around the site as there’s some really smart thinking on display; and b) I al SO IN LOVE with the minimalist photo on the homepage and the way it changes as you scroll in and out (try it, honestly, it is beautiful and I lost a couple of minutes to just zooming in and out on it just now). Oh, actually, while we’re doing ‘personal websites that I think are gorgeous’, here’s another one by Rauno Freiberg which is equally-gorgeous.
  • VIRL: Whilst I’m obviously a desperately cynical and jaded creature who long since lost the ability to feel anything approximating ‘joy’, I confess to continuing to feel a genuine sense of thrilled curiosity at all the different, interesting, terrifying things that we are on the cusp of being able to do thanks to The Machine – witness, for example, the theory behind VIRL, which builds on some of the work from last year around creating AI ‘agents’ and letting them loose in a simulated environment, to see how they would pursue goals and develop social bonds, etc. This is now all open source – you can get the code here, which means that ANYONE can now start to spin up their own modeled simulations USING ACTUAL REAL WORLD CITY DATA, which is frankly insane. It’s a bit hard to get your head around, but, basically, imagine that there’s a rough ‘digital twin’ of the world being created, using actual map data and images tagged to locations, to create a digital environment which tracks the real one – now into this world, anyone can drop AI agents, give them specific motivations or goals, AND SEE WHAT HAPPENS. Honestly, just typing this stuff is giving me proper, brainfizzing futureoddshocks – I can’t stress enough quite how astonishing the potential and theory here is, and if you can be bothered to wade through the *slightly*-techy explanations on the site then you will start to get a glimpse of some really quite astonishing potential just around the corner (and some obviously-terrifying implications, fine, but let’s not dwell on that right now).
  • Lina: Do you believe it’s possible to scry the very depths of someone’s soul based on a half-ar$ed doodle? Do you think that the best way to REALLY understand someone is to pay incredibly close attention to the stuff they scribble in the margins? Are you happy to entrust that INCREDIBLY IMPORTANT analysis to The Machine? Er, great! Lina sells itself as an AI art therapy app, and will basically tell you your personality problems to seven decimal places based on its ‘analysis’ of one of your drawings. This is, honestly, almost impressive – total b0llocks, marketed as ‘insight’, with in-app purchases! Based on this I am slightly amazed that noone’s attempted to launch a ‘Phrenology, but AI!’ startup (I am willing to sell this AMAZING CONCEPT for a low six-figure sum, form an orderly queue).
  • The Tearing Curtain: Without a shadow of a doubt the most satisfying bit of code I have seen in YEARS. Honestly, click the link and play with the curtain – it will make you feel like a cat faced with a particularly-wiggly piece of string (trust me, that will make sense, promise). I strongly believe that ALL websites ought to have something like this hidden on them somewhere as a gentle stress reliever.

By F Scott Hess



  • Tangara: When my brother killed himself, I got his iPod – there is something genuinely…odd about a device being able to tell you exactly what the last songs someone listened to before they topped themselves are, let me tell you – but it bricked itself years ago and I’ve been mp3less ever since. I am sorely tempted, though, by the Tangara, a fully-funded project (which still has about a month to go) which will later this year ship an MP3 player which, basically, is like the original iPod but with a bunch of quality of life updates, It features the gorgeous scrollwheel – still one of my favourite interfaces ever (what, you mean you don’t have ‘favourite interfaces’? weirdos) – it’s open source, and highly-moddable, meaning you should be able to keep it going pretty much forever by swapping out the battery when it carks it. It’s not the prettiest device in the world, fine, but the sense of nostalgia this stoked in me was such that I can overlook the shonky aesthetics.
  • The Album Cover Bank: ALL OF THE NIGERIAN ALBUM COVERS, EVER! This is a wonderful repository of design and musical/cultural history – from the blurb, “Cover Bank is a digital archive of 5300+ Nigerian album covers from 1950 to date. It provides a unique view into the evolution of Nigerian music. At its core, the archive is a research project that hopes to establish the history of Nigerian graphic design. Through this website, we hope to highlight cover artists as important cultural producers. Cover Bank is also a platform for design and storytelling inspiration. Our goal is to become a valuable tool for artists, enthusiasts, educators and researchers of all kind. The archived is built and maintained by wuruwuru, a collective for independent creators in Lagos.” Even better, all the albums (or at least all the ones I’ve clicked on) link to Spotify, so you can listen to any whose aesthetic takes your fancy (I was so taken by the artist name ‘Lifesize Teddy’, for example, that I am listening to them right now – it’s pretty good!).
  • Sabio: I can’t vouch for how good this actually is – and, personally-speaking, I would trust any ‘AI powered fitness coach’ about as far as I could throw an actual, IRL fitness coach (to be clear – that is no distance at all) – but I think there’s the kernel of a clever idea in here. Sabio basically lets you hook your Strava data up to some sort of ‘AI’ (no clues on whose tech this is using, but let’s assume it’s GPT-3.5-level), which will then give you personalised fitness recommendations and the like based on what the app tells it – so, in theory, it will adjust your regime based on the sort of workload you’re putting yourself through already and adapt based on your performance. I have…doubts about the efficacy and value of this stuff, but it’s an nice example of the ‘pipes’ theory of AI utility (er, that’s something I just made up, to be clear, rather than an actual theory) whereby it’s often more useful to think of ‘what data sources can I give to The Machine and what might it be able to do with them?’ rather than just thinking about prompts, etc.
  • Creative Fest: A *lovely* YouTube channel which mainly features really fun little making/crafting/science-y projects that you might want to experiment with – this feels like the sort of thing that might be WONDERFUL for a very particular type of child, although perhaps my saying that simply indicates how little I know about children and what they might or might not be into in 2024. It’s all very homespun and a bit shonky, and I really love it for that – also, the videos are wordless and faceless, and I don’t think they are North American, which is another reason to appreciate them (sorry, North Americans, but we’ve had quite enough of you for the past few centuries).
  • Carl’s Friends: I appreciate that for the vast majority of you ‘sourdough starters’ are a conversational theme which you’ve long since confined to the Bad Memory Oubliette that is ‘The COVID Years’ – but I am SO THRILLED that this exists that I absolutely must share it with you. Would you like to receive a sample of sourdough starter that has been going for NEARLY 200 YEARS?!?!?! Of course you would, you’re not made of stone! Carl T Griffith, after whom this site is named, died in the early-00s after a life apparently characterised by generosity – Carl apparently gifted samples of this sourdough starter, which had been in his family since the mid-1900s, to everyone he met, and since he died a group of people have maintained his legacy via this site, where you can find instructions on how to order your own sample of INCREDIBLY OLD BACTERIA for practically-free. Honestly, this is SO LOVELY, and were it not for the fact that I don’t think I could cope with the guilty of receiving and then promptly-killing a near-bicentenarian organism I would totally get involved with this myself.
  • Chronolog: Via Kottke comes this excellent site which has apparently been going since 2017 and which exists to let anyone post timelapses of a specific location, tagged to a map, to help environmentalists and researchers develop an impression of how a given area is changing over time. This is a really nice idea – anyone can apply to get a physical ‘mount’ sent to them, which they can then put up in the location in question; passers-by are encouraged to take a photo using the mount (which ensures each image captures the same field of vision) and then upload it to the site using a QR code; each image is added to the timelapse, creating a crowdsourced record of how the landscape changes and evolves over time. This is a North American project and as such the locations being captured are overwhelmingly in that geography – there’s obviously one set up near Oxford in the UK, but I want more please thankyou.
  • Entrances 2 Hell: “A constantly updated catalogue of entrances to Hell in and around the UK“ – this is obviously a joke, given the fact that the UK *is* Hell, lol!, but it’s quite a funny one and I admire the long-running commitment to the bit on display here.
  • Stickers To Manage Replies By: This feels very much like an artefact of an era or two ago, when people still actually shared stuff on social media and Twitter was still alive – I won’t say ‘good times’, because, honestly, lol, but there’s a certain degree of masochistic nostalgia attached to the memories. Anyway, Dan Hon has designed these stickers which are conveniently available on Flickr and which you can append to your social media posts for added clarity – they say things like ‘This is an observation, no reply necessary, no need to help’ or ‘Do not reply to tell me to use Open Source software’, and Dan was apparently motivated to make these because of A Particular Type of Person he’s finding a lot of on Mastodon, and if you want a decent reason why that platform is never going to be anything other than a niche concern then, well, here you are.
  • Florence As It Was: I confess to not really liking Florence very much as a city – yes, it’s beautiful but it’s all too white and, honestly, Rome’s just *better* – but I did rather enjoy this project, a joint venture between a bunch of US academic institutions and the University of Florence. “ “Florence As It Was” has as its mission the gradual reconstruction of this major cultural center, one structure at a time, city block by city block. Paintings produced by artists during the early period that feature buildings constructed before 1500 – including those that no longer survive – have been used to recreate the exteriors of churches, municipal offices, and city gates. Using extant architectural evidence, documents, art images, and contemporary representations, we will build a three dimensional model that will allow viewers to circumambulate a structure, venture inside it, and see stitched onto its walls and furnishings the images that once adorned it. Embedded into this reconstructed structure will be texts –in their original language with English translations –that help explain how and why a specific object looks the way it does, why a specific artist was commissioned to produce it, and/or the motives behind its commission from the vantage point of its patron. Literary passages, tax records, even musical performances will be accessed through clicks on ‘hotspots,’ allowing writers and notaries and musicians to speak for themselves. Hyperlinks to other reconstructed buildings will allow us to make connections, literally, to other spaces and the people who occupied them.” This is an ongoing project and as such very much incomplete, but there’s something lovely about the fragments of history you can explore already, and the way it stitches together maps and imagery, and any art historians or simply Fans Of Florence will rather enjoy stepping back in time.
  • I Have No TV: I have to give a big upfront caveat here, in that I have watched literally NONE of the content on this site and can’t therefore promise you that it’s good rather than, I don’t know, being seventeen different shades of whackjob conspiracy content. That said, the potential here is vast – I Have No TV is a site which exists to pull together ALL OF THE FREE DOCUMENTARIES on the web – you can search them by theme, or title, or just pick a title at random, and there are apparently over 4000 on the site for you to choose from…I just hit the ‘random’ button and got sent to watch something called ‘My Love, Don’t Cross That River’ which according to the short blurb is “a story of a couple in South Korea who share intimate moments after 76 years of marriage” and now I am crying, so thanks a fcuking lot ‘I Have No TV’.
  • Moonlight World: This is a fun idea – whilst I personally have no truck with Tarot and the occult and all that jazz – although, and apologies for the diversion here, but I did once have my cards done at a small esotericists in Covent Garden c.2003 (I was with a friend, she was into it and insisted I get mine done too) and despite the fact I was very much dressed like a 15 year old boy from 1994 (muchlike today, in fact) the reader accurately guessed that I at the time worked in politics, which honestly scared the sh1t out of me – I understand that All This Occult Stuff is quite popular these days, and it’s entirely possible that many of you live your lives by the Major and Minor Arcana. Moonlight World is on the one hand a company flogging online tarot readings, but it’s also offering ‘digital tarot spaces’ for free, meaning if you want to conduct a reading for your friends you can use their ‘virtual tarot room’ to do so, complete with digital cards and some quite nice animations. This is rather nicely done, and a smart bit of promo for the main business imho.
  • SandCastle University: I don’t know what you do for a living, but I am going to guess, based on your readership of this newsletter, that most of you have the sort of stupid, made-up, largely-pointless sort of white-collar, early-21stC media-or-’creative’-adjacent jobs as I do (genuine apologies to any readers who actually do make the world a better place on a daily basis via the medium of their toil – sorry to lump you in with me and these other useless cnuts!) – the sort of job that it’s famously impossible to really explain to your parents. Spare a thought for the people who founded ‘Sandcastle University’, though, who presumably at some point or another had to explain to their parents that they were jacking it all in to become, er, professional sandcastle building trainers – although on the flipside this does appear to be their actual job now, meaning they probably win. Sadly I think this is a solely US concern, but if you happen to be in North America and want to pay to have a bunch of people teach you how to, er, build better sandcastles, then MERRY FCUKING CHRISTMAS. I am not sure why, but I think there is something ASTONISHINGLY BLEAK about professionalising and monetising the act of ‘playing in sand’.
  • Film Secession: Very much one for the serious cinephiles amongst you, this – you have to pay to get access, the whole thing being subscribers-only, but read the description and see if it sounds like it might be up your street: “The singularity of cinema lies in its unprecedented capacity to transform the energies of the other arts into an integrated audiovisual experience. This synthesis makes cinema particularly engaging, immersive, and resonant, although, precisely because the constituent elements are organically fused together, it can easily be taken for granted. Film Secession creates new ways of exploring the ideas and artistic currents that have shaped different filmmakers, periods, and art forms. Subscribers will discover nonlinear pathways through the histories of the arts, be able to watch rare films provided by the world’s preeminent studios, production companies, and archives, and have special access to events held worldwide.” Does that appeal, or does it sound achingly pretentious and a surefire way to destroy the magic of an artform? YOU DECIDE!
  • Curated Design: SO MANY DESIGN EXAMPLES IN ONE PLACE! Searchable, filterable and INSANELY VOLUMINOUS, this is a wonderful resource for anyone who needs inspiration or examples or just a bit of a creative nudge.
  • Flaming Hydra: While 2024 sets to be yet another nadir for most media organisations, and the collective Big Beasts of publishing compete to see who can fcuk up the most spectacularly, there are interesting shoots of recovery and germs of alternative models – 404 Media this week announced that its in the black after only 9 months, the Patreon-funded YouTube gaming channel Second Wind seems to be going from strength to strength, and now there’s Flaming Hydra, which might not end up working but which, I think, has a really interesting model behind it. “We’ve invited several dozen noted writers and artists to join us, for a total of 60 members. Each member agrees to contribute a minimum of one original piece per month to an ingenious, brief and captivating daily newsletter, in exchange for an equal share of the subscription proceeds, payable monthly. Flaming Hydra members retain the rights to their work absolutely. Subscribers to Flaming Hydra will receive articles and essays, comics and criticism, humor, literature, photos, and reviews, with surprises each weekday. Because there are a lot of us, only a small amount of work is required of each member. We’re sharing audiences, work, and resources, so that all can benefit and thrive.” It’s worth checking out the list of contributors who’ve signed up – it’s a really impressive selection of names, and if they all commit to this for the long-term then it stands a chance of thriving, I think. Subs are about £2.50 a month, which feels…reasonable, I think, so if you can afford it and like the concept it could be one to try out.
  • No Vehicles In The Park: Have you ever had to think about community rules and boundaries? Anyone who’s ever been in a position of moderator responsibility ANYWHERE knows that it’s a peculiarly-horrible role, mainly because it is SO FCUKING HARD to set working rules and principles that do what you need them to. “Nonsense!”, I hear you scoff, “you’re a moron, Matt! I could totally come up with a set of universal principles to mandate accepted conduct on a theoretical social platform!”. To which I respond “well try this little game and see how you get on then, mr fcuking big boll0cks” – honestly, this is SUCH a clever and nicely-made exploration of how community moderation works.
  • Periodic Tables of Almost Everything: Would you like a Pinterest board collecting hundreds of different images of different periodic tables, displaying information about EVERYTHING YOU CAN IMAGINE (apart from chemicals, as that would be BORING and TRADITIONAL)? YES OF COURSE YOU WOULD! I genuinely have no idea why people have made all of these things – WHY DID ANYONE FEEL THE NEED TO BRING A PERIODIC TABLE OF HARRY POTTER INTO EXISTENCE, WHAT IS WRONG WITH YOU??? – but (aside from that specific Potter-y example) I am happy that they exist.
  • Hurt Party: I’ve featured one of Amy Baio’s card game Kickstarters in Curios before, and this is her latest – Hurt Party sounds like a genuinely fun little game, perfect for quick sessions over drinks, in which players compete to offer the best, sincerest apologies for slights delivered. It’s about halfway there with a month to go, so there’s a good chance it will get funded, and if card games and light roleplaying are your thing then you will enjoy this I think.
  • 500 Meals: This is the photo website of one Jack Baty – or, specifically, the part of it on which Baty for reasons known only to him, documents 500 meals he has eaten over the course of (more or less) a year. I LOVE THIS SO MUCH – honestly, there’s something sort of brilliantly-deadening about seeing one of the essential elements of life laid out so baldly, so unsentimentally, so…poorly-lit. Also, I do slightly worry that Jack might not be getting as much fibre as he ought – is…is it good to eat that much meat and eggs? I am not sure it is. Anyway, this is a GREAT project and I am honestly tempted to replicate it (I won’t, but the temptation is very real).
  • Part-Time Hermit: This is quite remarkable, and oddly-beautiful – were I a less miserably-cynical person, who has dust where their soul should be, I might even say ‘almost spiritual’ (but I am, so I won’t). This is the YouTube channel of (I think) a Portuguese Friar, who spent a year living as a hermit in Italy, and who is releasing a series of videos documenting his experience, and, honestly, they are BEAUTIFUL and peaceful and even I found myself slightly-mesmerised by…look, there’s no other way of saying this and so apologies for the hokiness here, but by his sort of ineffable peacefulness…each one is about 40 mins long, but give ‘January’ a try and see how it grabs you, you might be surprised.
  • Fartographics: FREE COMICS! Specifically, free comics from a bunch of young Croatian artists who are publishing their work as a collective under the charming ‘Fartographics’ brand. There are two editions online so far, each helpfully translated into English for the non-Croatian speakers, and they’re available for free (though you can chuck them a quid or two by way of thanks should you feel so inclined), and there’s a really nice mix of art and narrative styles on display here which are worth exploring.
  • Lemmings: Taking up the coveted ‘game at the end of Curios’ link this week is this ABSOLUTE GIFT of a link. Are you English? Did you grow up in the 80s? Did you or anyone you know have a computer on which to play games? GREAT, in which case you will doubtless be familiar with Lemmings, one of the first genuinely brilliant puzzle games whose simple premise (‘help all the lemmings get from one end of the level to the other without them all dying in a variety of comedy slapstick ways’) masked level design of quite fiendish complexity, and whose titular characters are one of the most incredible examples of ‘how to design an iconic character with approximately 9 pixels’ work you will ever see. Honestly, I can’t stress enough how AMAZINGLY GOOD this game is – if nothing else you all need to play it at least once, because the small vocal sample you hear when you set the ‘Armageddon’ command, whereby all the Lemmings explode after a short countdown, is honestly an audio tic that soundtracks my internal monologue EVERY FCUKING DAY and has done since approximately 1991 when I first played this.

By Amy Sherald



  • WebCam Tears: Videos of people crying to webcams. Absolute, 100% pure digital ART, this one.
  • Pog-A-Day: This is now-defunct, and hasn’t been updated for about 7 years, but, well, WHAT A LEGACY! SO MANY POGS!


  • Chocolate Bob Ross: Genuinely-horrible AI-generated imagery and animation, in the classic ‘plastic body horror’ style – these are properly unsettling, in a good way.


  • The World in 10 Years: On the one hand, all attempts at predicting the future are doomed to failure and ridicule – on the other, if you restrict your futuregazing to the short-ish term, you might occasionally hit upon something accidentally-prescient. In that spirit, then, have this FASCINATING bit of work pulled together by the Atlantic Council’s Scowcroft Center for Strategy and Security, which last November asked 300-odd people in government or civil service or academia or the private sector for their thoughts about the likely shape of the world in a decade’s time – it, er, may not wholly surprise you to learn that the projections aren’t wholly rosy! It’s a lot more interesting than that, though, promise – although it’s a *bit* US-Centric for my tastes, and I think that skews some of the projections here (I don’t, for example, think that anyone in the world outside of some VERY optimistic people in San Francisco think that America is going to be the global tech leader in 10 years time), there’s a lot of really interesting speculative thinking (and only a very small bit of it is about AI, promise). This came via Sentiers, by Patrick Tanguay, which continues to be a superb resource for smart, interesting thinking about potential futures and systems.
  • AI and Planned Economies: NO WAIT COME BACK I PROMISE THIS IS INTERESTING! Erm, ok, fine, it’s an academic paper and it’s also QUITE dry, but I promise that the questions it asks around computerised central economic planning, and how we might possibly expect AI to impact the discipline, are fascinating – it is VERY TECHNICAL in places, and I confess to skipping over the sections that contain equations, but there’s a load of fascinating sections which address the potentially-unknowable ways in which AI might choose to administer an economy if granted autonomy and a clear directive. Can…can anyone hear the sound of the paperclip machine firing up?
  • Hello Sora: So overnight OpenAI announced it’s forthcoming text-to-video model – it’s not in the wild yet, so all we have to go on is their blurb and the obviously-cherrypicked examples of the model’s outputs, but it’s safe to say that based on these and the general spec (vids up to a minute long, style transfer, animate stills, etc etc) that the people at Runway and the rest are probably feeling a bit assailed right now. It’s worth taking the time to read through the various capabilities they outline on the page and looking at the sample vids embedded throughout – obviously all of this needs to be heavily caveated with ‘don’t believe the hype, or at least don’t wholly believe it’, but there are some hugely-impressive outputs on display here, not least in terms of the ability to maintain object permanence and to deliver reasonably-lifelike human movement. It’s impossible to tell whether we can expect the improvements to continue arriving at pace with text-to-video, but the difference in quality between this and what was being churned out by Runway even as recently as ~3m ago is startling, and if it continues like this then MY GOD is YouTube going to be absolutely ruined in a couple of years. BONUS TTV CONTENT: this is a decent enough overview of the current state of the market (as of yesterday, at least), pulled together by the awful cnuts at A16Z.
  • Preferable Future Habitats: As a general rule I don’t find AI-imagined futurescapes particularly interesting or worthy of scrutiny, but I was amazed by how…incredibly soothing I found this visual essay, in which Pascal Wicht explores some urban design and architecture ideas using Midjourney…I don’t know exactly why, but the combination of visuals and words here is soothing to an exceptional degree, maybe because some of the urban imagery imagined by The Machine in this piece is oddly reminiscent of an area of Rome near where I lived, and there’s something lovely and conversational about the way Wicht talks you through his thinking and what each image prompted him to consider in terms of design and planning.
  • Pluralistic AIs: Yes, ok, fine, it’s an academic paper but I promise that it’s a (mostly) readable one, and it’s pretty short, and the questions it raises – about a new way of thinking about AI alignment more connected to the concept of ‘pluralism’ in political thinking. Here’s a summary – this, honestly, is a really interesting series of potential principles of setting guidelines around model performance: “In this piece, we propose a roadmap to pluralistic alignment, specifically using language models as a test bed. We identify and formalize three possible ways to define and operationalize pluralism in AI systems: 1) Overton pluralistic models that present a spectrum of reasonable responses; 2) Steerably pluralistic models that can steer to reflect certain perspectives; and 3) Distributionally pluralistic models that are well-calibrated to a given population in distribution.”
  • AI Is Coming And Noone Cares: Last year the artist Fred Wordie launched a project called ‘Dear AI’ (which I featured in Curios, as it happens), a piece of ‘design fiction’ which presented “a ‘speculative’ company that allowed users to generate personal letters and send them as faux handwritten cards. It told of a near ‘future’ where Generative AI would further erode what it means to be human. Like many of my Design Fiction work, I chose to place it in front of the public as if real, with little context” (I, er, didn’t clock it was a spoof – I AM A MORON AND I AM SORRY). Wordie expected it to elicit a wave of commentary about dehumanisation and the insertion of the machine into personal relationships…and it didn’t (except from me, obvs). Instead, people are using it (and occasionally complaining to Fred that it doesn’t work properly) – because, it turns out, NOONE ACTUALLY VALUES THE PERSONAL AT ALL. I can’t stress enough how important this ‘insight’ is, and how much it ought to inform your own personal perspective on the likelihood of The Machine eating everyone’s jobs.
  • Nozick and the Hedonism Machine: Robert Nozick was very much one of the Bad Old Guys when I was doing political theory – seriously, have a read of his thinking about intergenerational responsibility if you want a chilling example if very smart selfishness – and it was nice to be reminded of the fcuker’s name this week in this piece, all about the question of whether, given the opportunity, we would strap ourselves into The Machine were it able to gift us a sensory experience indistinguishable from real life. Nozick famously argued that we would not – that there is a specific value that we ascribe to ‘real’ experience which a simulation cannot replicate – but this article does a decent job of presenting some decent counterarguments; I don’t know about you, but based on some light observation of ‘how the world seems to be going’ I’m personally betting against ol’Bobby on this one.
  • Why Social Is Dead: Yes, I know, you have read enough ‘social media is dead!!!111eleventy’ articles to last you a lifetime. I KNOW. That said, I can recommend adding this one to your repertoire – it makes the standard arguments, but I really like the way it couches them in market-based terms, specifically when it comes to the ‘over time, everything becomes broadcast media’ trend – this is a nice, cogent and eminently-sensical explanation: “Social as a model works when people have about as much to offer as they want to receive along a given axis. But no trait is distributed uniformly; there are are outliers in the nice-to-look-at, nice-to-listen-to, nice-to-read, nice-to-get-stock-tips from axes, there’s a population that can offer a respectable performance with these traits, and there’s a substantial majority with below-mean performance. So, over time, most platforms end up with a more consolidated list of suppliers and a dispersed set of consumers.” Basically speaking, you could subtitle this ‘why statistical distribution and probability mean that the creator economy was always a total fcuking lie’. If you want a more-anecdotal, less swivel-eyed-rationalist (sorry, but it’s true) take, you could read this piece in Dazed instead, which basically tells the same story via a series of personal anecdotes – the same point applies, though.
  • Evidence Maximalism: Or, “how the infinite quantity of information now available to anyone about anything has turned every single person on earth into one of those weirdos with the photowalls and the red string”, or “with enough datapoints you can prove literally anything”. The whole piece is good – Charlie Warzel is always readable – but the central premise is neatly encapsulated as follows: “all of the information online—news, research, historical documents, opinions—has conditioned people to treat everything as evidence that directly supports their ideological positions on any subject. He calls it the era of “evidence maximalism.” It’s how we argue online now, and why it’s harder than ever to build a shared reality.”
  • Noone’s Reading Anymore: This week’s ‘man, this really has been EVERYWHERE’ essay is from Slate, and is basically a teacher bemoaning the fact that none of his students appear capable of, or willing to, read anything longer than two pages of text in a critical or close manner – obviously I’m not a teacher (lol thank God) and I can’ only speak to my own experience, but IT’S NOT JUST KIDS, IS IT? I can’t tell you the amount of times in the past few years when I have done work for people and had them complain that the output runs to several pages of text – LOOK YOU FCUKING MORONS SOME THINGS SIMPLY AREN’T PARTICULARLY COMMUNICABLE VIA TRIANGLES AND ARROWS, AND SOME THINGS SIMPLY DON’T WORK AS DIAGRAMS, CAN YOU NOT TRY FCUKING READING YOU FCUKING LAZY CNUTS? Ahem. Your regular reminder that I’m available for freelance engagements, and am often described as ‘a pleasure to work with’.
  • What’s The Face Computer For?: As mentioned a bit earlier on, there’s been a flood of people who bought the Apple Vision Pro taking it back to stores as they realise that what they have actually bought is a piece of speculative technology with no practical purpose whatsoever – I’m still broadly bullish on AR/XR/WhatevR, but I think we’re a good few years away from the tech being small and lightweight enough to encourage the sort of mainstream adoption that will lead to the development of mainstream usecases.
  • The Robot Travel Agent: Or, “how people are turning to GPT to plan holidays and the like” – or “how to guarantee you’ll have a mathematically-average experience when you travel”, depending on your point of view. This piece features a selection of people waxing lyrical about how good GPT is at spitting out travel itineraries and planning trips – which is particularly interesting given the forthcoming likely release of AI ‘agents’ into the wild in the next year or so, but which also sort-of completely fails to acknowledge the massively-problematic elephant in the room here – to whit, because of the probabilistic nature of LLMs, when you ask it for suggestions like this IT IS GOING TO GIVE YOU THE MOST AVERAGE ONES. Which means that if everyone starts using these systems for these purposes, you can expect all the joyful side effects of TikTok tourism (endless queues! Price hikes! Environmental degradation! Really angry locals!) to expand exponentially. There is a CAMPAIGN IDEA buried in this for the right travel brand imho.
  • Recruiting With AI: This piece is all about how some companies have started using software by a company called Paradox AI as part of the employee screening process, and how that means prospective workers are being asked to undertake surreal, AI-generated personality tests whose answers (again assessed by AI) will determine a candidate’s suitability or otherwise for a given role. This is darkly-fascinating – the implication is that this is all black box stuff, so only the AI ‘knows’ what personal qualities some of the questions relate to, and how it scores them, and how that links to overall suitability, and that as such we’re in the process of handing over control of quite important things (cf, for example, ‘employment’) to  systems that operate in ways we simply cannot understand in any meaningful way. Does that sound good?
  • The GenZ Bone Spreadsheet: I have a sneaking suspicion that this is one of those trend pieces where it’s not actually a ‘trend’ at all – just because half-a-dozen sociopaths are doing something does not a ‘trend’ make, after all, although obviously it depends on who exactly said sociopaths in fact are – but I would be curious to hear whether any of you know of anyone who’s doing this; the ‘this’ in question being ‘keeping a spreadsheet tracking everyone you fcuk, ever, along with helpful notes to remind you of the experience’. Look, I’m not going to judge, but it’s important to remind you that the first person I ever heard of boasting about this sort of quantitative approach to life was one Milo YIannopoulous, and we all know how HE turned out.
  • AI Tablets and Educational Advantage: This is not wholly surprising, but it’s interesting to see it playingout already in real life – this is Rest of World on the scramble amongst Chinese parents to help their kids succeed in the already-ultracompetitive scholastic environment by ensuring they have access to the best AI-augmented learning assistants. If you don’t think a significant proportion of future social opportunities will be determined by the calibre of Machine Assistance you can afford then, well, you’re significantly more optimistic than I am (also lol, have you literally not been paying attention to ANY of the past century?!).
  • The Pope’s AI Advisor: A gorgeous profile of Paolo Benanti, the Cardinal who advises the Pope on matters pertaining to AI (and who, in the article’s most interesting casual asides, appears to have an interestingly-close relationship with Microsoft – honestly, fcuk my personal ethics and the rest, I would honestly give me right fcuking kidney to get a gig flogging AI solutions to the Holy See) and whose story of arriving at the faith is genuinely lovely.
  • Ricky Gervais: You will doubtless have opinions on Gervais and his humour and his standup, but regardless of what those opinions are I can highly recommend this review of his current ‘work in progress’ show in London – Rich Johnston at comics website Bleeding Cool does a really good job of unpacking why, to his mind at least, Gervais’ ‘edgy’ work doesn’t quite have the hard work underpinning it to make it justifiable. This really is a good piece of writing about comedy, and it’s a far more positive and balanced review than you might expect given the short explanation I’ve just given you.
  • The Trump Emails: This is LOVELY – a close stylistic reading of Donald J Trump’s campaign emails and their…unique style and cadence. “Where does one even begin? I guess we might as well start with the e-mail’s inexplicable ransom-note-style font decisions. Why is YOU capped but not italicized, whereas WE is capped and italicized? Why is the first paragraph bolded (and in red type), whereas the second is not? Why is “corporate death penalty” both scare-quoted and italicized? The typographical chaos mimics the legal, political, and psychic chaos in which Trump operates; and yet his relentless energy seems to emerge from this very chaos, as he paranoically and insistently narrates his woes in a kind of stream of consciousness, by turns slinging mud at the so-called haters, proclaiming his perseverance, and flattering and wheedling his supporters. He is Jesus on the cross, but he will survive!” This is honestly glorious, and I would personally happily read a close analysis of the differing styles of e-campaigning employed by the Tories and Labour in the hopefully-imminent election here in the UK/
  • Dr Alex Comfort: You will of course know the name from THE JOY OF SEX, a book whose title and homely illustrations of bearded 1970s coitus are weirdly familiar to seemingly everyone, despite the fact that I don’t think I’ve ever actually seen a copy of it in the wild, but you probably don’t know what a genuinely bizarre and polymathic person Dr Comfort was – this is a lovely profile in the LRB, effectively a review of a new biography of the man, and it is FULL of great details about his incredibly polymathy and commitment to alternative lifestyles – I loved this particular paragraph especially, with its uniquely-British combination of low-grade kink and very brown sadness: “And so the three of them entered into a secret arrangement where Comfort spent weeknights with Henderson and came home to his wife and Nick at the weekends. As Laursen concludes, ‘it was really a fairly ordinary affair, confined to a London flat and clearly destined to make all three participants unhappy.’ It was sustained not by a ‘commitment to some radical ideal of open relationship’ but by the participants’ Englishness. In Sex in Society (1963), published when the affair was a few years old, Comfort argued that monogamous marriage is best for childrearing, but that it was certainly possible to love more than one person at once, and that adultery could be a useful ‘prop’ to keep a married couple ‘on their feet’. Conveniently for him. At the time, Comfort took pride in his arrangement, which persisted for more than a decade. Not until 1994 did he admit to a journalist that ‘it didn’t work very well,’ and that both women were ‘in eruption the whole time’.”
  • Somewhere There’s Cheese: This is an essay written in tribute to the Aardman Animations classic ‘A Grand Day Out’, which first introduced the world to Wallace and the dog Gromit, and, honestly, it may be my favourite thing of the week – the love for the animation is palpable, and I promise you that you will find yourself actually, properly smiling as you read this. Really, truly wonderful writing by Zoe Kurland that communicates the very particular magic of the characters and the style quite perfectly.
  • The Sensitivity Reader: Our final longread of the week is a gorgeous and beautifully-crafted short by Andrew O’Hagen in Granta – I don’t want to spoil anything, so just trust me and read it (please).

By Jill Mulleady