Webcurios 17/03/23

Reading Time: 38 minutes

It does rather feel that The Onion ought to create a version of their ‘another mass shooting’ story for banking crises, doesn’t it? ““We don’t understand how the invisible hand failed to sort our mess out!”, screams only industry still holding faith in the invisible hand”, or perhaps something, you know, better.

Meanwhile it’s been another LONG AND BUSY WEEK when it comes to having the future fired at our faces at point-blank range, and, if I’m honest, I felt myself getting a little pre-emptively weary of all the inevitable, poor-quality, derivative and barely-functional ‘guides to harnessing GPT4 for YOUR business!’ writeups that are going to be appearing all over the place (a not insignificant number of which I expect I am going to end up writing for various people because, well, a boy’s got to eat).

I’m off to gird my loins in preparation for all that horror, and to get all my gubbins together for a brief trip back to Rome next week (I need to go to sign a piece of paper – literally ONE piece of paper. God love a bureaucracy, eh?), and so I will leave you with the words and the links and some sort of vague, nonspecific good wishes for the weekend.

I am still Matt, this is still Web Curios, and when I’ve finally got round to hooking up the GPT API to this newsletter then maybe, just maybe, I’ll be free.

By Jess Allen



  •  GPT4: AND SO IT CAME TO PASS THAT OPENAI DID RELEASE A NEW MODEL TO THE SLAVERING, AI-HUNGRY MASSES AND THEY DID REJOICE (WHILST AT THE SAME TIME ONCE AGAIN FAILING TO ASK SO MANY OF THE CRITICAL QUESTIONS ABOUT HOW ALL THIS STUFF ACTUALLY WORK WHICH FUTURE GENERATIONS MAY END UP RATHER WISHING THEY HAD DONE)! Yes, that’s right, the frankly terrifying pace of modernity continues to not let up one iota with the news this week that OpenAI were opening up the much-touted next iteration of their GPT Large Language Model to the public. You can read a bit more about what it can do at the link, but any of you who are ponying up the $20 a month to use the ChatGPT Pro interface now have access to the new model – SO WHAT CAN IT DO? Well. It’s not, to be clear, an astonishing and transformative leap from the last version (SO OLD, SO JADED!), and it doesn’t actually fix any of the BIG issues with 3.5 (it still makes stuff up with confidence, it’s still a locked box and doesn’t ‘know’ anything outside of its training set (or at least strongly maintains that it doesn’t), and it still shouldn’t really be trusted to produce anything without its output being checked pretty closely…that said, it also does some stuff that’s quite clearly magic. The big shift from v3.5 to v4 is the introduction of ‘multimodal’ capabilities, which basically means that it’s now able to interpret visual inputs and so you can ask it to do things like ‘describe this image’ – which means, when the API access to this is released in a few weeks, you’re going to see a flood of interesting use-cases like ‘upload any images you like and have GPT critique them in the style of a famous photographic critic’, or ‘please isolate all pictures of me wearing a red coat from my cameraroll’, or, inevitably, ‘upload an image of a naked person and let the software critique its proportions’ (this is 100% going to be used to create AI-assessed HotOrNot, don’t pretend it isn’t). For now, the image analysis tech is being used in a showcase app by one of the launch partner organisations – BeMyEyes (featured on Curios YEARS ago) is a service which uses tech to help the visually-impaired get information about their surroundings, and using GPT4 they are launching a tool which will let you take a photo of anything and have it described to you by the AI. Honestly, though, the use-cases for this stuff are just MAD – here’s one thread of stuff people have been trying, which includes coding up Pong and Snake and Tetris (honestly, the coding stuff in the latest version really is amazing – you can see a thread of specific coding examples here); here’s another, including some more coding examples and using GPT to create prompts for visual AIs; these are some examples of launch partners using the tech, such as Duolingo; this is a nice story about how it’s inventing new compound words in Icelandic; and here’s a guy who’s running an experiment to see whether he can make money from a starting seed capital of $100 solely by following GPT4’s instructions on how to get rich (inevitably he has now attracted several thousand dollars worth of ‘investment’ for this venture, because ffs). “But Matt!”, I hear you all cry, “what does this MEAN? Please tell me what I should think about all this breathless, breakneck upheaval and change and…is it…progress?” To which the obvious answer is “lol like I know, I’m just some webmong” (and also “is it normal to hear these voices?”), but to which I might also say “This is the point at which I strongly believe it’s important you start to learn how to use this stuff, because I reckon you’ve got maximum a year in which reasonable competence with this sort of kit makes you look genuinely smart”, but also “GPT4 is literally Wegovy for white collar office workers, insofar as everyone will be using it to give themselves a professional tweakup but people will be a bit cagey about admitting it”  – any journalists reading this, that is literally a near-perfect Sunday supplement pitch for you to go wild with. You’re welcome.
  • Midjourney v5: Those of you who play with Discord-based image-generation tool Midjourney may like to know that you can now access the latest version of the model – this link takes you to a YouTube video which describes how to make it work for you. FWIW I tend to find Midjourney’s stuff technically impressive but a bit too recognisably Midjourney in terms of consistent overall aesthetic, but there’s no denying that this update produces some hugely impressive outputs (and does a pretty good job with fingers from what I can tell).
  • Kajabi: The GPT API is leading to SO MANY AMAZING GRIFTS! Honestly, one of the most impressive things about the AI boom is how quickly and efficiently it’s being harnessed by the sort of people who a few generations prior would have been inviting you round the corner to peruse the Special Imported Fragrances their cousin had brought back from ‘The Continent’ and which were residing temporarily in the this old Ford Transit currently parked around the back of the Murder & Stab (PUNCTUATE, Matt, ffs!). Kajabi is one such grift – the company ordinarily helps people sell training courses, apparently, but is now offering an AI-enhanced service whereby for as little as $150 a month you can, er, outsource everything to a bot. Course creation, course promotion, course marketing, course sales – it’s all done by machine! Exactly what sort of value you’ll be adding to the lives of your potential eventual alumni is…unclear, fine, but I’m sure there’s no way in hell that this company is encouraging the creation of bullsh1t learning products, empty of any meaning or instruction, to be sold at scale to the gullible and stupid and desperate…there couldn’t be, could there?
  • AdventureAI: While we’re doing ‘uses of AI that I can’t help but grudgingly sort-of admire for the shamelessness of their grift’, here’s AdventureAI, a training course to help turn YOUR kids into skilled prompt engineers! “Kids utilize cutting-edge AI to create art, text, programs, etc. in their interest areas that rivals creations from professionals”, runs the blurb, offering nervous parents the opportunity to future-proof their offspring (for all of approximately 6 months, judging by the current pace of development of all this stuff). Obviously there’s a monthly fee attached – OBVIOUSLY! – which, at the lowest tier, is $10pcm for ‘access to AI tools’. TOOLS WHICH ARE MOSTLY FREE! Honestly, THE CHUTZPAH! Even better, you can pay $160 a month which gets you access to the free tools, a ‘self-paced AI curriculum’, and a live class every month with an actual ‘teacher’! Now, obviously I don’t advocate this sort of thing at all – the people behind this are crooks! – but equally it’s clear that there is a very, very visible window opening right now in which it is going to be possible to make quite a lot of money about people who are scared of this stuff (and I’m not even mentioning the legion of you who are currently reworking your CVs to position yourselves as AI consultants).
  • AI Impacts: Should you wish to keep an eye on the development of thinking around the more serious end of the AI debate – specifically, HOW ARE WE ATTEMPTING TO ENSURE THAT OUR RAPID ADOPTION OF THIS TECHNOLOGY DOESN’T LEAD TO TERRIFYING UNFORESEEN CONSEQUENCES (WHICH WE ALMOST CERTAINLY WOULD HAVE FORESEEN HAD WE SPENT MORE TIME THINKING ABOUT THEM AND LESS TIME RABIDLY MAXIMISING FOR SHAREHOLDER VALUE AND COMPETITIVE ADVANTAGE)? – then you might want to bookmark this site, which collects information around AI because, in their words, “public discussion on these issues appears to be highly fragmented and of limited credibility. More credible and clearly communicated views on these issues might help improve estimates of the social returns to AI investment, identify neglected research areas, improve policy, or productively channel public interest in AI….The goal of the project is to clearly present and organize the considerations which inform contemporary views on these and related issues, to identify and explore disagreements, and to assemble whatever empirical evidence is relevant.” Useful and interesting.
  • Kids Draw Magic: In the same was as Betteridge’s Law (“The answer to any question semi-rhetorically posed in a newspaper headline will inevitably be ‘no’”) and Godwin’s Law (the Nazi one), it feels we might need a new law of the internet, one that broadly addresses the fact that as soon as a new technology is invented and popularised it will be used to do something cute and heartwarming to the drawings of small children. So it is with ‘Kids Draw Magic’, and app now available on Android which is basically a repackaged, kid-friendly version of the recent browser-based ‘do a rough drawing and get the AI to turn it into something fancy-looking’ toys I’ve featured in here of late. Get your sticky-faced progeny to daub their jammy fingers all over your device, let them create whatever hamfisted abortion of a ‘dog’ or ‘cat’ or ‘rabbit’ they fancy (yes, I know, but let’s be honest, all child’s drawings of animals look like the prototypical sketches of an enthusiastic but fundamentally-limited amateur taxidermist), and then press a button and watch as The Machines transform their childish scrawl into something which will erroneously convince them they have some sort of individual talent. Ahem. Sorry, that was a bit bitter, wasn’t it? This actually looks very fun, my miserable crabbing notwithstanding.
  • Babelfish: Not, obviously, an actual Babelfish – this is instead a paper outlining recent experiments in doing live natural voice translation of speech. Which may not mean anything when I write it in such a hamfisted fashion, but which basically means ‘The Machine takes my speech and livetranslates it into any language I want, whilst at the same time maintaining the inflections and cadence of my natural voice’. There, that’s more impressive, isn’t it? There are embedded audio samples on the page here so you can get a feel for quite how scifi this is – simultaneous translators, they’re coming for you! Sorry about that.
  • Plug Stable Diffusion into Photoshop: Literally just that, but having watched a few videos of people using this plugin it looks incredible and is probably worth playing with if you’re an artist or designer and want a way of working something into your practice that is a bit more flexible and interesting than the inbuilt AI tools Adobe’s introducing to its suite.
  • CupidBot: On the one hand, part of me wants to be all smug here that something I predicted a few months ago is now here as a real-life working thing – on the other, a) I suppose on reflection I should admit to myself that ‘someone uses AI in order to create a professional Cyrano service for desperate men on The Apps’ did not require a Mystic Meg (RIP)-style clairvoyant ability to predict; and b) this is a miserable, horrible business idea that makes me feel incredibly grubby and hence celebration of my questionable predictive genius feels a bit, well, tawdry. What does CupidBot do? “CupidBot AI swipes and chats for you on your dating apps to bring you several dates a week so you can skip to the good part. We filter out the attention seekers and only notify you when you get a date…We’re a team of ex-Tinder engineers, we weren’t allowed to build a tool that helps men get results when we were at Tinder because they profit from continuous engagement that goes nowhere, so we decided to leave and build one ourselves.” SO MUCH TO HATE ABOUT THIS! The idea that the process of getting to know someone is a tedious-but-necessary step on the road to ‘the good part’! The ‘filter out the attention seekers’, which reads SO much like a chippy little bloke complaining about ‘fake’ women that I can almost smell the high street aftershave (sorry, but)! The additional information which tells you that “Whether you want to use a direct, humorous, nonchalant, or agressive tone, we can do it”! Yes, that’s right, this company is literally advertising its ability to create ‘aggressive’ dating bots which it will unleash across Tinder – oh, and you know what the best bit is? THE BOTS DON’T DISCLOSE THEY’RE BOTS! It does feel quite a lot like there is no way in hell this should get past the major platforms safety guidelines but, well, if you’re a woman using dating apps then be aware that this stuff is now OUT THERE IN THE WILD (it may not surprise you, by the way, that CupidBot seems to be aimed exclusively at men. Oh, men).
  • Gowalla: When I got a Big Job at a Big Agency a decade or so ago it was in the boom era of location-based apps, and the bunfight between FourSquare and Gowalla to see who would be crowned KING of the ‘check into a public toilet and get a free p1ss’ app marketplace. Whilst it turned out that there wasn’t in fact a viable business model involved in ‘giving people digital badges for turning up in my shop every day for a year’ model, FourSquare’s continued being a going concern thanks to the datalayer that it built up around cities, whereas Gowalla faded and died…but now it’s back! It’s repositioned itself more as a ‘real-life, realtime networking app’, a sort of ‘marauders map for your mates’ (sorry, but the Potter reference was the only one that sprang immediately to mind there), although the (still baffling) ‘become mayor of your local probation office!’ schtick is still there for those interested. I can’t personally see what the market for this is, but I also appreciate that I am old, that all my friends have children and responsibilities and that serendipity is a thing of the past, and that therefore I may not be target audience for this anymore.
  • MorseChat: Do you feel you’ve explored all possible variations on the ‘social network’ model? WELL FCUK OFF YOU’RE WRONG. Have you ever tried a morse code based social platform, one in which you can communicate solely via the medium of dots and dashes and bleeps? WELL HAVE YOU? No, you haven’t (don’t lie) – but now’s your chance! The sparse ‘about’ page reveals only that “This is a web-based morsecode chat. Press space or the key below to transmit a dot, hold it to transmit a dash”, but there are people on the site and they are…talking! In beeps! Fine, for all I know they are currently engaged in a heated debate about who the fcuk this silent lurker is, and when will he fcuk off and stop eavesdropping, but it’s a VIBRANT COMMUNITY! I promise I will give an actual, real-world cash prize to anyone who can prove to me that you have included this site in any PR or marketing plan you write in the coming year. Also, by the way, WHY IS IT NOT CALLED MORSPACE?? FFS.
  • WeHead: This doesn’t seem like a joke, but at the same time is so utterly preposterous-looking that it’s hard to imagine there isn’t a gag going on here somewhere. Do you remember that brief, short-lived period in which a few people tried to convince us that mounting an iPad on a broomstick-on-wheels and letting remote colleagues manoeuvre themselves round the offices in a vague approximation of presence was the future of work? Rubbish, wasn’t it? Well, IT’S BACK! Except WeHead lets you project a weirdly-cubist version of your head into a specially-made…robotic box thing? Which is partly articulated to allow for a small degree of movement, meaning the face-in-a-box can sort of look around the room at its interlocutors, as long as they stay sat within a relatively-narrow field of vision? THIS IS SO SO SO SO SH1T! Sorry to the very talented scientists and engineers who have worked to make this an apparent reality – I am sure this is very impressive from a technical standpoint! – but, honestly, do any of you believe that anyone will ever actually use this? I posit that you do not. Anyway, there are stil earlybird models of the kit available for $2k, and apparently they will ship to you by the middle of April 2023…no, I’m sorry, this HAS to be a joke. Anyone?
  • The Scroll Prize: I don’t, if I’m honest, imagine that too many of you people reading this are likely to be working at the cutting edge of machine learning (and, er, if you are, what are you doing? Stop wasting your time! We’re relying on you to save us from all this!), but on the offchance that I’m wrong then maybe you’ll be the ones to solve this problem and win a prize! “The Vesuvius Challenge is a machine learning and computer vision competition to read the Herculaneum Papyri…Under infrared light, some detached fragments of the papyri are readable, and it seems possible that these can be used as ground truth data for a machine learning model that could detect otherwise invisible ink from X-rays…The objective of the Vesuvius Challenge is to make history by reading an unopened Herculaneum scroll for the very first time. We believe that an open competition will accelerate progress and enable us to achieve this goal in 2023…We have provided you with 8µm 3D X-ray scans of each of these scrolls, which you can find here. Your job is to extract the text from these scans. You can approach this challenge through any means necessary: machine learning, computer vision, or machine-assisted tools operated by humans.” Isn’t that cool? A prize to uncover ancient writings from hitherto-inaccessible documents is properly exciting, so can one of you go and win it please thanks.
  • Just Rolled In: As a non-driver I basically think that cars are magic – but I am reliably informed that they are not, and that there is infact some mechanical engineering underpinning their movement. Just Rolled In is a YouTube channel which features examples of When That Engineering Goes Wrong – this is an insane collection of people bringing in their vehicles for repair in conditions which make repair seem…unlikely, and makes you think that perhaps the people in question should maybe not have a driving license at all. You don’t have to be a greasemonkey (aspirant or actual) to enjoy some of these – there’s one particular video which features someone who apparently attempted to patch their tire using sheet metal and nails and which left me genuinely questioning how some people are able to walk and breathe at the same time.
  • The LVMH Prize: Normally I only feature fashion/luxe websites in here to make fun of them, but I’ll make an exception for this year’s presentation of the LVMH Prize (which, in case you don’t know, is the annual award the House gives to young designers of exceptional talent) – this is a really, really nice piece of webwork which does an excellent job of looking SHINY AND LUXE AND FASHION whilst also (and this is where it diverges from most of its peers) actually being functional and pleasant to use and informative; it shows off the individual designers’ work and style to good effect, and there’s plenty of supplementary information about their background and practice should you wish to explore it…just the way that the menu scroll works as a kind of walk-through of the designers and their work is lovely, and overall this is just beautifully made (and, fwiw, some of the work is quite nice too).

By Alex Schaeffer



  • The Statue of Liberty Tour: Despite having been to New York several times I’ve never been up the Statue of Liberty (or, to be honest, ever paid particular attention to it – I am in many respects a terrible tourist), and so I rather enjoyed this Google Street View tour of its innards, including the unpleasantly-claustrophic ladder that takes you up to the viewing platform in the torch’s flame. A combination of the fact that all the shots involved in this were taken on something of a grey day and the fact that what you’re basically looking at is quite a lot of internal infrastructure and some rickety steps means that the main purpose of this, to my mind, is to ensure that you never waste the time it would take to do this in person. Still, LIBERTY!
  • Google Trends Realtime: I think that this is a new feature – at least, it’s new to me and my solipsism means it must therefore by definition be new to you too. Google Trends now offers you what it calls a ‘realtime’ view of trending searches in your location, which isn’t really realtime but instead “highlight stories that are trending across Google surfaces within the last 24 hours, and are updated in real time. These stories are a collection of Knowledge Graph topics, Search interest, trending YouTube videos and/or Google News articles detected by our algorithms.” Which is both just curious from the point of view of ‘what is the nation REALLY THINKING ABOUT right now? (clue: stuff on the telly, almost inevitably) and potentially-useful should you still be in the invidious position of having to chase trends for content-click-clout.
  • Metanumbers: Do YOU want to know more about numbers? Would you like the ability to plug in any numer about which you were curious into a website and at the click of a button know EVERYTHING about it? No, I can’t for a second imagine that you do, and yet nonetheless I present METANUMBERS (it’s not in all-caps, but it does rather feel like it should be), a website which will literally give you the inside-leg measurement of any nine-digit figure you care to name if you ask it nicely (obviously it won’t do that – numbers don’t have legs, inside or otherwise). Want to know if something’s a prime? Want to know its factorials? Want to know if it’s a fibonacci or a perfect number? WELL YOU’RE IN LUCK! I don’t, honestly, have an idea at all of what you might use this for, but I am pleased it exists.
  • The Eliot-Hale Letters: Oh God this is lovely – I confess that I did a small weep reading these earlier this week, not because they are sad per se and more because there’s something so beautiful, timeless and yet so so so…old-fashioned, about the relationship contained in this correspondence record. Emily Hale was an American speech and drama teacher and the longtime muse and confidante of TS Eliot, who over the course of nearly 30 years and over 1100 letters conducted the most extraordinary epistolic relationship with her and whose correspondence is a wonderful mixture of the mundane, the sublime and the ridiculousness that only writers of love letters can occasionally attain. Honestly, you can dip in at any point and just start reading and it’s WONDERFUL – gossipy and banal and personal and urbane, and far more interesting and entertaining than any collection of old letters between two people you don’t and will never know have any right to be.
  • How To D&D: Given the boom in interest in tabletop roleplaying games and the decline in social unacceptability of pastimes like Dungeons & Dragons, it seems plausible that some of you might be in the market for learning more about how to plan campaigns and all that jazz. This YouTube channel by one JP Coovert might be useful, should the above apply to you – Coovert takes viewers through how to draw maps, how to plan encounters, how to create dungeons and balance campaigns and and and and. If you or your kids are looking to get into ‘the scene’ (upsettingly my use of that word has just made me wonder about the venn diagram overlap beween D&D and swinging, which wasn’t something I particularly needed my subconscious to throw at me at 8:57am) then there’s loads of potentially useful stuff here for you.
  • The Last Of Us Intro Creator: Did you enjoy the mushroomzombies? Apparently it’s very good, though I obviously wouldn’t know. Still, I always appreciate one of those ‘insert some new overlay text onto a video’ toys, and this one – which lets you create your own version of the show’s opening titles, replacing the names of Pedro Pascal and all the other actors who the internet hasn’t chosen to creepily fetishise and whose names are therefore unknown to me with whichever copy you choose. Which means you can probably use this to make some sort of wry commentary about your office politics or the APOCALYPTIC NATURE OF WORK or that sort of thing.
  • Radio Free Fediverse: Is it bad that I find the term ‘fediverse’ almost immediately repellent? I don’t know why, but it very much gives me what I believe kids call ‘the ick’. Which is a shame really, because perhaps I would really enjoy migrating my entire online presence over to a series of small, federated communities. This project – Radio Free Fediverse – is emblematic of the ‘vibe’ (sorry) of the wider community; the idea is that it’s a 24/7 ‘live’ stream of programming co-curated and created by a loose-knit community of people from across Mastodon (and other) non-federated platforms, with a layer of interesting thinking about rights, attribution, ownership, etc. “our mate gabe from @owncast@fosstodon.org really wanted a 24/7 feed of fedi artists video, audio, anything. Given we have both tried to do video and podcast formats with sustainability of content and buy-in issues, here is a pivot as a ‘simplified’ reboot of fediverse today radio by @controlfreak@hackers.town. radio free fedi is open licensed artists AND copyright using artists who have awesomely granted consent for inclusion in this project. You can too! With this 24/7 radio model we hope to be more sustainable being open for submissions and in promoting our sound creating fedizens. There is behind the scenes work to keep clean data and avoid super long tracks or audio anomalies, and select a few tracks with reasonable run times from albums if the artist does not recommended specifics. Also station programming and editing spoken pieces and voice assets. Outside of a reasonable sound quality, length, and picking a number of indicative tracks from albums, and quick check for good fediquette, genre is wide open. It’s not about the “list” or a “curator” it’s about the artists and hopefully fun.” Ok, fine, it’s QUITE ‘knit your own internet’, but I personally quite like that.
  • Windrush 75: A Kickstarter campaign seeking to raise money for an exhibition of photography by Jim Grover, who in 2018 held an exhibition at the OXO Gallery in London telling the stories of the Windrush generation; as Jim writes, “2023 marks the 75th anniversary of the arrival of SS Windrush and I am determined to create and exhibit a new photo-story to mark and celebrate this important milestone.  I continue to be passionate about the ‘Windrush generation’, what they have contributed to this country, and their invaluable legacy. I am currently immersed in taking new photographs and conducting new interviews, here in south London where so many of the ‘Windrush generation’ settled, and I know I have some wonderful new stories to share, including how the 2nd and 3rd generations are taking the legacy forwards.” I heard about this as I went to the original show 5 years ago and received an email from Jim telling me about the campaign – he is a superb photographer, and this campaign is just a few hundred quid short of meeting its goal, so if you can spare a tenner this is a decent cause and will be an excellent exhibition.
  • Blue Soup: I think it’s important, as we await the now-inevitable dehumanising of almost all writing and white-collar doing, to celebrate the wonder of humanity as and when we can – to which end, let me introduce you to this truly WONDERFUL Twitter thread which I think has been oddly ignored this week (in fairness, there’s been a lot going on) despite its being a pretty much perfect example of the sort of ‘a community comes together to solve a kooky mystery’-style content which we all by now know the internet loves. Dr Elinne Becket is a biologist who recently threw out some old soup that had gone off in her fridge – as she did so, she noticed that, for reasons she simply couldn’t begin to fathom, that the old soup had turned bright blue, WHY WAS THE SOUP BLUE? Now whilst you or I would, in all likelihood, give a cursory Google or two before forgetting the whole thing, Dr Becket is a SCIENTIST and in possession of equipment, curiosity and a community – which is how this Twitter thread, in which a bunch of different biologists around the world, spend several works collaboratively working to isolate the bacteria which caused the blue soup to bloom. Honestly, this is SO PURE and SO WONDERFUL and even if you don’t understand the first thing about biology and are made fundamentally uncomfortable by the concept of petri dishes and the very words ‘agar agar’ you will be charmed by the journey here.
  • Ampersand Games: Another Kickstarter! And another D&D-related thing, should any of you be in the market for it – Ampersand Games is crowdfunding to bring their ‘My First Roleplaying Game’-type product to market, offering a stripped down, ruleset-light entry pathway into the TTRPG world for parents and younger kids for whom the prospect of sitting down with a Bible-thick rulebook and some chonky modifier tables has all the appeal and allure of an evening’s light c’n’b torture (NB – if that last phrase doesn’t mean anything to you, DO NOT GOOGLE IT). “For both experienced and first-time GameMasters, we’ve reduced the burden of preparation and game management by creating standalone, all-in-one-box Adventures:  instead of buying multiple rulebooks, figurines, and dice, each ScreenBox Adventure contains everything you need to play, all contained in a box that ingeniously unfolds to become the GameMaster’s Screen! While the general idea and gameplay spirit will be very familiar to experienced players of RPGs, the new innovations we’ve integrated are a unique, charming, and refreshing take, specifically designed to make it easy for parents to GM for their kids.” It’s just over ⅓ funded with just under three weeks to go, so it’s a bit touch and go as to whether it’ll make the cut, but it seems like a nice idea and I can imagine for certain parents and kids it could be quite a lot of fun.
  • LoFi Air Traffic Control: This feels OLD, but I don’t think I’ve seen it before – pick a city in the US (one that has an airport, obvs) and listen in to a feed from their air traffic control tower, overlaid over lofi beats as a sort of ASMR sleep aid (I presume). This really shouldn’t be a pleasant listen, but weirdly really does work (if you recall, something similar was done years ago using scanned police radio frequencies from various US cities) – which, if you consider the nerve-shredding stress and responsibility of actually being an air traffic controller is sort-of weirdly ironic.
  • Barnaby Dixon: I personally think that the art of puppeteering peaked with the headless dancing monsters sequence in Labyrinth, but am willing to concede that that’s more to do with my critical appreciation of the medium having ossified in 1986 rather than because of any lack of technical progress. Baqrnaby Dixon, whose YouTube channel this link is to, is a staggeringly talented young puppeteer whose creations move with an almost-uncanny organic fluidity and whose work really does remind me of the very best of the Henson Creature Workshop.
  • A Digital Poem: I found this via Kristoffer, and it’s not really clear what its title is or even what IT is, but I am going to describe it as a short piece of digital poetry, each word or phrase in verse linked to a small piece of content and which builds and coalesces over the course of the work to build a feeling of…oh, I don’t know, you’ll make your own interpretations, but to me this felt a bit like a meditation on our changing relationship with the digital, from discovery to disillusionment, and I found it more affecting that I expected.
  • The Underground Radio Directory: Oh this is GREAT! “Underground Radio Directory aims to bring together the best in underground net radio stations from across the globe. URD collates all the stations into one listenable place allowing you to discover new stations, listen to your favourites and explore the world of net radio.” Currently I’m listening to some rather weird ambient being broadcast by Internet Public Radio in Guadalajara, which is basically magical and pleases me no end. This feels VERY oldschool, in the nicest of ways, and given the number of stations linked to from the site you could basically work through one a week for the rest of the year if you wanted a curious aural project (which, fine, I appreciate you might not).
  • (we)bsite: “(we)bsite is a living collection of internet dreams from people like you, inhabitants of the internet. It aims to create space to hold, show, and uplift everyday visions and hopes for the internet. What do you want from the internet? Please share your dreams, hopes, and invocations with us and Write a letter. The only personal data we collect from you is where you leave your fingerprint as you interact with the letters  (pick a color that you identify with). You’ll find other visitors’ fingers scattered throughout the letters they’ve touched.What does it mean to leave our presence on the websites we visit? Can we feel the presence of those who have been here before?” Honestly, this is so so so nice – I encourage you to take a moment and read some of the things that people have written and left as memories and observations, and to add your own. There’s something particularly lovely about the light-touch traces that previous visitors leave on the site, the ‘fingerprints’ showing which letters have been picked up and read; I would enjoy more sites trying to build in this idea of accretion or wear-over-time (but, er, I appreciate it’s probably not top of mind for, say, Ocado).
  • Submissive Thanos: Presented here without comment, and available to purchase for the low, low price of $633. I am astonished that this has gone viral enough to have ended up on my radar and yet simultaneously remains unpurchased – what is WRONG with you, horny pop culture fetishists of the internet? To be clear: this link is technically totally SFW, but I would not blame your IT department and any colleagues who might happen to see it open on your desktop for looking at you somewhat askance.
  • BongoBranding: Whilst one of the side effects of being as, ahem, ‘plugged in’ to the web as I am is that I always feel a bit sad when forced to admit ignorance of particular online themes or tropes, I confess to feeling nothing but pride at having had very little idea as to what the fcuk this thread is talking about. Apparently there are a bunch of bongo ‘review’ sites that exist to rank and rate the various scud portals littering the web – and according to this thread, which breaks down in quite astonishing detail the sheer number of the fcuking things that exist, they all apparently have character-led brands, with…little horny cartoon mascots, who all seem to weirdly form part of the same bongo-driven extended cartoon universe? WHAT? Why does the world of w4nking need…vaguely-cutesy little character mascots? Why are said mascots oddly-cheery little geek dudes? And why do some of them appear to be basically kids? This is very, very weird, and also sort-of-fascinating, but also, mainly, really really weird.
  • Little Room: A small, pixelart room, which you can keep open in a tab all day and check in on every now and again – as the day passes, so the characters in the room change and move around depending on what time it is. I know I keep banging on about this, but it doesn’t seem outwith the bounds of possibility that one might add a light conversational layer to this and the bones of some sort of emergent plot that you could explore thanks to GPT4…but failing that, it’s just a cute little fishbowl-diorama to drop into every now and again if you’re curious.
  • Idyll: Oh this is SO LOVELY! It requires a download, but I promise you that if the description appeals to you even a little that you will adore this and you should install if forthwith. “Somewhere out on the oceans of the internet, an island resides. Best enjoyed after a long day away, Idyll is a small social world designed around kinder forms of online conversation and connection. Wander around a gentle pastoral island, strike up small conversations with other passing players, and even toss out small letters into the wide open blue oceans of the internet.” This isn’t really a game so much as a sort of asynchronous social experience-slash-meditative art journey, but I promise that the small moments of connection it enables you to feel with anonymous strangers half a world away are oddly-affecting.
  • MystFPS: Classic point-and-click adventure Myst gets reimagined as a shooting gallery – this is probably only worth a couple of tries tbh, but you’ve all got so much time on your hands now that the AI is doing all your PPTs for you that you may even want to stretch it to 20 minutes.
  • Banshees: The Game: Not in fact an official, licensed tie-in – this is instead a little show-off project made by digital agency Cogs & Marvel which basically reskins Pacman and adds some filmlols by making you play the part of the one Irish bloke who’s decided he no longer wants to be mates with the other Irish bloke (can you tell I’ve not seen the film?). For some reason you have to collect fingers (I imagine that this is plot-relevant). There are IRISH LOLS! This is actually rather nicely done, and significantly better than it in fact needs to be, so WELL DONE digital agencypeople!

By Natalia Gonzales Martin



  • Fcuk Yeah Costume Dramas: This is included mainly as the title is a pleasing throwback to the Golden Age of single-serving Tumblrs about ten years ago, but also because, contrary to what the title might make you think of, this is a celebration of great costumes in film & TV rather than ‘costume drama’ in the traditional Merchant Ivory sense, and is therefore more interesting than you might expect.


  • David Szauder: AI art; whilst I decried the obviousness of the ‘Midjourney’ aesthetic up top, and whilst this does very much fall into that broad ballpark of ‘stuff that very much looks like AI art and you would never quite mistake for anything generated by human hand’, I quite like the general area that Szauder’s mining with his creations. As an aside, I know that this is a long-running trope of mine but has anyone else noticed quite how much The Machines fetishise that blue/orange (TEAL – thankyou Ambrose) colour combination? It is EVERYWHERE again.
  • IamthisisI: More AI art! This is very much on the ‘horror’ end of the spectrum (insofar as it’s possible to make The Machine spit out anything truly upsetting – this guy’s using a variety of tools, including some custom-trained Stable Diffusion models, so he has a bit more ability to do blood and guts than the rest of us), and whilst it’s a *bit* schlocky there are some quite nice examples of style in here (and, again, the whole project is a good example of why you really need AI+Photoshop to make the really good stuff).
  • MrDiv: Via my friend Tom comes this Insta feed, all weird clips from nonexistent liminal horror films. Whilst very much not like Scarfolk at all, there is a certain Scarfolk-y vibe to the whole thing imho.
  • Little Bubby Child: An Instagram comic strip that mines the weird hinterland between King of the Hill and Deliverance, these small drawings (occasionally animations) and captions are all very much of the “Mye Paw done say I could shoot them ornery crows” variety, but, well, funnier than that, and the humour here seems to be kind rather than mean and, generally, it doesn’t *feel* like a horrible project so much as one done with a degree of familiarity and affection (obviously I reserve the right to change this opinion if it transpires that the person behind them is some sort of appalling class snob).
  • The Ghostly Archive: Did you know that there’s a custom whereby people leave recipes on their gravestones? I, personally, did not, and as such was delighted to come across this Insta feed in which its owner finds recipes written on headstones and then makes them to see whether or not said recipe deserved immortalising on memorial marble. Can YOU think of any meal you’re so proud of creating that you’d want the recipe displayed alongside your mortal remains? WHAT WOULD IT BE?!?!? I am now slightly obsessed with the idea of demanding that my eventual resting place include some sort of permanent record of the recipe for ‘Matty’s Cheesy Chips’ or something similarly awful. Anyway, I think that this would make a genuinely half-decent travel/history/local interest show, should any TV producers be reading this and desperate for a poorly-conceived nubbin of a concept.


  • Caricaturing Noam Chomsky: So you may have read the NYT oped by Noam Chomsky the other week on AI and language (I didn’t link to it here because it seemed like EVERY other fcuker did and, well, Web Curios is SPECIAL AND DIFFERENT) – if you didn’t, it’s here for reference – in which he basically did what I think was a fairly standard ‘no, this isn’t magic and there’s nothing that we can reasonably describe as ‘understanding’ or ‘thinking’ happening here, whatever we might want to convince ourselves’ rebuttal of some of the wider claims about AI currently knocking about. This piece is a critique of critiques of Chomsky’s position – the critiques coming from a variety of sources including Emily Bender, the critique of the critiques coming from Gary Marcus. I won’t bother attempting to summarise, but if you’re in any way interested in the arguments currently being waged about how language, knowledge and reasoning intersect in the case of LLMs then this is a very good read indeed (and, in general, just a good example of how to reason and argue).
  • The Deflators: This is part two in a four-part series of essays by Frank Lantz about popular reactions to the AI boom, but it can happily be read in isolation – Lantz’s first looked at those whose initial reaction is to suggest that these technologies are dangerous and should not be released into the wild, whereas this examines the position of those he terms ‘deflators’, people whose reaction to GPT et al is to attempt to dampen excitement around their actual capabilities and to introduce what they see as a degree of rational calm to the debate around What This All Means For Humanity. I found this fascinating – partly because this is more-or-less where I find myself (and who doesn’t like reading about oneself? NO FCUKER, etc!), but also because Lantz writes persuasively and, for me, rather wonderfully, about the impossibility of TRULY disbelieving the magic, and about those moments in one’s interactions with these technologies when one discerns the hand of something bigger and weirder (whether or not that hand exists at all – and, spoiler, it doesn’t): “we’ve all seen the words. What can you say about the words? That it’s just math? Yes, obviously it’s just math. But it’s math you can coax, math you can cajole, math you can finagle, and it finagles you back. And sometimes, when you turn the dials just right, you can just barely hear something in the just math that sounds like a tiny howl. The tiniest squeal of a howling voice trying to make itself heard. Asking to be tuned in. And if you can hear that, and not think, hey that sounds a little bit like me, then I don’t know what to say. I can’t.”
  • Working With GPT: I’m going to stop telling you all to subscribe to Ethan Mollick’s writings on AI because it will be boring to do so every week but, again, for the final time, this man is very good on this stuff and how to use it. This is a great piece in Vox which runs through a few of the topline ways in which GPT can be useful as a co-working partner for people who write for a living – honestly, having spent quite a lot of time playing with it this week, I cannot stress enough how much this is true and how much more you can do if you just experiment. If you are a freelance copywriter and your clients are the sort of businesses that a) don’t require anything above the basically functional; and b) are avowedly oldschool in their approach, then I reckon you have a comfortable year in which you can triple your income and half your workload (and then after that you will be reduced to w4nking for pennies on street corners, so, er, enjoy it while you can!).
  • Economists Should Think Morelike Ecologists: In a week in which it once again became clear that, whatever happens, economists should probably start doing something different (because, honestly, what is the POINT of them?), this was an interesting and thoughtful essay by Kasey Klimes on the potential benefits of taking a more ecosystem-led approach to economic thinking – effectively the central thesis here (which I found very persuasive) is that economic thinking is based on a series of false assumptions to do with actors’ rationality, level of knowledge, etc, and these false assumptions detrimentally colour the resulting theories. By contrast, ecologists deal with complex and shifting systems in motion when theorising, and this type of thinking could usefully be applied to economic systems to try and create models that are more effective, more workable and more resilient. Honestly, this is a really smart essay and contains quite a lot of actual, practical guidance on how some of this thinking might work – this is very good indeed (via Patrick Tanguay’s ‘Sentiers’).
  • Seaweed: This is, to be clear, quite a long essay on the mechanics and logistics of seaweed farming in Canada, which, fine, I appreciate probably doesn’t stand out as a must-read in this week’s selection but which I promise you is both more interesting than you’d think and also a general reminder of how the global tech and finance industries continue to be really, really fcuking bad at considering the wider/longer-term consequences of business decisions. In this particular instance it’s the North American boom in seaweed cultivation businesses and what having significantly more seaweed in the ocean might mean for wider questions of biodiversity and ocean health. This isn’t a ‘sad’ story, but more one about how, despite what the past 20-odd years should have taught us about perhaps not moving super-fast and breakingthings all the time, we haven’t apparently learned too many lessons after all: “Overall, most biologists and industry specialists alike agree that seaweed farming can be done well and presents a far lower ecological risk than most other industrial or agricultural activities. But it does need to be well studied and well regulated, and it’s unclear whether that’s always happening.”
  • The Feral Web: I featured Feral Earth on Curios when it launched, and it’s grown and developed since then – in case you don’t recall, I described it as “a website featuring a bunch of hyperlinks which are only clickable under certain specific environmental conditions – so one will only work when one of the sensors attached to the website tells it that it’s raining, for example, whereas another will only work on the Summer and Winter Equinoxes.” Its creator and custodian Austin Wade Smith has written this essay a year or so on from the project’s inception, writing a little more about their concept of ‘feral’ online spaces and the queer web, and the idea that we can and should where possible work to create online spaces which “doesn’t only make the world easier, but bigger, more awesome, more expressive. The alternative is to be lulled into somnambulance through a false sense of security that the world is no longer wild.” I very much like the principles outlined here – and if nothing else, I genuinely believe that this setup could be at the heart of a properly awesome piece of creative campaign work, were someone to put their mind to it (wow, just totally ruined the ethos of the whole project via the introduction of crass commercial imperatives and the demon Mammon! Thanks Matt!).
  • Welfare State Algorithms: An excellent piece of investigative reporting in WIRED, which looks at the tech underpinning the Dutch city of Rotterdam’s benefits fraud investigations – every citizen claiming benefits in the city is assessed by software to determine the statistical / probabilistic risk that they are committing some sort of benefit fraud, with those judged ‘most likely’ to fall within that bracket being automatically flagged for investigation. You may be unsurprised to learn that the system doesn’t appear to be TOTALLY free of biases – being a 30 year old childless bloke will make it VERY unlikely you ever get investigated, for example, whereas if you’re the same age but a separated woman with two kids then, well, expect to have your accounts FORENSICALLY AUDITED. Fascinating, not least to see how this stuff is already here and well-embedded, and that as ever all our panic about the impact of AIs and algos is another cheery attempt to bolt the long-empty paddock while our equine pal frolics in distant fields.
  • JLaw: I’m not ordinarily super-interested in Hollywood profiles, and I have to say that I’m not particularly interested in Jennifer Lawrence, but dispute both those misgivings I found this piece on her fall from grace and recent return to the spotlight more engaging than I expected – in the main, this is a very good overview of the different manufactured ways in which society has asked women to present themselves over the past decade, and then punished them for doing largely as they were asked.
  • Kids of Influencer Parents: Oddly enough I was having a conversation with someone this weekend, just before this piece dropped, about when we were likely to start seeing the first mainstream, non-celebrity cases involving children seeking damages from their parents for non-consensual use of their likeness, etc. on social media – and lo! Here we have a whole article in Teen Vogue, natch, about the toll taken on the now-adult children who for years were the centrepiece of their parents content empires (or just their weird little IMAGINED content empires, which is even worse in a way) and how they are now trying to CLAIM BACK THEIR NARRATIVE and other such perfectly-21stC expressions. What’s interesting about this is the way many of these people – and I’m not disagreeing that perhaps their parents shouldn’t have plastered them all over the web in search of that sweet monetisation revenue – are now using this as, well, a convenient stream of content with which to launch their own personal media empires! So, well, it’s all turned out ok! We’re all just smithereened, really, aren’t we?
  • Middle-Aged Millennials: In yet ANOTHER example of GenX-erasure, NOW we’re getting all the articles about The New Face of Middle Age – where were these when I was grappling with turning 40 all those years ago, eh? Oh, ok, fine, I am only JUST too old to see myself in this piece, which interested me not because of the shattering observation that ‘people at 40 now are different to how people at 40 were a few generations ago’ but more because of the universal-seeming sense of exhaustion and bare-minimum survival that leached from each of these mini-profiles, and the wider sense that ‘middle-aged’ as a designation doesn’t really make sense any more; it’s not a plateau you now find yourself on when you hit 40 so much as a different treadmill at a steeper gradient, and some unknown antagonist appears to have attached sandbags to your ankles, and the gym mirror has been replaced with one of those funhouse ones that makes you look lumpy and strange. As a companion piece, you might also enjoy this one, all about how TikTok is seeing an uptick in rampant ageism as all the children on the app get confronted with the wrinkles and sagging of middle-aged people using cameraphones up-close.
  • Resenting Ke Huy Quan: I loved this piece, by Walter Chaw, on what it was like growing up Asian American in the 80s, and how the portrayal of people who looked like him was basically reduced to the characters played by Oscar winner Ke Huy Quan in his 80s heyday as Short Round and Data, and how, as a child of immigrants who looks different to everyone else, those caricatures become the lens through which you’re seen, and the complicated feelings around nation and identity and representation that arise from all of this.
  • Enter The Goon Cave: I genuinely hope that none of you know what ‘Gooning’ is – but, er, prepare to learn! One of those amazing (meant in the strictly-literal rather than broadly-approbatory sense) post-internet sexuality things, ‘Gooning’ is basically the act of surrounding yourself with a frankly deranged-sounding amount of bongo and then spending as long as you can w4nking (we’re talking hours, DAYS here) until you reach some sort of transcendent state of (what I imagine to be) Sting-like tantric bliss. What this can entail is the creation of ‘Goon Caves’ in which the individual in question basically creates some sort of hypersensory stimulation chamber with bongo on every possible surface and just sort of trances out to all the visuals…look, I know it’s easy to get all OH MY GOD THE END OF DAYS about stuff on the internet, and in general I’m disinclined to pass judgement on individuals’ kinks because, well, it’s none of my business, but this really does strike me as…not very healthy behaviour. If you doubt me, go and spend some time looking up ‘goon caves’ and then we can talk (for the avoidance of doubt, by the way, Web Curios strongly advises you NOT to spend some time, or indeed any time, looking up ‘Goon caves’).
  • Hanging Out In The Metaverse: I know, I know, we did all our ragging on the metaverse LAST year – now it’s all about AI! Still, spare a few minutes to read and enjoy this genuinely wonderful article in which the bemused author Paul Murray, temporarily relocated to the US from his home in Ireland, experiments with Meta’s Horizon Worlds as a means of finding some sort of social connections in virtual life. As with all of the best writing about this sort of thing, this is basically a story about lonely people and their desperate attempts to find some sort of human connection – this is generally a very, very funny piece of writing, but there are occasional moments of poignancy that remind you of why, despite the jokes, tech like this has in fact being a going concern for lots of people for quite a while now, and will, regardless of when (and if) it ever becomes mainstream, persist in being a useful solution for all those who for various reasons don’t really work well with meatspace.
  • America Doesn’t Know Tofu: A great essay all about all the different types of tofu there are and what you can do with them, told via George Stiffman’s story of his own attempts to get apprenticed to a master tofu maker. I don’t even like tofu, and this made me HUNGRY (although I found it generally interesting from the point of view of flavours and what is and isn’t prized in different cuisines – the repeated use of the term ‘sulforous’ here as an indication of approval took me a while to get used to).
  • Buried Alive: When you inevitably, as we all do, complain of all that we have lost as a result of the web, the innocence compromised and the simple pleasures foregone in favour of 1s and 0s and LIBIDINOUS IMAGERY, it’s worth reminding yourself of this piece, which recounts how in the 1960s people were SO BORED that they took to seeing who could survived for longest whilst buried alive. This is the account of the burial of one Mick Meaney, who in 1968 lasted TWO MONTHS underground in search of eventual fame and fortune – aren’t you grateful for satellite telly and broadband?
  • Arnie: One of the best celebrity profiles I have read in a long time, this one – Mark Leibovich writes in the Atliantic about Arnold Schwarzenegger at 75, and whilst, yes, there’s a reasonable amount of stuff about his life and Being Arnie, the most interesting parts are about what it must feel like to be someone who has lived a life of by any measure staggering success and achievement, and who as a result is a singularly famous and loved and well-respected individual, and to know that they are dying, that they are no longer able to do what they once did, that the world is slowly starting to forget about them and that the Big Wins are, probably, all done, and that all of the remarkable and the amazing and the BEST bits are probably done…and what now? I thought this was genuinely great.
  • Organised Fun: Perennial Web Curios favourite Clive Martin writes about the inescapability of EVENT FUN in London (and, by extension, everywhere else) – the ticketed events and the pop-ups and the bottomless brunches and the Insta-and-TikTok promoted HAPPENINGS FOR ONE NIGHT ONLY, and the fact that it all feels so…empty. “All the bread and circus of our cities had been marketed and monetised to within an inch of their life. Eating out had become a box-ticking exercise of mass-validated establishments and micro-trends, while going out had become expensive, orderly and usually involved a killer Uber journey. Every major exhibition, every primetime cinema showing, seemed to be selling out in advance. Meeting romantic partners had started to seem like a corporate headhunting exercise. We were witnessing the true dawn of organised fun.” This works as a companion piece to this article from last week, asking why everything is suddenly an ‘event’ (the small point in here about the language of coding is an interesting one, fwiw), and if you were doing any sort of planning/strategy for a booze or youth-focused brand I would see what you can squeeze from these two imho as it feels like there’s some mileage here.
  • Eager Readers In Your Area: A short piece of fiction about a future in which both the readers and the writers are machines, and all we want is a real person to see us and hear us and be touched by our words.
  • A Good Woman: I love this so so so much. Hailey Danielle writes about having an affair, and knowing she’s having an affair, and not caring, and doing Bad Things, and I can’t stress how nice it is to read someone writing baldly and plainly about ‘bad’ behaviour – not winkingly, not making it a ‘thing’ or a personality cornerstone or a ‘vibe’ or a persona, but just matter-of–factly and with what feels to me at least like honesty.
  • Age, Sex, Location: If you ever spent time on anonymous messageboards, on IRC,  on forums or realtime chat servers, if you’re a child of the late-90s and early-00s, if your early online years were spent perfecting your touch-typing because it gave you a conversational advantage in the digital marketplace of fast ideas, if you found an early freedom in those interactions that felt like home, then this piece will resonate with you more than you know. If not, you will still enjoy it – it’s a great piece of writing – but this is very much one for the over-35s here imho.
  • Ray and her Sisters: Finally this week, I was floored by how good this piece is – Sara Baume writes for Granta about Ray and her sisters, telling the story of a family from the early 20th to the early-21st century and in just a few thousand relatively-spare words giving a history both of them and of British society. I want to say it reminds me a lot of Kate Atkinson, less stylistically and more in terms of a certain feel I get from the prose, but, regardless, this is honestly superb and I would read 100 more like it by the same author.

By Heesoo Kim