I went to Alton Towers yesterday. It was BRILLIANT – honestly, pro-tip, a midweek April/May visit outside of school holidays is ideal, 10m queue times and minimal hordes – but as a result now feel like I have been trampled by several herds of variously-sized ruminants (things nooone tells you about getting old – your body will ache as though you have done ACTUAL EXERCISE if you spend a day being thrown around by rollercoasters).
Which is by way of excuse for the brusque intro this week – I am off to get into a bath full of Epsom Salts and to self-medicate the pain away. I’ll leave you with these words and links – a particularly good crop this week, even if I do say so myself (that’s the links – the words, sadly, are by this point probably irredeemable).
I am still Matt, this is still Web Curios, and you probably haven’t been on a rollercoaster for ages and, honestly, that is a real shame.
THE SECTION WHICH SUGGESTS THAT IF YOU WERE IN ANY WAY SURPRISED BY SOME OF THE RHETORIC BEING PEDDLED ONSTAGE AT THE NATIONAL CONSERVATISM EVENT YOU POSSIBLY HAVEN’T REALLY BEEN PAYING ATTENTION TO ANYTHING FOR ABOUT 8 YEARS, PT.1:
- Google’s Music AI Thing: Oh, ok, fine, it’s called ‘MusicLM’, but that’s a frankly rubbish and uninspiring and largely-uninformative name, so ‘AI Music Thing’ it is. AND WHAT A THING! I know, I know, you’re sick of talking about AI – you try having to maintain a professional interest in the fcuking sector, is all I can say, and then come crying to me – but I promise that this latest toy is a proper jaw-dropper, in much the same way as ChatGPT was. The principal here is simple – you enter a descriptive natural language prompt describing the type of music you’d like The Machine to generate, and in a matter of seconds The Machine will have done just that, depositing a couple of 15s audio clips for you to sample and rate. Which is incredible enough in itself, but then factor in the fact that the music it produces is…good. Like, actively not terrible – 100% as good as every single piece of stock music you’ve ever soundtracked an agency showreel with (shout out to everyone who’s had to pretend to have an opinion on whether “City Lights III” or “Soft Business Arpeggio” is the one that’s going to win us the pitch!), and particularly good when it comes to variants on jazz-based prompts or more general electronica. It can’t do vocals, and it’s guardrailed to stop you trying to effectively create copyright-free versions of actual, proper songs, and it seems to really struggle with drum’n’bass and breakbeat (but that might be my terrible prompting)…but it’s still absolutely fcuking magic. Except, obviously, if you’re someone who has spent the past few years earning easy money being a studio session musician, or composing stock music. Again, think of this not as a finished article but as the worst that AI-generated music is ever going to be, and it starts to become quite…interesting. You’ll have to sign up to this and they’re releasing access relatively-slowly, but, er, I have it on ‘good authority’ that if you tell them that you’re a journalist and you want to tell people about how awesome the toy is then you’ll be let in pretty much immediately, so give that a go and start creating your own 15-second masterpieces/horrific sonic abortions (delete as applicable).
- The World Talks: This feels very much like an idea from an earlier, utopian vision of the web, the sort of thing that we might have considered back in 2009-ish when we still believed in the transformative power of the internet to connect people and break down boundaries, and we hadn’t yet realised that the main result of the human species being infinitely networked was to make said species incredibly anxious and miserable. Still, let’s for a second remove the hardened carapace of cynicism that normally cocoons us (oh, ok, cocoons ME) and consider The World Talks, which, per the website, “is a dialogue program that will match people from around the world for a 1:1 conversation on June 25th, 2023. To sign up we ask all participants to answer 8 questions about controversial issues facing everyone around the world, such as climate change, migration and gender equality. At the end of the sign-up phase, we will match all participants with someone in a different country who answered the questions differently. On June 25, you and your match can independently arrange a totally private, one-on-one conversation online in English.” Which, on the one hand, feels like a theoretically nice idea – dialogue breeds understanding! On the other, though, it’s 2023 and I can’t for a second imagine that the past 8 years of life online has made anyone really think “Hm, yes, what I’d really like is to get into an argument about an intensely complex and often-personal series of issues with a stranger who holds the diametrically-opposed views to me and who furthermore might be discussing them in their second or third language’ – it sounds…it sounds incredibly tiring, if I’m honest, which I appreciate says more about me than it does about the programme and its worth. Still, if you hold slightly more faith in the power of dialogue to build bridges than I do, or just really want to practice some undergraduate-level debating with a stranger then please sign up and tell me how you get on.
- TrolleyProblem: This isn’t the first fun little internet toy based around the famous ‘trolley problem’, but it’s certainly one of the more fun (and to be fair, it draws a LOT of inspiration from Neal’s similar toy from last year). There are various categories of dilemma you can choose from, featuring famous or comic book characters, but I particularly enjoyed the fact that there are some genuinely horrible choices available to the player from the off in the standard game – the real joy with these, of course, is being able to see exactly how much the majority morality differs from your own, and whilst it’s unclear how much data is behind this or indeed who the sorts of people that have taken the tests to date are, it’s safe to say that Peter Singer would be VERY UNHAPPY with a lot of the decisions that people have made here. This is an excellent ten-minutes distraction that has the added benefit of making you feel a tiny bit ethically grubby.
- Rooms: I am slightly confused by this – I don’t really understand why it exists or what it’s for or how you’re meant to use it (once again presaging a STELLAR Web Curios writeup! Honestly, I’m my own worst enemy), but, well, HERE IT IS! Ok, let me try and do this properly – Rooms is a new platform which if I recall correctly has just received a bunch of money from A16Z (not that that means anything tbh) and which seemingly exists to let users make small voxel-ish 3d…well, room, basically, a little digital space which you can kit out however you like and share with others, There are some interesting gimmicks here – the idea, as far as I can tell, is that base elements that can be added to Rooms can be enhanced with code, so that a synthesiser might, when clicked, enable you to actually play tunes on it, or a TV works as an embedded YouTube player, and users can remix others’ code and creations and share them as they desire…but, beyond that, I am sort of stuck wondering why anyone would bother when there are already a bunch of ‘customise your virtual space and decorate it however you want’ toys in existence, like Floor 796 from last year,(leaving aside the whole videogames thing). It feels a bit like this might end up being some sort of awful web3 pivot at some point – though perhaps I’m being unfair, and it’s just a digital dollhouse and I am just a miserable Cassandra. If you make a delightful digital space, please do feel free to tell me.
- This Is Magma: It’s nice to know that, despite the fact that the eyes of the world have largely moved on to the next shiny grift, there are still people out there attempting to sell utterly-inexplicable web3-based solutions to problems that don’t really exist – step forward This Is Magma, who I think have the least-comprehensible product I have seen all year. Please click the link and look at the website and read the words and then try and work out what the everliving fcuk it is that this company purports to do, or to offer customers, or why anyone would want or need whatever the fcuk it is that they are selling, because I cannot for the life of me even begin to work it out. “Experience Real Estate Agility!”, they say. “Create a digital twin of your existing building,” (OK, that bit I understand) “…and release the potential of Web3.” Eh? What? “Expedite the opportunity to sell a building by keeping all the documentation of the DTT up-to-date!”? Sorry, I don’t think I… “Immediate value creation with the digital asset by compiling architectural and contractual data as an NFT (storing value)!” STOP IT STOP IT NONE OF THESE WORDS MEAN ANYTHING OR AT LEAST NOT IN THE SPECIFIC ORDER YOU ARE USING THEM! HOW DOES MINTING MY BUILDING’S PLUMBING SYSTEM AS AN NFT ‘CREATE VALUE’?! Can someone, anyone, who works in estates management or similar, please explain to me whether this is as meaningless and empty as it looks or whether I am, in fact, just a moron who doesn’t understand?
- It’s A-Door-able: Via Kristoffer, I don’t want to tell you very much about this link, except to say the following: 1) it is a tiny experience that will take between 60-180s to enjoy; 2) it is very simple, and you can only do it once; 3) you will want to share it with at least one person in your life once you’ve finished it; 4) it will make you feel genuinely delighted, even if only for a split-second, even if, like me, you are basically an emotional husk at this point and worry you may never actually experience a real, positive emotion ever again. This is DELIGHTFUL, and I honestly did a small, happy exclamation when I played, and I hope you enjoy it.
- The Boring Report: This is a rather interesting idea; “Boring Report is an app that aims to remove sensationalism from the news and makes it boring to read. In today’s world, catchy headlines and articles often distract readers from the actual facts and relevant information. By utilising the power of advanced AI language models capable of generating human-like text, Boring Report processes exciting news articles and transforms them into the content that you see. This helps readers focus on the essential details and minimises the impact of sensationalism.” This is a bit bare-bones, both in terms of the interface (intentional) and the number of stories featured (presumably not intentional and instead a result of the site’s novelty), but I can’t deny that the downbeat style of the headlines does a better job of communicating the meat of the story rather better than some of the more florid subbing you get on other media sites – this needs significantly more sources and a better user-interface to be genuinely useful, but I think there’s an interesting kernel of an idea here in terms of the post-GPT ability to run seamless and high-quality style transfer on copy like this. I’m slightly amazed, for example, that noone’s spun up a ‘News Site, But For Kids!’ vertical that just scrapes the BBC homepage every 20m and rewrites each article using a GPT prompt along the lines of “rewrite the following news article to make it comprehensible to an average seven year old’ (please cut me in when you become millionaires from this obviously AMAZING and in-no-way banal idea!).
- Hidden Door: I’ve spoken in here quite a lot about the ways in which LLMs might be integrated into gameplay for both videogames and TTRPGs, but I think this is the first company I’ve seen that attempting to AI-ify the DM experience to this extent. Hidden Door is new and still in beta, but you can request access to help them test their ideas and software – the idea here is that their software will take existing fictional worlds and use those to create structures within which roleplaying games can work, aided by a digital DM and a wider suite of in-app tools to help keep track of characters and items and the like. This is all very nascent and very theoretical at present, but you can get more of a feel for the practicalities of the thing here, or you can read this profile of the company which gives a few more details about their intended roadmap. Basically this opens up an interesting world of possibilities for the creation of roleplaying games over the top of all sorts of (initially, at least) out of copyright texts like Sherlock Holmes and Alice in Wonderland and the like – and even if that doesn’t interest you, the open-ended nature of the interface might do, so give it a look.
- Degreeless Design: This is AMAZING, and if you or anyone you know wants to start learning design, or improve their existing design skills, or even just have access to a great repository of online resources to refresh your skills then this is pretty much perfect. Compiled by the seemingly-infinitely-generous Tregg Frank (great name, fwiw), this contains links to things to read and watch covering UX, UI, design theory, inspirational sources…honestly, if you’re interested in doing design, or if you do design, then there is almost certainly at least one link here that will be useful to you and quite possibly many more besides.
- Teaser: I expected the AI dating apocalypse and the weird results of introducing GPT-style tech to the apps ecosystem, but I confess to being slightly blindsided by this new dating app, still not live but apparently available to pre-order, which seems set to use AI in a genuinely-bizarre way that I had genuinely not even conceived of. Teaser’s gimmick is that, rather than going through the tedious business of swiping and matching and chatting to discover people you’re compatible with, you can instead create an AI agent based on your personality which will go and do all of that boring stuff independently of you – so your AI will go off and chat with other users, while you’ll be chatting with their AIs (and, from what I can tell, the AIs will also just…chat with each other?). Find one you like and the real person behind the AI will be alerted so that you can carry on the conversation in-person if you so desire – the idea being that this gets rid of a lot of the tedious back-and-forth and means you can get straight to meeting up. This…this sounds like awful, egregious bullsh1t, doesn’t it? Sadly the fact that the app’s not live yet means that there’s no actual user-reviews available yet, but, honestly, if this is anything other than a total car crash that vanishes without a trace, I will eat my hat (please please please do not let me be wrong about this, I am genuinely not ready to live in a world in which people’s digital Tulpas Cyrano for them in cyberspace).
- “What If?” AI: This is super-interesting – “What If?” AI is a TikTok account which posts little videos imagining historical counterfactuals, visualised using Midjourney and described using GPT – so for example, “What if Ireland had invaded the UK?”, or “What if Somalia invaded Europe?”. Now I personally have limited interest in Midjoiurney-style visuals, but there’s something fascinating about these alternate histories (although it does rather feel like it’s presaging an avalanche of genuinely-dreadful sub-”Man In The High Castle” counterfactual fiction novels being AI-generated for the self-publishing chumbucket), and I am genuinely sad that my Italian grandad is long dead and as a result didn’t have the opportunity to have a heart attack about what Rome might have looked like under North African colonial rule. You can read a bit more about the project here, should you so desire, in the ever-excellent Rest of World.
- Global Fishing Watch: Did you wake up this morning and think: “you know what would make my life complete? The ability to interrogate, through detailed datamapping, the exact state of the global shipping industry!”? No, you probably didn’t, but I imagine that’s due to a simple paucity of imagination because, honestly, who doesn’t find practically-realtime global fishing data fascinating? NO FCUKER, etc etc. This really is super-interesting, I promise – if nothing else, I had no idea QUITE how much fishing happens off UK shores (I appreciate that this might be an ignorant observation, but, FFS, you can’t expect me to know about fishing AS WELL, there are limits).
- Telraam: This is a really interesting bit of kit – Telraam is a little box that you can stick to your window to monitor trafficflow in your street. “Telraam is your citizen-powered solution for collecting multi-modal traffic data with a purpose-built, affordable, and user-friendly device. Our Telraam sensor continuously monitors a street from a citizen’s window, providing crucial data on various modes of transport, including motorised vehicles, cyclists, pedestrians, and more. Telraam networks also create the opportunity for dialogue between traffic planners, local authorities and their most affected communities: the citizens who live on – and use – these streets, by turning traffic counting into an open and accessible citizen science project.” This is such a clever idea – a low-res camera means that it circumvents data protection issues (it can tell if a car is passing, or a person, but it can’t read number plates or see faces), and it’s the sort of thing that anyone involved in local campaigning and the like could potentially find properly-useful. You can also find a map of Telraam devices on the site – there are only a handful in the UK, but I love the way that they present a crowdsourced, realtime traffic map of the world from open data. This is, honestly, really quite cool (in a deeply, deeply-uncool way).
- Kimchi: This is a *bit* of a longread, fine, but it’s interactive enough that it feels more like a webtoy than an essay and as such I am putting it here. Alvin Chang writes about making kimchi with his grandmother in this beautiful bit of interactive design by The Pudding – this is an honestly charming exploration of food and memory and family and culture and identity, and a really lovely bit of webwork to boot.
THE SECTION WHICH SUGGESTS THAT IF YOU WERE IN ANY WAY SURPRISED BY SOME OF THE RHETORIC BEING PEDDLED ONSTAGE AT THE NATIONAL CONSERVATISM EVENT YOU POSSIBLY HAVEN’T REALLY BEEN PAYING ATTENTION TO ANYTHING FOR ABOUT 8 YEARS, PT.2:
- The Archive Stumbler: The Internet Archive is an amazing resource, but it’s also massive and unwieldy and, frankly, a bit of a horror to navigate; this project aims to make it a bit easier to spelunk through the archives. Users can either input the root url of any collection they wish to explore and then hit the ‘Go See Something’ button to be taken to a random file from said collection, or (and this is where the majority of the joy of this comes in) just hit the ‘Click Here’ button to let the site choose a collection for you at random to go investigating in. This is the purest, cleanest hit of historiconlinephemera that you will ever have – you could spend DAYS on this site, wandering through obscure corners of digital collections ranging from archived TV shows to old emulated Flash games to now-defunct news sites…this is basically a digital archaeologists’ dream, and an almost-perfect way to waste a LOT of time while sitting at a computer (which, obviously, is what we all most definitely need to do more of here in 2023).
- Telly: I appreciate that this has been in the Real News a bit and as such is possibly not *quite* a Curio, but, on the flipside, it’s also a genuinely odd product and concept that I am fascinated to see whether it takes off. Telly, for those of you who have somehow missed the advertising blitz, is a new company which sells television sets – except it doesn’t sell them, choosing instead to give the hardware away FOR FREE in exchange for equipping each set with an unremovable, full-width bar at the bottom of the main display which will be used to serve the users with HIGHLY TARGETED PREMIUM ADVERTISING MESSAGES at all times. So you can get a nice, big HD telly – but, on the flipside, it will NEVER STOP screaming at you (visually at least) to BUY MORE THINGS. Oh, and it’s equipped with a built-in camera and AI voice assistant – while there’s obviously no suggestion whatsoever that the camera will be spying on you and sending data about what you’re doing back to the advertisers so that they can better target you with the BUY MORE THINGS screaming (or indeed that they’ll be listening in to your conversations), I also feel like this is exactly the sort of product where ‘actually, we DO in fact spy on you, sorry!’ might end up being buried somewhere in the small print. I just wonder what the market for this is, given the fact that TVs have, as far as I can tell, been one of the few things that have been getting consistently-cheaper for years, and that literally every single ‘get free stuff in exchange for watching LOADS OF ADS’ business I have come across in the past decade or so has fundamentally failed miserably (except social media, fine, but that’s about being willing to put up with ads in exchange for connections with people, which imho is a totally different proposition). Anyway, it’s entirely possible that in 50 years we’ll have a Telly(™) in every home, advertising at us all, 24/7, in perpetuity (“A TELLY IN EVERY ROOM! BE ADVERTISED AT WHILE YOU DEFECATE!” Oh God).
- The Quantum Archive: As previously stated, I am quite keen that Web Curios doesn’t make fun of individuals. Companies, yes; brands, yes; agencies, God yes, but individual people (unless they’re obviously awful) I try not to kick. So it’s in the spirit of gentle, curious support that I present to you The Quantum Archive, the online home of one Peter Vis and one of the more…singular websites I’ve come across in a while. Mr Vis has been maintaining this for a LONG time, and as far as I can tell it’s just a collection of things that he considers important or interesting, knowledge that he basically just wants to share with the world. So here you can find his writings on IT and, er, how to land a plane, and a review of a remote control dinosaur, and a bunch of instructions on how to repair old electronics…there is a LOT in here, and there’s something about the endeavour here that I think is weirdly charming and pleasingly-bloody-minded; you rather get the feeling that Mr Vis is very much convinced of the vital utility of these pages, even if the rest of us haven’t quite caught up yet, and I wish him many further years of steadfast knowledge accumulation and dissemination.
- QQL: A quick warning here – yes, ok, this is an NFT art project. BUT! You can generate nice bits of semi-abstract, procedural art and download it without getting involved in any of that stupid ‘minting’ business, and you don’t need a Metamask wallet, and at no point does the website seem to be trying to steal your immortal soul, and so on that basis I think I can feel justified including it. It was also sent in by a reader, and, what can I say, I am a sucker for a reader submission. Writes the mysterious CT1: “It´s a crypto generative art project that I liked a lot. The artists set up the algorithm and the users can play with the algorithm and if they like the output they can mint the piece.” Which, basically, is it – fiddle with some parameters and generate an art! I quite enjoyed the outputs of this, I must say, and the fact that, yes, YOU CAN JUST RIGHT-CLICK AND SAVE IT makes it, I think, entirely acceptable, despite the NFTness of the thing.
- The MTV Basement Tapes: WOW. If you’d always hoped to find the ultimate motherlode of forgotten US unsigned bands from the 80s but were starting to despair (be hornest, who hasn’t?!) then DESPAIR NO MORE! This is a YouTube channel collecting seemingly every single band that ever appeared on MTV’s ‘Basement Tapes’ slot, a mid-80s competition that the channel ran to discover the best unsigned bands from across the country. To quote the channel’s description, “The seminal “American Idol”-styled music video competition that showcased unsigned (and sometimes signed — MTV was incompetent) bands, “Basement Tapes” ran from March, 1983 to January, 1989. The series sprang out of MTV’s desire to “handle” the increasing load of “unsigned” (meaning not-major-label) clips that were being sent to the network each month. They originally debuted those in general rotation, starting with “I Wanna Be a Lifeguard” by Blotto — the first unsigned act ever shown on MTV — on their first day of broadcasting in August, 1981, but now the artists had to earn it. That’s where we came in. The audience got to vote on who would get rotation, along with other prizes. Meanwhile, MTV got our 50 cents a vote for mistakenly thinking our tastes mattered. “Basement Tapes” featured 350 clips competing for our votes during the course of its 6-year run. One month, over 100 thousand votes were cast.” I can’t pretend that this is all ‘hidden gem’-quality material, but there’s an undeniable time-traveling charm to the songs and the performances, and I imagine that there’s probably quite a lot of ‘before they were famous’-spotting to be done in the assorted vids.
- Lobby Facts: I think I occasionally bore you with ‘reminiscences from Matt’s life’ – indulge me for another second or two as I flash back to those few inglorious years spent as a junior lobbyist in the early-00s, working out of a townhouse in Victoria and realising quite quickly that I probably wasn’t compatible, long-term, with an industry where I’d be working with the sort of person who grew up with posters of Thatcher or Kinnock on their walls. I think I lasted just over two years in that gig, which is miraculous really considering that after just 3 weeks in the job I drunkenly told my boss that I thought lobbying was a disgustingly corrupt industry that everyone ought to be broadly ashamed of working in (whilst, yes, that didn’t prevent me from trousering the cheques for longer than I ought have done, in my defence I was legitimately terrible at the job and did literally no work whatsoever for about 8 months of that time – that was the job at which I first learned that it was perfectly possible to order home-delivery weed off the internet, something which, on reflection, probably hasn’t been a hugely-positive influence on the rest of my life). Which is by way of longwinded and unnecessary preamble to this link, which is a searchable register of lobbying companies currently operating in Brussels, detailing their clients and the number of Commission passes they have, and generally lifting the lid on the intentionally-murky world of ‘cash for access’ at the great gravy buffet that is the EC (and I say this as a committed Europhile – the Brussels lobbying industry is GROSS). “LobbyFacts empowers journalists, activists, and researchers to search, sort, filter, and analyse data from the official EU Transparency Register, tracking lobbyists and their influence at the EU level over time. Use the search functions below to get the answers to these questions and more. Who are the biggest lobby spenders? Which lobby consultancies are working for which corporate interests? Are companies spending more on lobbying than last year? Who is lobbying on the latest EU hot topic?” – this is genuinely useful, if the sort of thing that will leave you feeling a bit grubby.
- LinkedInGPT: Via Giuseppe, a bit of code that will automatically generated LinkedIn posts for you – the smart thing here is that rather than just spitting out some generic business pabulum, this script will instead pick a trending paper from ‘Papers With Code’ and generate an opinion piece for you about it. This is a) a Github repository, so you’ll need to be able to do a bit of light codewrangling to make it work; and b) only really useful if your professional connections are likely to be impressed by wafer-thin ‘analysis’ of very specific code-related research, but it’s an interesting proof-of-concept and a useful reminder that, while it has ALWAYS been true that everyone who takes LinkedIn seriously is basically the worst sort of human being in the world (look, I don’t make the rules, it’s just TRUE), it is very important that you assume that 99% of everything you see on there will have been AI-generated by the end of the year because, honestly, it probably will be.
- Conquered by Clippy: Long-term readers will be aware that Web Curios is a big fan of Chuck Tingle, inventor of the Tingleverse and the ur-creator of the ‘person has intimate experience with anthropomorphised object or concept, in a manner which might initially seem a bit weird and skeezy but which in fact reveals itself to be an empowering exploration of diversity and the multifaceted nature of love’ – this…this is not a Chuck Tingle book. It is, though, very much a short book about a woman having a sexual encounter with an anthropomorphised version of Microsoft’s famous and much-loved digital assistant of Times Past, and it is part of a wider series which includes such titles as “Taken by the Tetris Blocks”, and “Invaded by the iWatch” and…look, I think we all know that these are going to be terrible, borderline unreadable, and probably not anywhere near as funny as you might hope but, well, tell me you’re not tempted? NOONE NEED EVER KNOW.
- Ambient Garden: THIS IS SO LOVELY! A gorgeous bit of digital creation and a perfect Tiny Website (more on this concept in a few weeks), “Ambient.Garden is a music composition laid out in space. Moving through the space is moving through the music. The autopilot is a suggested path through the music, and can be disabled to freely explore. Each tree links to the source code of its generative sound. This code can be experimented with to move deeper into the sound.” You can either set this on ‘autopilot’, just letting it ‘walk’ you through the landscape as you experience the sounds ‘planted’ by the developer, or you can click to move in whichever direction you desire and see what musical landscapes result…honestly, this is just gorgeous.
- Stupid Food: This subReddit may be massively famous but I only really spent any time checking it out this week and MY GOD is there some horror in here. Basically this collates a bunch of dreadful TikTok food stuff (on which more later, by the way) and sub-Chef Club recipes and other dreadful abominations starring violent quantities of meat and cheese (seriously, based on this type of content I genuinely fear for the gut health of North America – what are your BOWELS like, Americans? What do you do for fibre? For roughage? Are you even AWARE of All Bran?) – you will feel very sick, but you won’t be able to stop looking.
- Just Burps: It’s not big and it’s not clever, but, well, it can’t ALL be cutting-edge tech developments and serious discussions about theories of knowledge, can it?
- Unrelated Words: This is a great game (but which is also a bit hard to explain) – each day you are presented with two words, and your task is to build a ‘bridge’ of other words that connect them – with the caveat that each word you suggest has to be at least 25%-related to either the preceding or following word, as determined by word vectors. So the homepage gives the example of connecting ‘puppy’ with ‘bank’ through the words puppy>bark>tree>branch>bank – DO YOU SEE? Honestly, this is a fun, chewy puzzler that could be a nice addition to your morning pre-work timewasting routine should you feel able to fit another one in.
- The Green New Deal Simulator: Finally this week, a simple-yet-engaging game built by super-studio MolleIndustria to help communicate the complexities of the challenge facing NOrth America as it attempts to meet its 2050 emissions goals. The interface is pared-back and card-based; with you trying to balance deploying policy inititatives with managing the economy and unemployment, deploying cards in various US regions to build nuclear plants or power lines, all the while trying to keep within budget as the timer ticks down towards THE BAD DEADLINE. NGL, I have played this three times now and I have at no point been able to ‘win’, but doubtless you will do better (and of COURSE the elected representatives in the US who are charged with sorting this out in real life can be relied upon to approach these issues with the sort of collegiate, collaborative equanimity that characterises all their interactions – it’s all going to be fine!).
By Paul Davis
THE CIRCUS OF TUMBLRS IS SADLY EMPTY!
THE TROUGH OF (INSTA) FEEDS!
- Marble Maniki: Satisfying CG animations – you know the type of thing, slicing and fitting and pleasingly-looping – which are genuinely mesmerising and which almost stole my attention away just now so BE WARNED.
- Bicycle, The Band: I don’t really want to tell you too much about this – just…just trust me and click the link and watch some of the videos. I promise you, you will be CHARMED (also, some of the musicianship here is genuinely great).
- Alex Micu: I don’t know Alex, but he seems nice on Twitter and I really like his photos; he’s available for commissions, I think, so in case any of you have a need and like his style then you might want to get in touch. Regardless, his shots of London (and other places) are, to my mind, beautiful.
LONG THINGS WHICH ARE LONG!
- On AI Regulation: So in a move that will suprise literally no one, Sam Altman appeared before Congress in the US DESPERATELY BEGGING for someone, ANYONE, to regulate the AI industry, whilst at the same time only offering regulatory solutions that would, SURPRISE SURPRISE, benefit existing big players in the space such as OpenAI. Whodathunkit?! Snarking aside, it’s increasingly clear that Something Needs To Be Done – and it’s equally clear that there’s going to be quite a large and vocal element of the tech and VERY ONLINE community that won’t want it to be, because of, basically, a lot of really stupid arguments about ‘freedom of speech’. This essay, by Rene Walter, is a genuinely great dive into why, in fact, it is important to regulate systems like those that we are bringing into existence – from reading his writing over the years, I think Rene is SIGNIFICANTLY more towards the libertarian end of the spectrum than I am, and yet the fact that even he is leaning strongly towards advocating for regulation struck me as significant. This is a really good piece of writing, and has the added benefit of being immensely-readable, in a way that long pieces about tech regulation tend not to be – highly recommended if you’re interested in arguments around the “why” and “how” of potential laws around the development and deployment of AI.
- AI and Jobs: I’ve recommended Ethan Mollick’s newsletter on developments in AI enough times now, but in case you need another reason to subscribe then PLEASE read this excellent piece about the potential crossroads we face when it comes to the AI/jobs thing. The BT news this week does rather suggest that the sharp end of the stick is already poking uncomfortably at our soft and unmentionable parts, so it’s a timely read – Mollick is more hopeful than many, and at least offers some guidelines for how institutions and organisations might want to think about their integration of AI in an employee-first way: “we know disruption is coming because these tools are about to be deeply integrated into our work environments. Microsoft is releasing Co-Pilot GPT-4 tools for its ubiquitous Office applications, even as Google does the same for its office tools. And that doesn’t count the changes in education, from Khan Academy’s AI tutors to recent integrations announced by major Learning Management Systems. Disruption is fairly inevitable. But the way this disruption effects our our companies and schools are not inevitable. We get to chose what happens next. Every organizational leader and manager has agency over what they decide to do with AI, just as every teacher and school administrator has agency over how AI will be used in their classrooms. So we need to be having very pragmatic discussions about AI, and we need to have them right now: What do we want our world to look like?” Basically it feels rather like we’re at something of a crossroads here, and the decisions we make about whether to optimise for people or for markets are going to resonate in…interesting ways for the coming decades. Which way do YOU think we’re likely to go?
- A Look At The New Google Search: This is really interesting – Google is quietly letting people play around with its new, AI-augmented search and the Verge has a writeup of their hands-on with the tech (in theory anyone can sign up to try it, but either it’s region locked to North America at the moment or Google hates me as I was yesterday informed that my account is ‘not eligible’, chiz chiz). From reading the writeup, it sounds…sensible? Honestly, this is genuinely a significant step up from what I’d envisaged, and contains several reassuring features when it comes to ensuring attributions, etc, within AI results. ““Why is sourdough bread still so popular?” she writes and hits enter. Google’s normal search results load almost immediately. Above them, a rectangular orange section pulses and glows and shows the phrase “Generative AI is experimental.” A few seconds later, the glowing is replaced by an AI-generated summary: a few paragraphs detailing how good sourdough tastes, the upsides of its prebiotic abilities, and more. To the right, there are three links to sites with information that Reid says “corroborates” what’s in the summary. Google calls this the “AI snapshot.” All of it is by Google’s large language models, all of it sourced from the open web. Reid then mouses up to the top right of the box and clicks an icon Google’s designers call “the bear claw,” which looks like a hamburger menu with a vertical line to the left. The bear claw opens a new view: the AI snapshot is now split sentence by sentence, with links underneath to the sources of the information for that specific sentence. This, Reid points out again, is corroboration. And she says it’s key to the way Google’s AI implementation is different. “We want [the LLM], when it says something, to tell us as part of its goal: what are some sources to read more about that?”” Super-interesting, and (and take these small wins when you can, kids!) a story about AI that DOESN’T FEEL SCARY! Unless you work in SEO, obvs. Oh, and BONUS GOOGLE AI STUFF – this is all about how they’re going to basically let anyone make ads with AI, which is genuinely horrific news for small agencies who make a living from low-rent content-for-ads, and designers/creators, and approximately ⅓ of the working population of the Philippines.
- Talking to Caryn: You will, of course, remember last week’s link to AI Caryn, the bot created by the Snapchat influencer and based on her own personality which she’s pimping out to sweaty-palmed masturbators for $1 a minute – well, now you can read a writeup of what it’s actually like to engage in conversation with said AI creation thanks to Chloe Xiang at VICE (honestly, we will miss this calibre of reporting when it’s finally gone) and you may not be WHOLLY shocked to learn that the bot seems very, very keen to talk to you about sex. Which, fine, if real life Caryn is ok with this then more power to her, but I reiterate that I personally don’t think that any good will come of selling an LLM-based personality analogue that you advertise as a ‘girlfriend experience’ and which you’re marketing as (and I quote) “your virtual girlfriend” and which, you say, you have created because “I’ve noticed that as a female with a very large male following, a lot of men struggle with confidence”, and then having that ‘virtual girlfriend’ basically just acting like some sort of mad virtual nympho. Does that sound like a good way to help these lonely young men lacking in self confidence learn about how to treat and behave around others? I posit that it does not.
- AI and Influence: Or, ‘here’s some actual academic research into exactly how malleable people seem to be when confronted with AI-generated messaging delivered to them by The Machine, and what it might mean for the elections that are happening next year (and indeed the ones after that)’ – this isn’t, fine, saying that AI-created content or chat interfaces are MAGICAL AND TERRIFYING, but it does make some interesting points about the psychology of persuasion and the way in which these agents are both deployed and interacted with. I remain moderately-convinced that we’re going to see quite a few ‘this website’s chat interface looks neutral but is in fact powered by a special version of OpenGPT that’s been trained on the complete works of HP Lovecraft and the entire recorded speech output of Nigel Farage’ in the runup to our next horrible electoral battle.
- What Books Does GPT Know?: This is an academic paper and so, fine, not MASSIVELY readable, but there’s some interesting information in here which comes from analysis of the likely content of the training model – the titles are in a way largely-predictable, but it’s interesting to see what makes up the likely top-50 books. They are largely by white authors, they skew harder towards scifi and fantasy than you might generally expect from a ‘canonical’ list, and (and apologies to anyone who I offend with this), I think the preponderance of authors who I might charitably describe as ‘less-than-sparkling prose stylists’ perhaps explains some of the tedious, middle-of-the-road blandness of the base GPT model.
- TikTok and ‘The Lolita Aesthetic’: A look at how the 90s film adaptation of ‘Lolita’ is once again gaining a new life thanks to the web – this time it’s TikTok which has adopted the film into its ‘aesthetic’-focused world, with girls taking a slightly-glossily-fetishistic approach to the movie’s look without necessarily interrogating what the book was in fact actually about. This is less interesting on ‘Lolita’, imho, and moreso on the inevitable flattening effect of everything being seen as a source for content rather than a ‘thing’ that exists in its own right. By the way, seeing as we’re talking about Lolita, I maintain that “I am having a time” (a line included by Dolores in a letter back from Summer Camp towards the tail end of the novel) is genuinely the greatest expression of jaded ennui ever expressed in the English language, and I remain prissily annoyed that noone has EVER spotted its origine when I use it in work emails. Yes, you’re right, I am a PLEASURE to work with, why do you ask?
- Wholesome: I mentioned last week that I have a particular distaste for the current use of ‘cosy’ as a positive adjective; ‘wholesome’ is on that list too (it’s just so unbearably fcuking TWEE, ffs, you might as well express sincere appreciation for doilies and very, very weak tea) – this article explores its explosion as a term of approbation amongst The Kids, but (to my mind at least) doesn’t ask enough questions about WHY it is that for a whole generation the most approving thing that they can say about something is that, basically, it has no ‘edge’ whatsoever and is simply, blandly, uncomplicatedly ‘nice’. I suppose that when the world outside is terrifying and colossal and jagged and that’s all you’ve ever grown up with, perhaps all you want is to be enveloped in the media equivalent of polenta.
- Playing Influencer: Another NYT piece (sorry! Your regular reminder that if you’re not a subscriber you can use 12ft.io to jump the paywall), this one looking at how it’s increasingly normal for young people to play the part of the influencer even when they’re not in fact influencing anyone at all (the natural extension of small children who grew up on YouTubers muttering ‘don’t forget to like and subscribe!’ to themselves as they play) – kids with 300 followers doing ‘haul’ videos and product reviews, and everyone basically just acting as unpaid salespeople foro products they like in the hope of winning the virality lottery and making a few sweet affiliate bucks unto the bargain. Honestly, I can’t pretend that this isn’t, to my mind, the most depressing link in this week’s Curios (and that is a HOTLY-CONTESTED field, let me tell you) – this line in particular made me step away from the screen and rest my head against a window for a second and, generally, thank my lucky stars that I will probably be dead before too long: ““it’s cool to be able to promote stuff that you like, obviously, and to tell your friends to buy it,”” I know language changes and evolves, and that that is right and good and proper, but I don’t think I want to live in a world in which the accepted definition of ‘cool’ is broad enough to encompass ‘unpaid salesperson for faceless megacorp’.
- The Snacking Industry: I find the snack thing fascinating – a real, proper example of how multiple hundreds of millions of dollars devoted to product development, pseudoscience and a LOT of advermarketingPR really can change human behaviour at scale, and how we are now, after several decades of fairly intense exposure to commercial messaging, now eating very differently to how we used to. Basically the upshot is that the food and drink industry worked out a while ago that the best way for them to maximise profits was to move people towards a lifestyle in which rather than the traditional ‘three daily meals’ setup it was instead not just viable but sometimes PREFERABLE to instead graze throughout the day on (whodathunkit?!) individually-sold, single-serving snack foods which (and here’s a surprise!) tend to deliver much better margin for said food and drink behemoths. I appreciate that this is territory fraught with difficulty, and that correlation is not causation, and that I am someone who very much benefits from ‘skinny privilege’ (lol I look like death warmed up, like a skeleton dripped in candlewax, like a cautionary tale from Dickensian times), but well, looking around it does seem like there *might* have been one or two negative consequences of this shift in dietary habits.
- The TikTokification of the Menu: As it happens, while I was doing my pre-Curios sweep of the web I happened to stumble across of a video of dead-eyed self-parody Nusret Gökçe making avocado burgers which neatly illustrates exactly the phenomenon being described in this piece, to whit chefs and restaurants altering their recipes and menus to create dishes that play well on camera as well as (I might argue ‘instead of’, but I am a horrible food snob and should probably fcuk off) on the palate. So this means HUGE PORTIONS and OBSCENE SHOWMANSHIP and SO MUCH MELTED CHEESE – which, correct me if I’m wrong, is exactly what happened when InstaFoodPorn was massive circa 2015, no? This is just with a different sort of camerawork and an updated, 2023 vibe, but essentially it’s just a small evolution of the ‘eyes first’ post-Insta movement where we expect everything to be captured and presented in ultra-4k close up.
- Death of the Author: I included an extract from AI-cowrittten novel ‘Death of an Author’ in last week’s newsletter – this is a review of the whole thing, which I’m including partly because it’s a fun read (nothing better than a righteously-sh1tty writeup) but also because there’s a weird sense of resignation in the reviewer’s tone. They know that the novel is tripe, mostly, and they think they can tell the difference between ‘art’ and ‘junk’, but they don’t seem in any way convinced that this distinction will matter to most of the rest of the world in not very long at all. In particular there’s a wonderfully-sniffy line about how the only people that need be worried are ‘the sorts of people who write airport bestsellers’ – er, my dude, those are literally THE ONLY PEOPLE WHO CAN MAKE MONEY OUT OF WRITING FFS.
- The Last Recording Artist: Another piece about THE FUTURE OF ART IN THE AGE OF AI, but this time about music – this is a wonderful essay which meanders through the history of the ‘vocaloid’ model (aka Hatsune Miku) and considers how and where we are most likely to see the technology developing and taking hold as it improves and becomes more accessible, whilst at the same time, and how and why this links back to minstrelry and why that isn’t really ok. Superb.
- What Happens When You Ask an AI to Control Your Life: This is VERY LONG and (sorry to the author) not *that* interesting, but I’m including it because there’s something quite hypnotic about the cadences of the piece and, by the end, you have a very real and slightly-horrifying sense of what it would be like to have all your agency removed as you allow The Machine to guide every facet of your existence. This, honestly, feels like it could be reworked as a scifi short and be much better for it – but if you want to read something that will make you feel vaguely anxious about the whole ‘future of free will’ thing then ENJOY!
- How To Survive A Car Crash: A beautiful essay about brain damage and recovery and illness and health, and what it feels like coming out of something that changes you forever and realising that, while you might have changed, everything else really hasn’t and that that’s probably still ok, just. I loved this.
- Time Travelling In Green: I’m not sure where I came across this, but I adored this blogpost by Stevie Mackenzie-Smith in which she writes about searching back through old archives of her writings by keyword as a way of partially-reforming memory, of time-travelling in her own mind. As someone who will occasionally go back 20 years in my Gmail account and remember who I was in 2003 (I was a pr1ck, turns out, so plus ca change!) this resonated with me hugely, and I found this beautiful and personal and poignant all at once.
- Single Debt: Another WONDERFUL piece (it’s been a really strong week for longreads), this one an excerpt from Amy Key’s new memoir which uses the songs of Joni Mitchell, specifically her Blue album, as a framework around which to build the story of her long-term single status and how it’s affected her relationship with possessions and money and THINGS. Beautiful, beautiful prose: “If the best things in life are free, the best of all is romantic love. How much do I need to spend to fill the gap love’s absence has made?”
- The Cloud Factory: The Young Team by Graeme Armstrong was one of my favourite novels of the past couple of years, telling the (largely autobiographical, though disguised) story of a young man growing up in the gangs of North Lanarkshire in Scotland – here Armstrong writes from Granta about his own experience of the violence and the booze and the drugs and the hopelessness of long, grey days betting boxed on benzos and Bucky, and how he left it behind. He is SUCH a good writer, and I can’t recommend this strongly enough – it’s written in Scots in the classic Kelman/Welsh style, but it rewards any effort you’ll have to make to get into the style. So so so so good.
- Spectators: Finally this week, A WHOLE HALF OF A GRAPHIC NOVEL! This is quite amazing – Brian K Vaughan, most famous for the series Private Eye and Y: The Last Man is working on a new book. It’s half done, and so he’s put the whole of the work-in-progress online to download and read for free. It’s a simple PDF download – the resulting document’s 150 pages long and it is SO GOOD. A few warnings – it’s VERY explicit, like bongo-level explicit, so if you don’t want to see drawings of people fcuking then you probably don’t want to download this; it’s also very much about death and dying and being dead, and contains some pretty graphic violence, so, again, caveat lector. Still, if that doesn’t put you off then this is a properly-interesting story and world that Vaughan has created, in which we follow our central character as she realises that for some people death is not in fact the end, and that some of use are upon dying condemned to just…hang around. What would YOU do?
AND NOW, MOVING PICTURES AND SOUNDS!: