BOX113: There’s No App for Everything

12.09.24

I took 10 days off a computer which was genuinely wild and came back to hundreds of emails, most of which were newsletters and marketing so I’ve been on one of my periodic unsubscribe binges and being pretty brutal about whether I’ve actually ever read a newsletter and if not, delete. I mostly read aggregating things because I just want the bottom like quick, not thousands of words of hot reckons – I save my PhD time for that kind of reading. Nonetheless here I am scribing out a few hundred words of hot reckons.

I am sort of constantly side-eyeing the app that I want to help me organise my brain but I’m not sure it exists. I never look at the Twitter or Threads feed (just post there) so most things come through the 50 or so newsletters I’m subscribed to. Between other things I scroll through these and, similar to books, have a pretty hard ‘I’m not obligated to read this’ rule e.g. if it’s not grabbing me; skip it.

If it does grab me, then the filing begins. For anything PhD-relevant, e.g. papers and journal articles, it gets downloaded and added to Zotero where it can be read, tagged and annotated properly. I try and read 1-3 things in Zotero a day to keep on top of it. Any pertinent quotes or ideas are dragged out and dumped in Scrivener for later sorting as part of the PhD process as described previously.

Then there might be stuff that’s more transportable or mutable or work-relevant; resources, signals or reckons which all get put into Raindrop via the bookmark bar applet. Resources are lists, maps, toolsets, kits of.. whatever that might be useful one day. Reckons I’ve been less good at looking after because there’s so many of them and they don’t tend to have much staying power but any long-form essays or articles. Finally, signals should be broadly familiar to people; little snippets of change or interesting things that are worth keeping hold of. These actually get used a lot for work.

Regardless, I have another applet that I wrote myself that will quote whatever I’ve selected to post to a tweet which I then copy to the Meta one which is about my only interaction with them and I do more as a sort of ‘I’m still alive’ to the Internet. The thing is, this is such a perfectly honed workflow for my reading and research that I’m not sure what I’m looking for or even if there’s a problem – how would I know if I’m missing or forgetting things? People recommend things and new ways of working and things pop up all the time that I scroll through. Like the ‘obsidian method‘ looks great but then does it just take up more time to learn a new way of doing things any maybe my way of doing things is fine? My issue isn’t organising methods its the time to do everything. And maybe friction is good?

Upcoming

A lot of things. I saw a thing about how taking on lots of bits and bobs is a sign of low stimulation which is something I probably need to do some more should searching with. But, presented for your consideration:

HCID conference 2024 – 18th September, London. With absolute heroes and legends Laura Forlano, Sara Heitlinger and Paulina Yurman amongst others. There should be a word for when you suddenly find yourself platformed with people you’ve looked up to your whole career but feel massive inferiority next to. I’ll be wittering on about design and futures and all you ahve to do is sit through that to get to the good stuff. 18th September
Wherever, Whenever Festival – 22 October, Cologne. This looks huuuuuuge. I love chatting with Rob who is the organiser of this shindig and a super lovely guy. I’m going to be looking at five speculative future of work scenarios. Then getting a train up to Eindhoven for:
Design and AI Symposium at Dutch Design Week – 22/23 October, Eindhoven. I think I’m doing something both days. Both a keynote thing and a panel chat on a different day. More folks I’m a big fan of at the event but I don’t think the full programme is released yet. I’m really looking forward to this one. Should be similar to the amazing panel we did in 2022.

Reading

How to Raise your Artificial Intelligence. I’m a huge fan of Melanie Mitchell and Alison Gopnik’s work and this light-tough chat seems to brush over most of their theories. It may have even been Mitchell’s ‘Why AI is Harder Than We Think‘ that firmed up my PhD thinking a few years ago with a sort of ‘ok, then why do people keep going on about it?’ This interview with them both seeks to redirect attention away form hysterical men talking about the end of the world or enormous profit and towards the still worrying but likely outcomes. Bonus points as well for pointing out that we already have a technology that kills millions of people each year; the car.
Poking Holes in Reality with Prototypes from Libby Miller references some amazing practices, including Strange Telemetry work but is essentially about how shonky, half-finished, sketchy prototypes challenge the inevitbailism of Silicon Valley technological visions.
The Politics of Scaling from Sebastian Pfotenhauer, Brice Laurent, Kyriaki Papageorgiou, and Jack Stilgoe. Yes this is once again going to start being more of a PhD blog and on my reading list was this paper on why and how scaling has become the normative political and economic imperative of social and political change, driven by technology.
The Simple Macroeconomics of AI by Daron Acemoglu landed a few months ago but I only just got to reading through it. Well, reading through the bits that aren’t dense economic calculations. It basically says productivity gains aren’t that great because it’s either automating things that were already automated to some degree, that the ease of integration for contextually complex tasks is overstated and AI can’t theoretically exceed human productivity on ‘hard tasks’ (like diagnosis) because it can only be trained on human historic performance and finally that the costs of fighting ‘bad AI’ (like crime and social media manipulation) dig deep into productivity gains. Finally he basically says ‘if you really want AI to have an (arguably) positive impact on society and economy then focus on making it good and reliable rather than messing around with sci-fi fantasies of human-like intelligence:’

To put it simply, it remains an open question whether we need foundation models (or the current kind of LLMs) that can engage in human-like conversations and write Shakespearean sonnets if what we want is reliable information useful for educators, healthcare professionals, electricians, plumbers and other craft workers.
AI Scaling Myths from the AI Snake Oil substack has some useful data to point to in the ongoing debate about pumping more data at LLMs to make them ‘better.’ What’s interesting about it is they quite succinctly identify two collective myths that AI people hold to be true. One is that ’emergence’ (surprising new behaviour, novelty, gestalt; e.g. do things not in the training data or infer) is a product of more and more data. I’m not sure on this one because I don’t understand the science but there’s essentially two sides as to whether you believe an LLM can surpass its training data or not. Logically I’d say not but AI people are banking on it. The second one is this idea that that benchmark capability is equal to social usefulness. This one is a pretty obvious no from me and a lot of my PhD is about trying to pull this one apart. There’s a sort of baked in assumption that with enough power and novelty, mass AI adoption and social change is inevitable which is just a fundamental misunderstanding of how the world works. This is why AI people spend so much time talking about power and only vaguely hand wave at what their AI is actually for.
Contraption Theory, Venkatash Rao’s love letter to helicopters explores the typology of ‘contraptions’ – objects of enormous complexity and highly limited design variability that defy the tendency to simplify or variegate through design because of their ‘One Weird Trick.’ Arguably, my first ever talk that I did at an event while an undergrad was about a similar thing but I think I called it ‘Designs that Don’t Make Sense’ and talked about tourbillons (we have digital watches now), the SR71 (a plane that barely flies made out of Russian titanium) and the Concorde (a giant stonking battleship at the end of battleships). I guess the point of all of those is that they’re more political or social objects than practical ones.

Listening

I listen to a lot of K and J-pop because it’s beaty, catchy, vibey and I can’t understand the lyrics so I’m not distracted. It’s that or techno or metal but at the moment I’m enjoying the good vibes.

Jeon Somi, Ice Cream (I have a real thing for like rising bridge-chorus things I think)

Alright that’s about it. Another brief one with no big hot reckons but I’m sluggish. Love you, speak later.

BOX112: WHERE WE’VE BEEN WE DIDN’T NEED ROADS

21.08.24

I suppose there’s no excuse for the delay in missives. Things got busy and intense and I had to stop for a bit. Back in March I was getting back to PhD work, a few weeks later I had to put that aside as work got too intense and I had to focus up. Then on Friday I was at an event and, following a nip out for a call, caught the end of Alex’s bit where she said something like ‘blog, just blog’ and I sort of internally sighed and said ‘ok’ to no one in particular.

I am constantly trying to find ways to squeeze the most out of the time in the day but tend to shirk ‘productivity hacks.’ I enjoy doing the work and spending time with things there’s just never enough time so I’ve found rhythm really important – regular beats of physical and mental exercise that refresh the flesh machine. I suppose everyone is different: A post from Jay was about the pressure to produce weekly and he talked about some interesting folks that publish when they feel like it without regard for the algorithm and then there of course is Matt’s advice is to publish when it feels right (which for him is multiple times a week, usually at night).

But I like and rely on rhythms, beats, deadliness and regulation to know what I’m doing moment-to-moment and I think I’d felt ‘I’ll get back to the blog when I have something to say.’ But instead, I just went around saying those things because sitting down to write is a more strenuous process. But that’s sort of the point; putting effort into forming up thoughts and properly reflecting has always been really important and has benefits over time so here we are again. All that said, don’t expect big hot reckons today. I’m just reactivating all these muscles and stretching and yawning. And absolutely do not expect renders right away. Instead, this post contains bits and pieces from the last few months.

LinkedIn is for Boats

A few years ago, amidst the furore over Cambridge Analytica, PRISM and the creeping idea that maybe this whole social media thing was really potentially not very good over all, a new outfit called Ello briefly popped up which the design/futures/art/tech community gleefully hopped on for a few weeks. The premise was simple – no ads, no tracking – but this was before that sort of thing raised suspicions that one day they would start to figure out a way to extract money from you through subscriptions or similar.

Anyway, Ello had a bunch of curious features in its beautiful simplicity such as only being able to integrate the bread and whale emojis instead of like thumbs and hearts which thus became stand-ins (🍞🍞🍞 = very good, for instance) but the whole thing was like busting down the door to a massive empty warehouse with your friends, looking around and wondering ‘…what is this for? What are we supposed to do with it?‘ A sager group of people might have considered that this heralded the end of social media, that without the cajoling and prodding to force interaction, we were kind of fine just hanging out and talking and maybe should have disbanded. But Ello had a very particular design feature; a full-width feed. Yes, instead of a narrow vertical scroll, it had a full-width scroll like an old forum and so… Ello was for boats. Or more accurately, pictures of boats because the aspect ratio of the scroll suited boats.

I was talking with friends the other night (mostly from academia) about how increasingly resigned they are to accepting that LinkedIn is now all that’s left and that’s sort of sad but I think the reason it remains is because LinkedIn is actually for something. Sure, that thing is humiliating self-promotion but at least it has a purpose while the rest of them are sort of just ‘places’ or ‘communities’ in a busy, noisy world.

The fediverse is too complex and fractured to really resuscitate the feeling of early Twitter and the Meta stable are all horrifically over-engineered advertising psyops so really all that’s left is the blue jobs website. When I left academia, well I don’t mind telling you, I was shook by the seriousness with which LinkedIn is taken having spent years understanding it as this strange little interactive CV that you might update once a year while getting in rolling arguments over on Twitter. But while all these others are scrabbling around trying to figure out what they’re for (Instagram is for ~~photos~~, ~~videos~~, ~~reels~~, ~~storie~~s, ~~rooms~~, ~~voice notes~~, ~~calls~~, spaces) LinkedIn is just for hustle. LinkedIn isn’t a network, it’s a service; sometimes literally serving you jobs, sometimes serving your ego. The vast majority of people aren’t going to LinkedIn for lifelong, deep, intellectually and spiritually enriching community engagement: You just take something you did, dress it up as some massive revelatory experience, tag a bunch of folks, lob it over the wall and move on with your life. Everyone’s doing the same so no one finds it particularly obnoxious anyway.

In the post-Muskian Twitter period, there was this frenetic squabbling over where this massive community would relocate, to various decentralised platforms that require a lot of active engagement and push to make work. Maybe that brief period of parasocial relationships with strangers on the Internet is over and that’s just fine. Maybe a super transactional, collective fiction where we treat the platform as a service rather than a community is healthier anyway.

Recents

Malina and I ran a second session on Speculative Design for Milan Politech’s service design executive masterclass. This time we took them through a whole futures and speculative design process from beginning to end, starting with working with trends, to scenarios, to worldbuilding to designing things from that world that they then performed back to the group in diegetic livestreams, court cases and adverts. It’ revealing how tight 7 or 8 hours is to get that stuff across to folks but they really gave it their all to get into it.
I did an interview for Design Decode over the last few weeks. A couple of friends have also been featured recently, Michela who puts it all together is an incredible and diligent researcher and a thoughtful interviewer. Having the questions weeks in advance and such well-crafted ones meant I probably massively over thought it but anyway, you can go there for similar thoughts to what you might get here but perhaps more concise.
A piece I wrote for Towards the Realm of Materiality on the interface from Minority Report has been published. The book invited various authors to explore the ficitonal technologies of Philip K. Dick. I wrote about how it serves as a sort of incarnation of computational imaginaries up to that point (2001) and has perversely incentivised the development of computational technology since. Twenty-three years later, we are still trying to build things from Minority Report. Julian also invited me back on the Near Future Laboratory podcast to talk about it briefly.

Reading

Our friend Paul Graham Raven has been turning his hand to content. Here’s the first bit of content on living with the mess of contradictory but equally true theories and resist the urge to eliminate one or the other. In the case of the titular theory; is doomscrolling causing a decline in mental health or is a decline in mental health causing more doomscrolling. Why not both?
I read this article on AI becoming a technology of faith. There’s nothing surprising in there but it articulates how grandiose promises are used to obfuscate current failures, in this case a promise of a smart health assistant based on technology that makes stuff up. The telling part though, which is not unusual is when the writer brings up Sam Altman on the risks posed by such intimate health data being potentially accessible to employers:

Huffington suggested that Thrive’s AI platform would be “available through every possible mode,” that “it could be through your workplace, like Microsoft Teams or Slack.” This led me to propose a hypothetical scenario in which a company collects this information and stores it inappropriately or uses it against employees. What safeguards might the company apply then? Altman’s rebuttal was philosophical. “Maybe society will decide there’s some version of AI privilege,” he said.

Which sort of summarises another aspect of this hubris; ‘It’s not my fault if my ideas cause harm, that’s society’s problem.’
Ok I didn’t capture any other reading. For your info I tend to ‘tweet’ things I read as I read them here (ew) and here (sigh) but also bookmark things here.

Listening

Aespa – Armageddon, The 1st Album

Ok, sorry it’s been so long. Sometimes feels like that’s all life is – stumbling along, tripping and apologising profusely even as no one actually cares that much. Sonder. Love you.

Box111: Corp Wars

14.03.24

DS099: Not very happy with this, 098 was a dud. Just couldn’t get the thing I had in my head out onto the screen and was singking more and more time into playing with lighting etc. to no real benefit.

One of the things I’ve been really emphasising about this new technological wave in talking to people is that we’re not in the ‘exciting’ and frenetic days of early social media or the Internet. This isn’t a time where some new technologies are emerging and smart, playful outsiders are coming in and showing us new ways we might do things. Generative AI is characterised by four or five of the world’s wealthiest companies, run by a few dozen of the world’s wealthiest men, focussing on the two wealthiest states, fighting to maintain the status quo.

Of course there are and will be, weird and interesting things that happen along the way but the incumbents are so powerful that they can just hoover up any competition. This was well analysed by Henry Farrell on the political economy of AI. He points out that, just as with the early Internet, a war over IP is emerging between the incumbent corporation that capitalise on culture and the artists and creatives who feed that culture, only this time the incumbents aren’t Disney, Warner Brothers and the record companies as with Netflix, Napster and Spotify but the big tech companies; Microsoft, Google, Amazon and so on trying to extend the living they’ve made off the back of the work of creatives. The point Farrell makes is there a future in which this just kills culture and the Internet; that the well is so poisoned by synthetic media and market disincentives that the whole enterprise of the Internet just sort of ossifies and collapses.

As we know from Gopnik, generative AI is a cultural technology, a way of organising and disseminating knowledge. It doesn’t create anything new but changes the way we order things and value them. The IP fights going on are a symptom of this shift and in fighting to maintain total supremacy and status quo over a speculative future market, the incumbents are likely smothering anything new that might emerge as a result.

In a sort of answer to the last’s posts provocation, (‘If someone tells you what something could do, ask them why it isn’t.‘) why would any of these incumbents seek to change the techno-cultural production machine that has made their bosses billionaires? AI isn’t a disruptive force to them, it’s a compliant one and the aim is simply to avoid letting any of your three or four competitors claim any space off you. Luckily for us, maybe, it’s actually going quite badly as Open AI starts to hit a ceiling, the numbers look unworkable and they keep launching things that flop or provide some novelty but little functional utility.

Short Stuff

I’ve been asked to do an interview for a thing but the thing I really like is that I’ve been given a long time to do it. I have the questions but now have two months to answer them which is really really interesting because it means I can actually think and sit with them rather than dash them off like often happens or as on this blog

I love the detail Ed Zitron goes in to to tell the story of tech. Here he is on how bad decisions are being made in the name of profitability.
A good piece on rituals, it’s a useful antidote to the sometimes lazy framing of ‘smartphones are now rituals’ that you sometimes see in popular reporting. Rituals have specific qualities and properties that are not present in most technologically-mediated content binges.
Following on from my last post, Dave Karpf’s review of Dixon’s book on blockchains: “He sees some problems with the Internet that venture capital helped build. The only solution he can imagine is more venture capital.”
Wes on tech’s delusional relationship to Star Trek. (Weirdly I added an overlong footnote about Star Trek to a recent essay. Probably won’t make the cut but it was basically ‘Star Trek is a silly thing to pin your colours to because it is politically infeasible.’)
Jay has an interesting point about ‘Law’ vs ‘Lore’ here. Made me reflect on my brief fascination with speedrunning although I’m not sure in what way.
The fascinating trap OpenAI has itself in as a result of its arrangements with Microsoft and as a nonprofit to prove that it has not got anywhere near so-called AGI.
This was a really great interview with Cameron Tonkinwise (and Okskar!). I nodded enthusiastically with most of his talking about designers in organisations. Was hoping there’d be more succinct and clear definition around Transition Design but there’s a lot of great content there.
I have little loyalty or connection to San Francisco but you know I detest the hubris and nihlism of tech culture. This is a great piece from Rebecca Solnit on how it has destroyed the city and the paradox at the heart of claims of democratisation while Silicon Valley increasingly lives and encourages isolation, alienation and separation. Putting it better than I did in the last post.
Something really sparked a circuit in this great article from Beth Singler about apocalyptism in AI. I’m paraphrasing but; apocalypses are a utopia for those that survive them.

Sorry it was very short this week. I feel like I’ve worked through a lot of stuff recently already and have been focussing on work to try and get various things finished and over the line. Ok, love you, bye.