Google’s Bard was undone on day two by an inaccurate response in the demo video where it suggested that the James Webb Space Telescope would take the first images of exoplanets. This sounds like something the JWST would do but it’s not at all true. So one tweet from an astrophysicist sank Alphabet’s value by 9%. This says a lot about how a) LLMs are like* being at the pub with friends, it can say things that sound plausible and true enough and no one really needs to check because who cares? Except we do because this is science, not a lads’ night out and b) the insane speculative volatility of this AI bubble that the hype is so razor thin it can be undermined by a tweet with 44 likes.
I had a wonder if there’s any exploration of the ‘thickness’ of hype. Jack Stilgoe suggested looking at Borup et al which is evergreen but I feel like there’s something about the resilience of hype: Like crypto was/is pretty thin in the scheme of things. High levels of hype but frenetic, unstable and quick to collapse. AI has pretty consistent if pulsating hype gradually growing over the years while something like nuclear fusion is super-thick (at least in the popular imagination) – remaining through decades of not-quite-ready and grasping the slightest indication of success. I don’t know, if there’s nothing specifically on this, maybe I should write it one day.
*This week LLMs have been like bullshit generators and lossy jpegs as well. Both good stories but ChatGPT is like ChatGPT really.
PhD update
Writing this thing is really coming in waves. Sometimes I’m tearing my teeth out, sometimes I re-read something I wrote a week ago and am like ‘damn that’s good’ then I’m like ‘but is it actually relevant or useful?’ Then It’s a downer again. At the moment I’m on a roll with it, I was quietly working away on little bits here and there, looked up at the weekend and I’d basically done most of a chapter and it was… good?! And obvously that feeling is contagious so I’ve just kept on rolling on.
Consistency is definitely key. The other weekend I took four straight days off thinking I would binge it and ended up firstly, not meeting my own over-inflated expectations of what I thought I’d be able to do and then feeling a bit grim and secondly; rushing, resulting in stuff I was unhappy with and was a bit incoherent. As a result of the binge I ended up taking a break for a few days, then when I came back I couldn’t remember all the half-finished threads I meant to get back to.
So I’ve stuck with just doing two hours or so a day (occasionally more where work allows) and it’s been really productive; small, confident steps.
Upcoming
I’m going to be delivering part of the Designing Narratives and Evoking Change course being run by the Danish Design Center, March 24th-29th. They’re a lovely bunch I’ve been talking to in an Arup capacity for a while and I’m sort of interested to see how we might build as a relationship seeing as my bag is really about the social utility of design and sometimes in a big company, where quite rightly the focus is on clients, that can be hard to demonstrate the value of or find time for.
Short Stuff
- Really great (and pretty scathing and hilarious) critique of Refik Anadol’s Unsupervised at MoMA. It’s interesting to see AI spectacularism and boosterism critiqued from the art world press. There’s a sort of incredible clarity to it when highlighting the superficiality of the work. I wonder if he’s read it.
- Matt picked up #breezepunk and ran a little further with it. He’s right, this is how things use to happen; you sort of keep the idea snowballing and people chew it over and share. I don’t think that’s how things happen now. I think you need to convince people they can get rich or something. I seeded breezepunk with Julian and co. and I’m hoping it takes root over there.
- Google Bard was released this week, weirdly I got a prediction right. There’ll be more to come as they all race to claim back the dwindling market cap for new digital products and services.*
- Dan McQuillan perfectly timing ‘We come to bury ChatGPT, not to praise it.‘
Gosh I have so many reading tabs open but there really is no time. I don’t even really have breaks between meetings to read things. I’m failing at the thing I lobby for all the time; time for reflexivity, thought and self-driven exploration.
Malina is setting up a blog for us to talk more about Arup design work. As if you could get any more but as this place continues to focus on my personal interests and practice (through PhD mostly I guess) it would be good to have somewhere to write about Arup stuff since people always ask.
Ok I do really love you. Bye.
*You know what, this is really the biggest story of the moment. It’s going to make writing about LLMs very hard because of how hot it is right now but I’m pleased that it seems to be conforming with my expectations at least.