On this week’s show we are joined by Matt Webb to talk about his approach to invention with emerging technologies.
You can subscribe to the AI Clock mailing list for updates and to find out about the upcoming Kickstarter here: https://aiclock.substack.com
You can find out more about Acts Not Facts, Matt’s new product invention studio here: https://www.actsnotfacts.com
And you can find Interconnected, his blog since Feb 2000 and place for thinking in public about technology and design, here: https://interconnected.org/home/
If you’d like to check out Christophe Weston, he can uncannily be found here.
Photograph of one of Tim Hunkins marvellous contraptions at the Under the Pier Show in Southwold.
This week’s show transcript (AI generated so treat with a little caution)
Matt: [00:00:00] Hello and welcome to episode 273 of WB40, the weekly podcast with me Matt Ballantine, Chris Weston and Matt Webb.
Welcome back to the show. It is time again for more excitement, intrigue and telling you that we are not going to do things that we never intended to do in the first place. Christopher, how has your week been?
Chris: Well, it’s been a good week, Matt. I have it’s been quite a nice week in many ways except for being sort of slightly ronered at the end of the weekend, [00:01:00] Friday, Saturday, feeling pretty rough.
Didn’t test, but daughter tested and she got, she was positive. So, I’m hoping that , if I was, , positive for coronavirus or COVID, , on Thursday that I didn’t give it to the many people that I met at the, , CIO100 event. But there was a lot of talk of it. It’s been around, right? So, . I don’t feel like I’m a super spreader.
Not this year. Anyway, the, , that was great. I mean, that event was really good. I have to say it’s been, you know, you’ve done this, , years ago, Matt, and I’ve done it more recently. , I really enjoyed. Seeing everybody at the event, meeting new people, lots going on, genuinely had a good time. And other than that, yeah, it’s just been a busy week with, with, with work stuff.
And, , and then me playing with, , some AI tooling that, Made me speak French, which was a bit, a bit weird and, and, you know, really impressive in many, many ways, [00:02:00] but opens up a lot of questions about, about how much we trust the tech, really. So, uh, so yeah, it’s been a good one. Did you see that?
Matt: I did. I did.
And I found it very disturbing. , so this is some, I don’t know what we call these things anymore. I mean, it’s, it’s presumably using some sort of machine learning kind of thing going on underneath it. But it’s basically some video
manipulation tool that makes you look like you are speaking French in A passable French accent but still very much a Chris Weston from Tamworth accent. Your lips are moving in sync with the thing. You’ve got no idea what it is you’re saying because you don’t speak French and therefore you could have been declaring war on Macron for all you know.
And I think this slippery slope started with allowing blurring of backgrounds and I said it at the time. that when you start to be able to allow people to be able to alter the reality in in the way in which we’re [00:03:00] interacting really dark and bad things might happen and you speaking you know more than passable french without a first clue about how you’re doing it is potentially bad
things
Chris: oh killer it’s not that’s that’s harsh anyway but i’ll also say you know you’re sounding a bit you’re sounding a bit puritan it’s a bit puritanical there’s like women women wear makeup To alter their appearance.
We should ban that. Ban it completely. We don’t know what they really look
Matt: like underneath. Okay, now this is true, and I’ve been, I’ve just finished Alice Sherwood’s wonderful book, Authenticity. I’m seeing if I can get to, , lure Alice onto the podcast at some point to talk about it, because it’s a fabulous book about what is real and what isn’t.
And, you know, one of the things she argues is that the idea of people impersonating others, or things impersonating others, is something that goes through… Evolution way before humans. There’s people, you know, mimicking and pretending.
Chris: I mean, like,
yeah, [00:04:00] I was just dying. I mean, he was
Matt: doing it way before he was doing it for many years as well.
But I think there’s a, I don’t know, there’s just at what points do we start to get weirded out by this stuff to the point where you can’t actually trust anything, madcap conspiracy theorists that are already empowered enough with enough. craziness without then having Chris Weston speaks French to add to the, you know, it will be on there with Pizzagate and 5G coronavirus Bill Gates and Chris Weston speaking French.
Now, this is the first sign of the beginning of the apocalypse or, or something. I don’t know, maybe I’m overreacting a bit. What was the technology that we’re using to be able to achieve this sleight of hand and.
Chris: It was a thing called, it was a tool called Haygen. It’s out there, it’s out there in the world.
I can, I’ll send you the link or something. I think it might have been a labs thing. But it was, you can pay to have it done. I didn’t pay. And I put it into the queue and then a week later it popped out with a video. It’s not [00:05:00] quick, but it’s probably quicker if you pay for it.
Matt: Yeah, I’d imagine so. , yeah, I found it all very disturbing and we’ll put a link in the show notes to that video as well, because it is worth seeing.
, but it is, I think,
Chris: I think it’s worth seeing if you know what I speak like already and you go, Oh my, that does sound like Chris speaking
French.
Matt: Yeah, quite bizarre. , so, so yes, and it sounds like the CIO 100 has seriously upped its game since I was either a judge or a competitor. , I see that they all got very glitzy.
Lumps of presumably glass or perspex presented to them by television’s famous Gabby Roslin.
Chris: Yes, indeed.
Where I got a piece of cardboard through the post presented to me by nobody in particular, but awarded by you. Which, , you know, they’ve upped their game. Yeah, yeah, to be fair. I think, I think these are, I don’t think it’s marble or glass, I think it’s, , I think it’s samples of the asteroid Bennu that they made of it, so it’s that
glitzy.
Matt: Wow, there’s a thing.[00:06:00] Joining us this week, , somebody who I’ve been trying to get onto the show, well I haven’t been trying to get onto the show, I’ve been wanting to get onto the show for ages, but I actually managed to get round to sorting it out only recently. Uh Matt Webb, how has your last seven days been?
Matt W: Well, it’s been pretty busy. I’ve been, I’ve been flying to Hamburg to go and speak at a conference. , but I think just to connect to what you were just saying, we’ve also in the last week set up a secret family passphrase to use in the event of, uh, imposter scam. Deep fake video calls.
Matt: Ah, so they, they say, well, what is the, uh, the third styling on the right?
And they’ll, they’ll respond with, it is flying tonight. And then that will mean that you can show that it’s a, it’s a real thing.
Matt W: Something like that. Well, I’m not gonna give you any hint. I know, obviously, which would kind of destroy the, the whole idea, destroy the purpose, but it’s, it feels true. There’s a, there, there’s a rising imposter scans.
, you can, you know, what was the one in Washington DC not long ago? You can train. Uh, deep fake audio of [00:07:00] 30 seconds of audio for less than a hundred bucks, and it is relatively common now just to kind of, you know, go and, uh, you get a phone call from, you know, what sounds like it’s one of your relatives, and they say, you know, I’ve been in a car accident, , they’ve, they’ve taken my phone, this is my one call, I need 5, 000 for bail, um, you’re gonna get a call in a minute from my lawyer who’s gonna help set this up, and it is common enough that the police know about it, and easy to do, and, you know, I know these kind of things have been happening for a while.
But the, you know, the ease of it means it’s worth setting up, you know, passphrases or maybe, you know, one day we’ll have two factor authentication for humans or something.
Matt: Yeah, yeah. That’s, that’s crazy, isn’t it? And I think there is, I mean, there’s this thing about, oh, we’ll put, some sort of watermark into everything that’s fake to say it’s fake.
And so I can see one big glaring… Flooring that plan. It’s the only people who do it aren’t the ones you want to not, you know, it’s crazy. [00:08:00] But, , yeah, I think, well, there we go. Passphrases for the family. That’s the next project. It’s not disturbing at all, is it? No, no, no. I mean, usually it’s fairly miserable at the start of the show, but this has gone, you know, to another level.
You know,
Matt W: people,
Chris: people beating their, , their… Beloved children about the head and neck with a spade because they’ve forgotten the password. It’s, it’s just pretty, I’m just, I’m pretty sad. It’s happened by the whole situation.
Matt W: But also positively, went to Hamburg, spoke at a conference called NEXT, , where Accenture had a main, , Sponsors, but of a conference of ideas, , talking about playfully prototyping to find out how we might be interacting with AI.
And that was a wonderfully positive experience. And I have totally forgotten how to travel, it turns out. Left for the airport late, forgot to take my noise cancelling headphones.
Matt: You know. Didn’t have all your things neatly arranged so that when you got to the security, you could just pop them out one after one, put [00:09:00] them all back in again.
Matt W: Forgotten how to
Matt: do it. Took something over 100 millilitres. Yeah, I know, there’s so many. , yeah, I’m travelling at the weekend actually, I’m going to see some friends in Prague. And, I think I’ve sort of remembered how to do it all. But this isn’t a work one, this is a pleasure one. So it’s, even that is confusing.
I don’t know, they seem to change the systems to… Anyway, that’s a good trail for what we are going to be talking about on this week’s show, your experiences in Hamburg there, so I think we should probably crack on.
Unless you take into [00:10:00] account my beloved co host’s slight affliction with limericks, this is the second time on WB40 we have talked about limericks. In, in part, poetry. The last time, you may remember, was with the wonderful poet, Mr. G. And he’d been doing some work with the Open Data Institute, and we talked about poetry and data.
And it was a wonderful conversation. This time around, we’re talking about poetry and generative AI and a, a playful experiment stroke product development that you’ve been working on for a while now, Matt. , what, what you’re doing.
Matt W: I have on my shelf a small screen about four inches across, and every minute it tells me the time in the form of a rhyming couplet. So to give away what time we’re recording this right now, I’ll tell you what it says at this second. On bookshelf I stand, time shall not be missed, with [00:11:00] rhyme and rhythm, 7 to 54 p.
m. persists. No, I didn’t say they were good poems. Just that there’s one every minute. Chris can
Matt: relate to the not good poems, but that’s another story.
Matt W: So the poem is… The poem’s generated by, well, originally by GPT 3 and then by Turbo, ChapGPT, , which… It’s a different kind of model, and maybe we can talk a little bit later about how the different models feel, because I’m tuning in a different way.
And it’s a, it’s a physical device, it’s a clock, and I put it out as a very quick hack. On X or Twitter or whatever we call it nowadays, , at the beginning of the summer, , where it went viral, uh, was in the New York Times, in the Verge, it went, it went viral in newspapers in India, ended up in Private Eye, which matters quite a lot for me.
Oh, wonderful. And so over the summer, I’ve been working on developing it as a hardware product. I’m going to be going to Kickstarter in a few weeks. And working with the [00:12:00] industrial designers an incredible studio called approach on what the you know, how to manufacture the thing how it’s gonna how it’s gonna look Yeah, it’s , it’s been an entertaining and entertaining.
A few months with this thing
Matt: and the the thing I really love about it is that an awful lot of what we’re seeing at the moment with , where exploration with particularly generative AI is happening is that it is, , being used as an approach to try to be able to drive efficiency and cost saving and the replacement of human endeavor with half arsed Software stuff and so devaluing humans along the way and what you’re doing is to be able to do something that nobody would ever Commission a poet to be able to write half arsed rhyming couplets every minute forever Because that would be insane and no poet want to do that well, I mean, maybe one would but but what you’re doing is it feels like it’s additive rather than Subtractive you’re [00:13:00] trying to be able to find ways in which , the technology and some of the hardware can be brought together to be able to do something fun and playful and, um, exploratory.
Matt W: I do think that’s why people like it. So when people, , have been in contact with me about what, what they like about it, and it’s often people who aren’t, you know, technical or deep into AI, uh, one of the reasons they like it is it, it gives them a way in to talk about what generative AI can be used for.
, which isn’t in the form of AI is going to take our jobs or something very, very obscure. It can generate small amounts, a live copy that sounds a bit like being human. It’s like, right, now we understand that, right? It’s not going to take our jobs. It’s a bit like having an intern. , the poetry is kind of strange.
It’s weirdly motivational sometimes. , it fibs. And all of these things, I think, help unpack something which is, you know, AI, a bit obscure. [00:14:00] A bit mysterious. Nobody really knows how it works. , and it gives a way in to have that conversation.
Matt: When you and I first met, now this isn’t the first foray that you’ve made into creating things. This is, this is what you do, isn’t it? I mean, the idea of being able to help people come up with ideas and to be able to shape ideas. , we first met in my… Difficult Microsoft days, as I now refer to them. And, , you, at that time, had something else that had gone pretty viral, from what I remember, which was a thing called Little Printer.
And Little Printer was a… little thermal printer but had been created in such a way that it was a bit anthropomorphic. It had a sort of outline of a face on it and, and feet and could be used as a target for all sorts of things. Essentially what it was was a very early Internet of Things device, it’s hard to actually realize that it’s only about a decade ago that very early Internet of Things things were happening, [00:15:00] how much that that world has developed.
Um,
Matt W: I mean, so that, you know, that was back then I co founded a Berg. And we’ve been working on, you know, connected hardware, sort of opening up the space. And I think, just the thing I want to say is that so many people at Berg were, you know, poured their talents and, , creative energy , and technical energy into Little Printer.
, so anything I can say about it is really just kind of, you know, what I took away from that. You know, experience of, you know, looking into connected hardware and then going through manufacturing, kind of, you know, what, what resulted from that. And it’s an incredible thing to realize that, you know, the physical world can kind of talk back to us.
And how do we relate to that? And what does that kind of mean? , and I think this is part of, , you know, I’ve always been involved in what I’d call like product invention. You know, how do we open up, you know, new, new ideas or new [00:16:00] places for imagination? , and I think what Little Printer embodied and now…
You know, the AI clock, the perm clock, whatever I end up calling it, is that One way of finding out those things is to roll your sleeves up, get your hands dirty, and just make things. And it’s not the only way of coming up with new ideas and starting these conversations. , but it’s, it’s my favourite way, and it’s the one I know how to do.
So for example, with Little Printer, the, the thing that as a studio it led to for us, weirdly, is working with Nespresso. So this is way back when. , Nespresso machines, I don’t know whether you know now, , they have a button on them. That lets you order more capsules from the device itself. So it counts how many capsules it’s used, and you can purchase it through the device.
, and that was a, that was by Berg. That was prototyped with Berg. It used the same platform that Little Printer did, which is a very kind of esoteric, toy like product, but here it is kind of, you know, being used for commerce. [00:17:00] And now this is in their machines, and that’s kind of how you purchase capsules alongside, you know, phoning them up and using the store and using the app and device commerce was a kind of a new thing back then it sort of preceded amazon dash so this kind of playful weird thinking opens up other ways of operating and you know i think that’s a you know it’s an You know, it’s a fun strategy, I think, as well as a useful one.
Matt: How easy do you think organisations, established organisations find it to be able to do this kind of approach? Because I think the other thing about with both of those examples, and I’m sure you’ve done more as well, but they weren’t a replacement. They weren’t a digitisation of the now. In both cases, what they did was that they found some elements of an emergent technology, Internet of Things, generative AI, and said, Let’s Create something that is [00:18:00] entirely new, not because it necessarily serves a, , a problem solving purpose.
But what it does is it helps us to be able to explore the idea. And it’s, but it seems that most commercial organizations are so obsessed with it having to be a solution to a problem. But that, is that a hard approach for them?
Matt W: I wonder whether this is to do with, where we are in the technology S curve as well.
The last 10 years, let’s say, we’ve kind of known, you know, in organizations what technology can do. If somebody goes like, you know, we’re going to make an app or a website, you can loosely say, Budget this, it’s going to achieve this. These are the metrics we’re going to use for it. And this is how people are, you know, going to get promoted.
This is what success looks like. And we sort of know that just intuitively, you know, engineers, designers, PMs, leadership, all the rest. Um, there are periods where we don’t know that. 10 years ago, we didn’t know that when, , when mobile was really taking off. Um, 15 years ago, we [00:19:00] didn’t know that. You know, at the beginning of digital, and I’d argue that we don’t know that again at, , this sort of dawn of generative AI, and the question is, what do we, what do we do with that?
How do we, how do we kind of find our way through, uh, these kind of things? Because like the strategy that’s worked for the last 10 years when we’ve known technology is the thing we have to discover is user needs, and we figure that out by iterating and, you know, with With kind of, you know, making decks and two by twos and post it notes, which are all very good when we don’t know what the technology is capable of, I don’t think that, and we have to find, you know, other approaches instead, just to ramble on a little bit more about that one approach that I’ve really kind of found very attractive is out of meta research and they call it pathfinding and it’s a kind of a combination of, , prototyping and strategic recommendations. Which go hand in hand. So you’re not you don’t you don’t just make a prototype.
You don’t just experiment. , what you have to do is you have to use, [00:20:00] design methods to make a recommendation to the organization and, you know, to build some conviction into we need to do more research here or we should build a product here or there’s an opportunity over here. And that’s why I try and frame the kind of product invention I do as fitting into pathfinding.
and that’s, you know, I think that’s that that’s where we are. You know, we’re imagination bottlenecked at the moment. and so we need, you know, we need kind of approaches like that.
Chris: There’s something to be said though, isn’t there, for having, if you’re inventing something, I think inventing is the right word in this context, that, that’s physical, that, you know, that you can, you can pick it up and put down as opposed to a piece of software that goes, that runs on your smartphone, which is.
You know, that’s the easiest way into create anything and it has been for some time is to write some software, but actually making something, which is a thing that does something that I think that has a different effect on people.
Matt W: It does. I think what it does is it, you know, you, [00:21:00] you just need to look at when you get a A few different people, different disciplines in the room together, talking about a deck or something on a whiteboard, the amount of misunderstandings, you know, people hear what they expect to hear.
And when you put something tangible in front of them, which can work with software, but is very, very effective with physical things, suddenly you give something, give people something, which is beyond. what they expect. Beyond language really, it is, it is just a thing, right? And it insists. And suddenly people start being able to have a conversation which is beyond, uh, you know, they stop talking past each other.
so I’m a big believer in artifacts when you want, you know, ideas to cross between engineering, design, marketing, leadership, product. And the rest, I mean, the, the problem I think with making artifacts, what, what makes it kind of difficult is how do you arrive at things which feel sufficiently new [00:22:00] and communicate.
And that’s a whole different, you know, set of difficulties.
Chris: There is something about, as you say, an object. And sometimes that is disappointing because it’s always going to be something somebody’s seen before or a derivative of or a bit like because that’s how objects are. We just can’t just magic new things up.
And when we’re talking about things and sketching things out on a whiteboard or coming up with ideas, there’s lots of kind of possibilities hanging in the air. And they’re all in different people’s heads and not… Maybe not everybody sees each other’s possibilities, but when you bring an object in, a lot of those possibilities collapse into what could be done that’s like this, which is good in a way because it, as you say, it cuts through all that stuff, but also it kind of removes a bunch of things that could have happened if that object hadn’t arrived.[00:23:00]
Matt W: I think it strips a lot of artifice away. It stops people kidding themselves that something might be possible that they’re imagining, which, you know, actually when you try and do it. The other thing it does is it opens up new possibilities. So one of the, you know, the parts of my practice is what I call software sketches, where I try and make…
You know, if I can imagine an interaction, like, let’s say one of the things I’ve been working with recently is how could we interact with AI as if the AI is like a fake user, an NPC, right? So on a, on a whiteboard canvas, the AI would have its own cursor, you would see it move around, you know, it talks to you in chat.
And, you know, I can imagine that, right? But when I start building it. What happens is it makes me feel very different, you know, certain things are very difficult to build. It’s like, Oh, okay, I need to do something that, , certain things surprise me and I’m like, Oh, okay, I could build that thing out a little bit more.
And so that’s part of the process as well. I think there’s a, there’s a kind of a, a sort of a back and forth with the, [00:24:00] with the material. I’ll tell you, I’ll tell you what surprised me about the clock, if that’s okay, because one of the odd things is I’ve been, you know, sitting, you know, I think like, you know, everyone have a kind of a hybrid practice now, which means I’m sitting at home in my office with my little clock on the shelf, and it’s, you know, giving me a poem every, every minute, all day.
So it’s in the corner of my eye. So I’m seeing a lot of poems generated by open AI. And that’s, you know, absorbing that. Sometimes, like I said, I use GPT three. And then I realized that actually it was, you know, 10th of the price to use chat GPT. So I’ll use that instead. So that was, you know, the kind of the surprise that wasn’t really a surprise is that the poems generated by chat GPT are rubbish compared to GPT three.
GPT three has much better vocab, right? Like it doesn’t, it does much better poems. So that was kind of, you know, but I’m going to go with the, I’m going to go with the worst poems because there are, they’re, they’re [00:25:00] cheaper. Like, let’s just say that, which is like kind of, I don’t know what that says Literary
persuasions, but you know, practical, practical, practical, practical creativity. The second, the second thing, which surprised me was that it fibs, it makes up the time in order to make the rhyme work about one time in 15, probably, you know, just kind of, you know, the hallucination is, is real. And this is also great.
Right. Because it means now I’ve got a way into very tangibly talk to people about the. The hallucination risk with generative AI. So that was, that was good. The thing that most surprised me is the poems that come out are motivational. They make me feel like, you know, there are poems that come out, which is, you know, you know, it’s such and such time go out and get things done.
[00:26:00] Right. Or, , time is short. Let’s get up and like seize the day. I mean, there are, there are poems like that. I did not put that in the prompt. It is like hanging out with LinkedIn influencers as a tiny screen. And the question for me is like, where does that come from? It’s not in the prompt. It’s in the, it’s in the, the, the tuning given to the turbo model.
This kind of enthusiasm, this kind of get up and go. The reason I kind of find it really intriguing is that do you remember years ago, Facebook did some, an unethical experiment in the newsfeed where they changed. The sentiment of what people saw. Do you remember this one? It was, yeah, they, they sort of like took a, took a group of people and they, they looked at the newsfeed and they selected the posts with the algorithm, ones that had a [00:27:00] negative sentiment and they took another group and everyone had a slightly more positive sentiment in the posts they saw from their friends in the newsfeed.
And then they measured the sentiment of those people in those two groups afterwards. And they found that if you see more negative stuff, you become more negative. If you see more positive stuff, you become positive. Of course, right? Not okay for them to run that experiment, but like now we know chat GPT 100 million people were using that the end of February I don’t know how many people are using it every day now Like I’m not not sure anybody’s put for the figure but it’s a lot right and all of these people are exposed to the exact same kind of experiment that those people in the Facebook groups were which is we are being exposed to something which has a A viewpoint, which in that case is, you know, the case of chat GPT is positive and linked in influence and motivational and go and get it.
And I’m not saying that’s a good thing or a bad thing, just that it’s something that [00:28:00] is worth saying out loud. And maybe, you know, maybe it’s fine that we all have a little tiny Sam Altman gently influencing the way we approach the world. Or maybe, actually, it would be better if we didn’t have that culturally, like, you know, that we had more time, you know, to purchase the kind of.
Meditation or thinking about things, but it’s worth noting, I think, and that’s something which came out of, let’s be clear, a very, you know, quick 30 minute hack assembling parts already had, which there’s no other way I would have come to that, you know, point of view or realization or whatever you want to call it.
Matt: That’s fascinating. And the, the idea that, , the culture that has gone into creating the training data that is then used to be able to power the machine, but then is tweaked and how those cultures feed on each other and how those cultures are [00:29:00] going to be very American, , you know, Western, but particularly American, very individualistic, is maybe that’s part of what’s coming through within that.
It would be fascinating to understand what would happen if you were to say, have the French version. Obviously, Mr. Weston here would be able to help advise and give consulting skills for you on the French version. But if you were to create the same experiment, but if you were to do it so that it was a French poetry clock drawing from.
the corpus of French, , data that has been used to be able to program a French generative AI, what would the sentiment be there? And actually you could try that for many different languages, that in itself would be a really interesting way of being able to start to be able to understand the way in which different cultures represent in online.
, and that becomes a, you know, out of that is this fascinating experiment into [00:30:00] understanding different perceptions, different, this also was inspired actually from a conversation I had earlier today with, wonderful thinker Mark Earls. And we were talking about some of the, the challenges of assuming that particularly things like behavioral science that, everybody thinks like.
American undergraduates because that’s mostly the only time that any experiments have ever been done in behavioral science It’s done on students in america. So there’s a huge cultural bias around in all of that huge types coming back to that point about , the uh, the Having it as a as a machine. I mean I speak as somebody who on my desk in front of me here I’m surrounded with pieces of hardware that are designed to be able to enable me to interact with software in different ways is music kit.
And they all are doing exactly the same thing in many ways But one has got big rubbery buttons and one of them’s got something that looks like a piano keyboard and one over there looks like a saxophone and various other things and and there’s something about the way in which Different hardware forms dramatically changes the way in which I interact with essentially the same [00:31:00] Pieces of software and just coming back to that idea of breaking away from just using a smartphone.
Smartphones are wonderfully amazing things to be able to quickly be able to put software into the hands of many people very quickly, but there’s a huge amount of constraint, I think now built into the smartphones that we know how apps work and they all essentially work the same. So to try to be able to explore something like.
, generative AI, I think it would be extremely difficult within the constraints of the medium of the smartphone now, because, you know, apps is apps is apps, then there’s not a huge amount of variation in them. I
Matt W: wonder if it would be possible to open up new ways of doing, I think it’s possible, right, you know, with social media and with, and with, the.
You know, games, [00:32:00] especially different, different ways of different ways of doing things. I think there are, you know, the, it takes a long time to explore all the possibilities of a medium and maybe, maybe apps aren’t done yet. One of the examples I sort of come back to is like how long it took for the worldwide web itself to be like understood.
You know, you kind of think the web was invented, what, in 1990. First commercial browser in 1994, then it took until 1998 to realize you could put a credit card number into a text field and press submit, and that was e commerce. You know, and then how many years did it take for SAS businesses to overtake boxed software businesses, like 10, maybe 15, like it was a long time.
And I wonder whether the same is, you know, also true a little bit of apps. There’s a couple of other, you know, aside from AI, there’s a couple of other things that I sort of like look at as potential disruptions. I think one is the move to real time multiplayer. The, the [00:33:00] digital world is becoming a social one where we kind of, you know, using things, things like not, not just collaboration.
Right. But like, and, and in games, but, you know, something where we can sense the presence of people around us that that’s happening. and the other thing is like you were saying with little bits of hardware, it’s becoming, you know. Ever easier for the internet to kind of break out of computers.
And I know I’ve been kind of banging that drum for a good kind of 15 years, but I’m still a believer right in the vision of ubiquitous computing and being able to use our, you know, peripheral, perceptions, to interact with. Interact with technology. And I think between all of these things, we may, you know, there might be, there might be life in the old app yet.
I think
Matt: I’m, we had a guest on a couple of weeks ago, who, was talking about a platform called soundscape. Which was originally a Microsoft research project, which was looking at using spatial audio [00:34:00] initially to be able to help people with visual impairment, but actually more broadly than that, to be able to do kind of augmented reality in sound so that you could have spatial positioning of objects in the real world, laid upon the real world through headphones, through, you know, modern headphones.
And one of the things we talked about in that conversation was how creating an entirely new pattern for interaction is really, really tricky. And if you look at what’s been going on with visual virtual reality, there are some conventions that Meta are starting to try to be able to put together of, you know, hand gestures so that you’re not using the controllers all the time.
But still, where they’re getting to is essentially a WIMP interface, where you point on screen with your finger, or you point on a virtual screen. Because it’s so difficult to be able to get people to be able to put the effort into using new interaction patterns. If you look at the way in [00:35:00] which, getting rid of the, the button and using gestures alone on phones, I’d love to know how many people still have a virtual button on their screen because they haven’t got their heads around the ideas of the swiping from the left or swiping from the right or swiping up from the bottom or down from the top.
And they just want to be able to stick with the, the patterns that they’ve got. So I think it, it’s not. It’s not that we can’t change interaction patterns on smartphones, but actually the effort to do it now is getting increasingly difficult. But moreover that it, it, it, it gives a whole load of preconceptions about what a thing is that if you want to explore something new, it gets, you’ve got a whole load of stuff to be able to overcome there to be able to make it not look like the thing that everybody’s used to.
And therefore you’ve got to break over those kinds of UI pattern to be able to then get to a point where you can explore. Something which if you stick a E Ink screen on an Arduino and then start [00:36:00] having poems coming out every minute is a totally different prospect and is immediately obviously different.
Matt W: No, I think you’re right. You need a kind of a point of disruption to get people to, to think about different interaction patterns because you’re changing the mode of interaction just for the sake of it, is not there. And maybe, you know, AI is our, is our way in. It’s actually interesting as well.
Like how, why the clock was a kind of a 30 minute hack, because I think this is also relevant for our discussion about, opening things up. So I, I’ve had a kind of an, I love ink, you know, everyone. Everyone does a little bit, I think, because it’s not a kind of a glowing screen. It kind of takes on the light of the room.
And I’ve had a word clock on my shelf for about three years. The Bartlett Connected Environments program put out some open code to make one on the kit. So I made one. I was building with a startup. I was helping out, prototyping some software about how we might do executive coaching, using generative AI over WhatsApp.
You know, because basically [00:37:00] one of the things about AI means you can get much, much closer to the user. so this was like another project, you know, recently, and I was going, well, how do I prove when I’m, you know, making my WhatsApp thing work, my prototype, what’s my proof of concept that shows that actually I do have everything wired up, you know, I send a message.
I want something to come back, which is almost like a proof of life in the AI. And for whatever reason, I decided it would send back the time and then for whatever reason, I decided it would make a poem. Out of that. Just in order to prove the thing works. And I, I think I’ve done that because I’d also made another software sketch recently about how, you know, you could have a little avatar in a Google Doc and it would tell you who had been there recently.
But, you know, it would do that as poetry. And I think. Why was that? That was because I’d talked somebody from Iceland recently on one of my, on Office Hours calls and we’d been talking about Elves or Trolls or something. [00:38:00] Yeah,
Matt: so,
Matt W: so we were, we were talking about we, you know, because they can’t build roads somewhere because they’re already, yeah, yeah, yeah, absolutely, rocks and all that kind of stuff, right?
So all of this, you know, the, the elves and there’s a nice emoji of an elf, which means that got me into like being Able to do my software sketch easily, which is like a big part of it. So I had this kind of poem, and I came downstairs one morning, and I looked at my clock, and I looked at this demo, and I was like, why don’t I plug those things together?
And I plugged the things together, and I couldn’t stop laughing for about 30 minutes watching this thing. And so I put it on Twitter. And that’s what happened. And, you know, maybe, like, I would, I would love to, you know, do a little shout out for the mailing list at some point, get some, get some folks on who are interested in this.
Maybe we can put a link in, in the show notes later. But, what I found fascinating about this is that, like, the way these things work is these little Lego bricks that you do for different reasons, either to kind of, try and illustrate an interaction, [00:39:00] or try and learn a new skill. or because. You have to do it because it’ll make you laugh.
I’m all just combining two other things and going, well, what will happen? and then you fill in some gaps and now you’ve got another Lego brick you can use in the future. And that’s what kind of carries the whole thing forward, I think. And like, I could say that one of the things that’s happened to technology in the last 10 years is that making those Lego bricks has got really complicated.
You know, the iOS SDK is vast. It’s not possible for anyone to be able to do that. but one of the things about AI, because it’s reasonably early, and, you know, integrating with WhatsApp reasonably early, and it is just text, is we’re making Lego bricks that are the, the right size to be manipulated and combined.
And that lets us break past, you know, the imagination bottleneck, which is where we are. You know, this is my favorite part of like, you know, the technology curve right this point where you can just get your hands dirty and make Lego bricks and combine things and that tells you [00:40:00] something new. And then, you know, you end up from that possibly with a new product.
But if not. You know, you end up with just something very odd instead, and that’s also fun. But yeah, there’s, there’s that bit of the process, and then there’s kind of how do you make it useful in business, and that’s, that’s a different kind of art, I think.[00:41:00]
Chris: So I’d best find out, Mr. Ballantine, since we’re running out of time, what are you doing next
Matt: week? In the week ahead, we have a series of events happening in London, which I’ve got much more active involvement in than I did in the ones in Manchester last week, so I’m running the whole day on Wednesday, which will be entertaining.
And then a trip away, for pleasure, not for work, to the beautiful bohemian city of Prague for a long weekend, which will be very nice. Then back to, back to things again by the time we speak again next week, I guess.
Chris: Hmm. And, how about you, Matt? What’s the, what’s life like for you this week? What have you got waiting for you?
Matt W: I’ve got a fascinating project at the moment with a startup called PartyKit that does real [00:42:00] time multiplayer internet infrastructure where I am… It’s a brilliant title, this. Inventor in residence for a few months, with the idea that by making new things, it can give some direction to the platform. So I am looking at, how groups of people will use AI together.
Chris: Very interesting. That’s certainly got, your work goes out there. It sounds like a lot of pressure being the inventor in residence. People keep knocking on your doors, asking what you’ve come up with today.
Matt: That sounds fascinating. how about you, uh, Chris, what have you got in the week ahead?
Chris: Oh, I’m down in that London again this week, various, things and.
Yeah, and pretty much it, really. It’s that time of year, isn’t it, where everything sort of starts to really roll between now and Christmas, where we can actually get things done, get projects completed.
Matt: Yeah, it’s two months now before December and everybody starts to get distracted
by Christmas [00:43:00] and the break.
Chris: All of that. Exactly.
Matt: Fabulous. Well, we will be back again next week. thank you again, Matt, for joining us on the show this week. Much appreciated. Thanks for having me. It’s been delightful. And, we will be back same place, same time next week.[00:44:00]
Matt W: Thank you for listening to WB40. You can find us on the internet at wb40podcast. com, on Twitter at WB40podcast, and on all good podcasting platforms.
