Near Future Podcast

Tom and Jonny reflect on their live workshop at Product Unleashed in London, explore the history of designers building their own tools, and dig into how AI can be used as a counterbalance to your own working style — not just an amplifier.

Chapters

Building a Design Tool in 15 Minutes at Product Unleashed
Tom and Jonny recap the workshop they ran at Product Unleashed, where they built retro design tools (Etch A Sketch, MS Paint) live in front of an audience — twice. They discuss how the challenge evolved, why they pivoted from serious tools to fun ones, and how they got the final result live on the internet using a QR code.

Why Designers Building Their Own Tools Matters
Jonny shares the narrative from the workshop's opening slides — a Gutenberg-moment argument for why designers are now able to build hyper-personalized tools for the first time, referencing Adobe, Figma, and the growing landscape of AI-native design tools.

Historical Examples: Eames, Lennon & the Apple Calculator
Tom walks through three historical examples of makers building tools to unlock their own creativity: Chris Espinoza's Calculator Construction Set for Steve Jobs, Ken Townsend inventing automatic double tracking for John Lennon, and Charles and Ray Eames creating the Kazam machine to bend plywood into new furniture forms.

The Value of Craft in the Age of AI
Jonny reflects on Karri Saaranen (founder of Linear) and his view that design is the thinking — and there's no shortcut to it. The conversation explores whether something made quickly can ever be truly iconic, and why craft still has real, felt value even as AI capabilities grow.

AI as Amplifier or Counterbalance: Knowing Your Working Style
Tom introduces the idea that AI shouldn't just universally amplify — for some people (perfectionists who stall), it should act as a counterbalance, nudging them to ship. For others (fast movers who skip deliberation), the opposite. Both share personal examples of how they've tuned their own AI workflows around their personality types.

Ooda: Instantly Sharing Internal Prototypes
Jonny shares an update on Ooda, his tool that lets teams publish prototypes, tools, and documents to a private, secure link in seconds — just by asking Claude. He explains the pivot from "run your dev server in the cloud" to "don't let your team's work disappear," and why the response from a handful of people last week convinced him the problem is real.

Building a Calculator with a 5-Year-Old
Tom shares a heartwarming weekend story of building a calculator with his five-year-old son Waltie — complete with a car button, a horse button, and a Lego-themed interface — and reflects on what it means to grow up in an era where building software is as natural as playing. Jonny draws parallels to how touchscreens changed children's expectations of interaction.

  • (00:00) - Episode #3:
  • (00:04) - Building a Design Tool in 15 Minutes at Product Unleashed
  • (07:45) - Why Designers Building Their Own Tools Matters
  • (12:13) - Historical Examples: Eames, Lennon & the Apple Calculator
  • (19:57) - The Value of Craft in the Age of AI
  • (22:45) - AI as Amplifier or Counterbalance: Knowing Your Working Style
  • (34:32) - Ooda: Instantly Sharing Internal Prototypes
  • (43:59) - Building a Calculator with a 5-Year-Old

Find us at nearfuture.works

Creators and Guests

Host
Jonny Burch
Co-founder of Near Future
Host
Tom Harman
Co-founder of Near Future

What is Near Future Podcast?

A regular podcast covering design and AI from the founders of Near Future, a boutique AI consultancy focused on teams that care about craft. We cover both what we're seeing on the ground and industry trends, ways of working and occasional guests from the design world.

Episode #3:
===

[00:00:00]

Building a Design Tool in 15 Minutes at Product Unleashed
---

Tom: hello. Welcome to the Near Future Podcast. I'm Tom, this is Jonny and we're gonna, give you a quick update on how we're thinking about what we're building, how, what we're exploring in AI and design right now. So, I thought we could start out by talking a bit about a workshop that we ran on Friday evening, Friday night, the prime time for design workshops, at an event called Product Unleashed in London, where we built a design tool in 15 minutes twice.

so it was a fun event. There were a bunch of, great talks that happened. there are four talks about designers exploring AI, and then we were the, dessert. We, we, we, we were the bit afterwards, in a separate room, in the top of this building where we ran a workshop on building your own design tool, twice, as I mentioned.

So, Jonny I thought we could talk a little bit about how we went about that, [00:01:00] any high-level takeaways or themes from the event, and perhaps give a bit of a behind-the-scenes look at, building out that workshop. So how does that sound to you?

Jonny: Yeah, sounds good. Yeah, it was a interesting challenge that I think we set for ourselves really. There were-- no one asked us to try and build a design tool in 15 minutes. decided to do that, and then went through a process where we realized that it was gonna be slightly more complicated or at least less predictable we thought it was. In fact, I think I probably should blame myself. I think I, responded to the email with like, "I th- I-- we should just build a design tool." And then as we actually started to run tests of what you could build in 15 minutes, I think we probably realized at that point that while it is possible, absolutely possible to build a design tool in 15 minutes, quite a lot of that time is just watching Claude or your LLM [00:02:00] and then write code, and it's not very interesting. and we actually ran a straw poll during the workshop of who was familiar with these tools, and pretty much everyone was using Claude in Terminal. So I think it wouldn't have even been interesting to people who hadn't seen it before. weren't, there weren't that many people that hadn't seen it before.

So, w-we sort of guessed that, and quite quickly realized that not only were we gonna need to build a design tool, we were actually still also going to have to present to think about a narrative and a presentation and all of the talk aspects. wasn't us getting out of doing a talk, this was just us adding an extra thing on top of doing a talk. also working out how to do it together in 15 minutes, which we're-- I think we're pretty good at now. We're like, we're, we're a fairly good double act. We know when to sort of hand it over to the other person or to butt in, but all quite complex. So, [00:03:00] We-- I think the first, the first ideas we had were sort of slightly more serious design tools or actually creating an artifact on the screen like a button or a bouncing ball or something.

We've done-- we've run these things before, and it works quite nicely. And then you go on and you prompt the sidebar or the, the controls to be able to manipulate whatever it is that you have on screen. And quite quickly, I think we tried that and just sort of felt like the temptation was gonna be to directly manipulate the object rather than build the tool around it.

Or at least that's what my fear was. And so switched to, "No, let's just build an actual tool." The problem is a lot of design tools are quite complicated, it maybe would've been a little bit dry to watch someone try and the easing curves on a hover state for fifteen minutes. That's not, you know, we're eating [00:04:00] pizza.

It's sort of eight thirty PM. We've had or two alcoholic drinks. This is not the kind of content that people are gonna be asking for. So we pivoted again to, building retro tools that we thought at least even if it doesn't work, we should get a laugh. So we ended up with, three options. Etch A Sketch was one, and that was, that was chosen. the, the audience chose which tools we would ultimately build. So Etch A Sketch, WordArt, and the third was Paint, MS Paint. And, that-- I think that worked really well actually. Like, and what's amazing about it and the sort of mind-blowing thing about these coding models and their training data and all of the stuff they already know about the world without you having to tell them is all you had to do to build a fairly complete version of one of these tools, whether [00:05:00] it's an Etch A Sketch or a MS Paint, is write build MSPaint, and actually within about a minute it's there. So the fifteen minutes was never at risk in terms of actually getting a sort of basic version of a thing that looked recognizable, and people could and understand what it was. I think the harder thing ended up being what do we then do next, and how do we get people to get excited about wacky ideas and, and feature requests?

And then how do we actually fill the time where people aren't doing that with a bit of why, a bit of a narrative? it probably ended up taking us a bit longer than we would have hoped to, to get it to the point where it worked. But I think then it worked. So, I think it was entertaining and sort of funny, hopefully we landed a little bit of a serious point in it as well.

But yeah, Tom, what did you

Tom: Yeah. Yeah, I thought it was, I thought it was really fun. Definitely agree with you that we, on paper it [00:06:00] looked like we, "Oh, we're just gonna have a terminal open and run through this thing. We don't really need to prepare very much." but in the end it's like that plus a slide deck, plus how do we hand off between each other.

I thought it was- I thought it went really well. I, And if you weren't there, if you didn't see that, then the Etch A Sketch we created the first time round was this trippy, rainbow-colored animating Etch A Sketch experience that, was kind of quite wacky, quite bizarre. worked very well on the mobile though.

and there was this neat moment at the end of the workshop where we, uploaded it to Udo, which is a tool that you've built, Jonny and that got it online very smoothly, showed up a QR code for the audience to scan with their phone, and then they could play with it on their, their phones. so yeah, I was quite impressed with how the technical details held up under that kind of environment.

[00:07:00] and the narrative was, was fun as well. Like, it, it was more, as we said, more work than we anticipated, but it was fun to dig into some of the, the narrative in a bit more detail, that allowed us to explore a bit more of the history of tool making in design, which I'll speak to in a moment. But yeah, I wonder if you might actually wanna speak to some of the slides that you shared at the start, which were about giving some context into why this matters.

'Cause I think if you just look on the surface of building an Etch A Sketch, in a group in 15 minutes, it was a fun exercise, but maybe not super relevant for designers practicing day to day and what they might wanna do. but I think there is a, a real depth to why this matters. So yeah, would love to hear your thoughts on that, Jonny

Jonny: Yeah.

Why Designers Building Their Own Tools Matters
---

Jonny: this isn't the first time we've sort of told this story and, and it-- I've-- I think that this, topic is something that I keep coming back to myself in terms of how I think about how this technology's gonna impact us. and I know in the [00:08:00] last episode we talked about there's three different ways in which, broad buckets in which a designer might be able to utilize AI in their day-to-day.

One is to production code, one is to build and non-production code but, but, enable their design work to be brought to life in a, in some better way, and then the third is to build tools to better allow them to design whatever it is they're designing, whether that's in code or not. and I think that that tooling one is basically the story that-- I think that I wanted to tell at the start of this workshop.

I'll see if I can get the slides up, actually. oh yeah, here we go. so I've just-

Tom: If you're watching this on YouTube, you, you actually have something to look at other than our faces. But if you're listening to this only, we will, try and narrate these so that they make sense without needing to look at a screen

Jonny: Okay, here we go. Yeah, so the fir- I mean, the f- the first, slide is just a [00:09:00] picture of a man that is actually Johannes Gutenberg. I think Gutenberg is probably a bit of a cliché to bring up as part of a narrative around technology inducing enormous societal change, but I did it anyway, YOLO. and I think that w- broadly, in, like, this Gutenberg moment now, this post-Gutenberg moment.

AI is the great enabler. There are second, third, fourth order consequences of this technology that we're yet to see. and things that were held in very few people's hands before the printing press appeared suddenly were more broadly available and books got, probably not cheap immediately, but, but became much more accessible and therefore knowledge was disseminated, et cetera, et cetera.

We're now in that situation, where we're no longer beholden to two or three companies that are very large and very far away, [00:10:00] very, with, roadmaps that you can't easily yourself for your own ends. And we're now actually able to choose from tools that already exist. I think, yeah, there's now I think nearly 100 tools on this website, designtools.fyi, that I'm, I, I have on a slide if you're listening. so there's lots of paid options still built by and, and individuals, but actually even more than that, we're now able to build our own. So that was, like, the sort of narrative that I wanted to go with. and I don't know, like I had about three minutes to explain it, as with this whole thing, is 15 minutes is nowhere near long enough, but maybe someone got it.

I think there were a lot of nods on the slide, on the Adobe and Figma slide. Lots of people agreeing hard with that at the very least. So hopefully, something happened. [00:11:00] and then I got to use my, my hardcore meme skills. I got to be a meme lord with, this is one of my favorite images to meme on, which is a man pouring a lot of olive oil on a salad, and the man is us, and the olive oil is features, and the salad is this tool.

So if you're not looking at it, it's not it might not make much sense, but, kept asking for more features throughout this process, so there's occasional slides with requests to the audience for more features. Anyway, yeah

Tom: great. Yes. and there were some, some good feature requests. I was impressed that the audience were pretty engaged with, suggesting ideas. the, I think the second time round we get the horse button in,

Jonny: Yeah

Tom: maybe. yeah, lots of rainbows and animation and, yeah, it was very fun

Jonny: So Tom, as part of this, [00:12:00] presentation, we also threw in some historical examples, and while I have the slideshow up, we could dig into them if you're up for it.

Tom: Sure. Sure, yeah

Jonny: interesting

Tom: Yeah. So,

Historical Examples: Eames, Lennon & the Apple Calculator
---

Tom: I think the reason we wanted to add a bit of history and a bit of context here is, is that all of this tool making seems to be exploding on the internet right now. Everyone's designing their own tool for their, their own process, which is really exciting to see. but it feels like maybe this is a new phenomenon, but actually it's-- there's a long history of people designing their own tools and designers designing their own tools.

So a few historical examples we spoke to-- spoke about in the narrative. One is the, the classic example of Steve Jobs, and Apple, where Chris Espinoza, early employee of Apple, got fed up of, being micromanaged by Steve Jobs' attention to detail and creative vision, and built him, the calculator construction set, which allowed Steve to edit and [00:13:00] manipulate every minute detail of the, the cal- calculator UI to get it just right for, to perfect that experience.

so that's one, one good example, the classic example perhaps. we also talked about another couple though. the second one was around the-- this idea of automatic double tracking, which was invented by or, inspired by John Lennon of The Beatles double tracking his vocal, which has this amazing, audio quality to it.

But John wasn't a big fan of needing to manually do two takes of all the songs they were recording. so the audio engineer there, Ken Townsend, decided to, invent this way of automatically double tracking vocals or John's vocals, which allowed them to keep this sound-based, signature move that they had.

you will hear it a lot in [00:14:00] Revolver if you listen to that album. but it allowed them to do it m-much more, efficiently and easily and sort of all- gave them that creative power, but without needing John to go through this gauntlet of painful extra work that he didn't need to do. So a cool example there.

The final example we, we spoke to the crowd about was, in, be-- even before this, in the 1940s and '50s, Charles and Ray Eames invented the Kazam machine, which allowed them to bend plywood into these new shapes so that furniture could be built, with these new forms that didn't really exist un-until this time.

So it wasn't that it wasn't available to them, it wasn't available to anyone to bend materials into these shapes before the Kazam machine. so I think it's just a really cool example of how the idea for a vis- or the vision for a product in someone's head isn't even possible to make with the tools available today.

You [00:15:00] need to first build the tools in order to unlock what those tools are capable of producing. so I think this sort of second order effect of how to think about designing products in this sequence is really exciting. So those are the three we spoke to the crowd about. Jonny did you have any e- extras that you wanted to mention that were left on the casting room floor or any other ideas for historical references for designers or creative people building their own tools?

Jonny: Yeah, I mean it, this sort of symbiotic relationship between a person and their tool, you like... I love the Eames one because you, created the thing that allowed them to invent the, their iconic chair, amongst other things. as far as I understand that the jig that they use for the Eames chair today is still the same one, the same mold as they created back in the day, you know.

The- there's lots of fake Eames chairs. You can buy, replica ones and, and d- for, for [00:16:00] considerably cheaper. But if you want to, if you want a chair that was made in the same mold that they were using in the '50s, then, you really can get that. That's, you know, a new Eames chair still uses that same simple tool.

You know, it's a shape, basically. That's all it is. that shape is iconic and has lasted for decades and, has led to this extraordinary sort of legacy of, of design. So I don't know. It's, it's a ve- I love this one. I think that, that the, the Kazan machine and then that methodology that then led them to, to create all of this wonderful furniture is just extremely romantic to me. And wow, just imagine if any of us get to create something that lives as long and large as, as, as this has done. don't know. I sort of feel like, you know, we, we, work in pixels and on the internet, and it's all a little bit more ephemeral than this, and maybe all [00:17:00] digital product designers just really want to design chairs and lamps and things because, you know, it'll still be around in hundreds of years. But, oh well. We're doing this. Our websites may get taken down one day, but, in the meantime, yeah, this is a, this is was my favorite one. I don't have anything else. I just like this one

Tom: Yeah. Yeah, yeah, yeah. It's a-- That's a good, a good comparison with the ephemeral versus the sort of physical world. And yeah, I do think there's something that is maybe a little bit more unique to this moment is the instant and temporary nature of these tools. Like they don't-- I, I love this example as well with Charles and Ray Eames, but it is, something that has lasted an incredibly long time, and it is its own mini form of manufacturing almost.

Whereas the idea that you can prompt a tool i-in 15 minutes that can directly help you shape not only the output [00:18:00] of your work, but the way you think about what is-- what your work is capable of through this new tool that exists, I think is, really interesting. And I think there's something that's powerful about that temporary and very hyper-personalized, hyper-specific use case that

Jonny: Yeah

Tom: think could, could unlock some really interesting things

Jonny: Yeah, I think like th-this, this being a jig or like a mold, you-- I think you would think of something like that as being really temporary to go back to furniture making or carpentry or something. Like, I would make something in order to put a in exactly the right place in all four corners of a bit of wood, and then I'll throw it away, and I never need that anymore.

And I think that that's m-a much closer parallel to a lot of what you see, like prototypes and, and you know, these sort of one-off tools and they have an interaction and then never look at the tool again. and I do think that there is this debate at the moment about whether anything that is so easily and cheaply [00:19:00] made can ever be really valuable. is there... there need to be pain and toil and and, you know, failure and all of the other things probably led to something like this and that led to all iconic pieces of design? Is-- Do you need that in order to create something iconic, or can something be created quickly that becomes iconic? I think, if you read-- There's this guy, Karri Saaranen, who founder of Linear, and he's like very-- having some very interesting nuanced take on all of this because he's very pro AI, and they build a lot of AI stuff into Linear and things. But as a designer, he's very skeptical of shortcuts around the design process.

He thinks design should be the work is the thinking, and there's no shortcut to the thinking, et cetera, et cetera.

The Value of Craft in the Age of AI
---

Jonny: I'm increasingly subscribing to that. [00:20:00] I think I went through this like hype cycle of like we're all doomed a little bit, because I couldn't see where the technology would stop being good. And now I think, probably I think like a lot of people at the moment, I'm starting to be like, "Oh, well, it's good at these few things, but What I really value is something that was created by a person and really thought through, and you can feel the difference. I really think you can feel the difference still. And maybe there'll be a point in time where we hit some singularity, but until then... Yeah, like stuff like this, I, I'm not convinced that, with all of, all of the design skills, as in AI agent skills and, amazing models, et cetera, et cetera, I'm not sure that the value of design will lessen. The value of, of something that has been crafted will lessen. Which sort of speaks to our mission, I think, as [00:21:00] a, as a business. Like we're-- we care about teams that care about craft, and so then our position on AI is necessarily more nuanced and not just throw, throw agents at it and throw skills at it, and then it's good enough. Like no, how do you actually-- how do you maintain that bar and, and still build things of value?

And given you still have the same amount of time, and time is necessary, where do you put that time and things like that.

Tom: Yeah. Yeah.

Jonny: This is-- That was a real, tangent,

Tom: Craft,

Jonny: Yeah

Tom: pension Yeah. But I, I, I like, I like the tangent though. I, I think, I, and I think to your example there, it w- it's everything you described there isn't necessarily a new thing that is now possible because of AI. AI has just opened up new leverage, new ways of amplifying human craft. I mean, I think there's maybe a separate [00:22:00] conversation about When the products that are being built are to be consumed by agents, maybe that changes the, the form factor of them and the, how, how you do-- what-- how you define good design when the audience is not human.

but, but assuming the audience is human, I am inclined to, to agree the sort of authentic human judgment and editorial, taste is, increasingly important. And I've-- actually, sort of an adjacent tangent to that is, I've been thinking more recently about, personality traits rather than AI being universally good or bad at something.

So,

AI as Amplifier or Counterbalance: Knowing Your Working Style
---

Tom: what I mean by that is using... In the pursuit of better craft and using AI to achieve better craft, there's often a narrative about AI being an amplifier. so it might amplify w- your range of creative [00:23:00] thinking or, and getting to a prototype faster to generate an idea for the thing you want to, to build.

And maybe those are two examples that are fairly universally applicable. But I think there's some interesting things around whether AI plays the role of the amplifier or AI plays the role of more of a counterbalance to someone's default instinct. And if you're a craftsperson who is perhaps a, a perfectionist who doesn't want to ship things, and you just wanna focus on, like, crafting the details, maybe there are ways that AI could be used to counterbalance that instinct and actually get you to, like, share things more, like, earlier or, reassure you that actually sharing this openly with a certain audience is going to be a positive thing to get feedback on your idea and get it-- im-improve it in the long, long term.

Whereas if you're a craftsperson who is maybe less obsessed with details and hungry to get something out quicker, you might need AI to counterbalance you in the other way. so I think there's, I don't know, potentially a, a [00:24:00] theme around someone's unique profile of how they think, how they make, using AI to not just universally amplify all the things but really target where it can amplify based on your profile as a, as a person.

And so this is, spoken of someone who overthinks things and, and Jonny you and I have talked before about maybe my bar for getting something out is, high, higher than it should be sometimes. and so I've been increasingly trying to use AI to, like, how do I lower it a little bit to, like,

Jonny: Mm-hmm.

Tom: get an idea out there or, like, just get me over the line of publishing something or putting out an idea or building a thing, with the view toward the longer term goal of that being to help inform and iterate on the thing, whatever artifact it is, to get it ultimately to a better place.

But I think that really requires you to know yourself and know how you operate to be able to intentionally [00:25:00] get AI to nudge you in the direction that is, taking you on the right path. 'Cause I think there's definitely a trap to fall into where you're a, a narcissist and you want AI to just tell you you're great all the time, and I don- don't think that ends up in a positive loop.

But if you're very self-critical and AI is a n- a sycophant, maybe, I don't know, maybe there's a bit more balance there. maybe it's not a universally bad thing

Jonny: Yeah, really interesting. I, I mean, I do think I've v- probably a slightly different use case, but found AI most useful in shoring up weaknesses of m- of mine, I think. Whether that's literally a technical weakness, like built things that I could never build because, you know, my coding ability is okay but not, n- not great, and definitely things like infrastructure and, and how to think through that is-- be challenging without AI.

There's definitely, definitely a, a, a gap for me there. But also, I'm a activator personality type that builds things and then generally struggles to stay excited about them over the medium to [00:26:00] long term. So having building routines that involve agents that do things and then surface them to me in ways that make me think, "Oh, actually, I'm quite in-- kind of interested in that."

You know, so I respond pretty well to finding out a bit of information that I want to read more about, whereas a blank sheet of paper or just like a to-do, I struggle with. So how do I build sort of agent infrastructure to pique my curiosity unpredictably every day so that it feels like something new, even if it's the same old boring thing?

Tom: Mm-hmm.

Jonny: for... to, to your point, I think you're in the right spot to leverage AI because I think AI is way better at encouraging speed than it is at encouraging slowness. I think, for people that, and I'm slightly more in this direction, for people that are naturally inclined to just yolo stuff out Unless you [00:27:00] very intentionally build ways for AI to, to review and stop what you're doing and, I don't know, like slow you down, it's just not its will.

It's not what it want. It wants to move fast. It wants to overproduce and get things done incredibly quickly. So working out how you would actually use it to Help you deliberate and quest- question your decisions and sort of, you from shipping the thing you wanna ship. For me anyway, I don't know. think it's slightly less good at that basically at the moment. That's not to say it won't ever be, and it's not to say it can't be if you work out how to make it work. But, I actually think humans are uniquely good at that or more good at that than, than AI today

Tom: I, I agree with you. I th- I also think that's kind of fascinating 'cause I w- went through this process of trying to create a Claude skill or I think it was a Claude skill to help me [00:28:00] write more publicly. But the way I originally wrote it was, or prompted it, was with slowness in mind. not-- It wasn't framed as slowness, it was m- framed more around rigor.

like,

Jonny: And then

Tom: less about posting publicly and more about sharpening my own thinking to the point that I would reach th-this stage where I had this well-formed, well-opinionated, well-defended idea of something that I could then share. Because then it wasn't really sharing anyth- like I w- I felt very bought into every aspect of sharing it.

but basically the focus was on interrogating the idea rather than just getting something live. and I think that might work if y- if I was writing a, I don't know, a book that was going to print and needed, much more rigor around it. But I, as we just discussed, have the opposite problem where actually I need AI to tell me this is good enough.

Like, you're just posting something on Link- [00:29:00] LinkedIn, like it doesn't really... No one's gonna read, gonna spend that much time reading it, like it's good enough. and so I sort of retuned it for less about slowness and rigor and more about compensating for my own bias, because I think I would sl-stall myself in getting something live without that.

but it does require some-

Jonny: it d-did

Tom: Oh, yeah.

Jonny: help your, yeah

Tom: Yeah, yeah. I, I generally have it... It's, I mean, it's kind of been interesting going through, through this now where it, I use it to, I use AI to help sh- push back on the ideas that I'm thinking about or like sharpen my thinking or challenge some of the thinking.

I've generally tried to not have it write very much, but it's more just an editor giving me feedback.

Jonny: Yeah

Tom: but then if I, if I give it like three iterations or four iterations, after that it will just stop telling me, giving me feedback and just say, "You should [00:30:00] post this now. Like it's good. It's good enough."

Jonny: good. Yeah, yeah. That was handy.

Tom: so, so it's-- I guess it's those kind of personality quirks that I think do require a bit of self-awareness or at least,

Jonny: Yeah. Yeah, I see

Tom: a,

Jonny: mean

Tom: at least a sort of clear goal in what I'm trying to get done.

Jonny: You have to understand why, why the thing you want to do isn't happening in a fairly crisp way

Tom: Yeah and I think maybe that comes down to the more you understand your own way of operating, the better you can prompt AI to help get you to either expose your blind spot or help accelerate you or amplify from what you're doing.

Jonny: Yeah

Tom: and I guess r- yeah, posting, writing publicly is, is one specific use case.

Might not be relevant f- to all designers, probably isn't relevant to all designers, but, but it is this sort of self-contained thing where you can really push-- have AI push against different angles based on your own in- insecurities or weaknesses or, or [00:31:00] just profile your own shape. and so for me, I found that quite helpful at meaning that the AI is less about...

It's a little bit of judgment, but less judgmenty, and I'm, I'm bringing a lot of the judgment, but it's helping me sort of light- lighten, lighten that expectation a bit on myself, which has been really useful

Jonny: Yeah. Well, that's great. I mean, you just said that these skills are very personal and, and it's about knowing yourself. But even so, I think it'd be kind of interesting to see-- put that somewhere and, and see if, it's more broadly applicable, like a writing assistant that isn't just producing slop because I think a lot of people are like, "Well, I don't want AI to write for me because it's...

The, the output is gonna be poor." But actually, what I would love is to just have something to me in the flow and get me feeling confident to do it myself on some level. So an AI editor rather than an AI writer, I think is pretty compelling. I'd use it

Tom: Yeah. Yeah

Well, I think, I [00:32:00] think there's, Well, thank, thank you. I think there's another a- angle of this, which is around just how you process things. Like how, like, I don't know, I feel like I need to talk things through. I mean, that's partly why we, to an extent, why we might- maybe why we start- started this podcast, to have more of an excuse to talk things th- through together.

But, but I find that really important to just processing information and, getting clearer in my thinking. so perhaps that's another angle of this, of how you, how you... Knowing how you reach that clarity is an important thing to bake into how you operate, and then find ways for AI to leverage that, to, to build on that, but not replace it.

And, and I guess that all comes back around to the, the craft point you were saying at the start, which is how do you use AI as this mirror or this amplifier throughout your creative process to make sure it is amplifying the things that extend your thinking or extend your capacity, or count- [00:33:00] counterbalance and push back, based on what you uniquely want to get out of it

Jonny: Yeah

and fundamentally not remove the things that you love about the work. And

Tom: Absolutely.

Jonny: there'll be some things that people love that will be automated away, unfortunately, over the next however many months and years because of this technology. But I think there's plenty that doesn't need to be, and I think there'll be plenty of opportunities to spend more time doing the things you love by making everything else more efficient around the, the edges of it. I think we are, doing extraordinary amounts of low leverage work as a society still that is g- is a great opportunity, that would... is a utopian, but, but would help make space for fun stuff that still is worth paying humans for. So that's, that's my optimism showing. I'm sorry,

Tom: Yeah. Yeah. No, absolutely. I think if there's, [00:34:00] if there's one thing we're able to help with or for in teams who care about craft, that one thing might be taking off, taking a lot of that low leverage work off people's plates so they can spend more time focusing on craft. Doesn't necessarily mean AI has to focus on the craft.

but there's potentially opportunities for that too

Jonny: Yeah, for sure. let's move

Tom: L-

Jonny: I've

Tom: yeah

Jonny: I've been showing this slide for ages and... But we were talking, we were talking around it for a long time, so that's, I think that's

Tom: Yeah. That's,

Jonny: so what's next on the docket, Tom?

Tom: So,

Ooda: Instantly Sharing Internal Prototypes
---

Tom: I would love to ask you a bit more about Ooda. So you mentioned and we talked about in that, demo on, on Friday, the last step of our demo was getting the design tool that we built with the, with the audience live on the internet. and that worked incredibly smoothly, and that was thanks to the tool that you had built, Ooda.

So could you tell us a little bit more about that and, and what has been, have you been thinking about it this week?

Jonny: Yeah, I think Uda had quite a big week last week [00:35:00] because of two things that happened. Well, no, I think it was just one thing that happened Was at the end of the week before-- I'd be- I'd been building this tool or, like, playing around with it and trying to work out if it's interesting for of months. And the original idea was around this concept of local host is hard for non-technical people and risky. Like, I want to run my, coding agent with no permissions because it's annoying, and I never read them anyway, and I don't know what they mean, so I'm just gonna turn permissions off. But then, you know, it could delete my file system, and I need to get all of these secret keys and API keys and passwords and things from an engineer, and then I need to ask them how to help me set it up.

And I've got to run all sorts of commands that look scary. And then after all of that, I've got loads of company IP sitting on my laptop, and, [00:36:00] I don't know if there's-- I'm adding enough value to really need it there. and so the original idea was, like, everyone should just be able to code in the cloud.

Like, local host will s-look silly in a few months' time, in a few years' time. So, and everyone will be in their cloud sandboxes, so, build those sandboxes, but specifically for non-technical people. So that was the, like, original and I actually got the technology working pretty smoothly, and I, I really enjoyed.

And to, to this whole like what is a designer and what do we design when it's for agents thing, designed a CLI and like loads of non-UI sort of design work. Had a lot of fun learning about how to, how to think through that. And it ultimately ended up with a piece of technology that s- that, like, worked and I felt pretty good at and, about, but of a black box, I sort of couldn't-- didn't feel that confident [00:37:00] that I could describe why the, why someone should care. And then a week or two ago, I sort of realized that the thing that really mattered probably was getting things-- the finished thing live somewhere for someone else to look at. We have all of these knowledge management platforms for, like, documents and, and, rest of our work. And then I was talking to a scale design team who said, "Oh, they tried out Cloud code," but they didn't have anywhere to put the things they'd built, so they went back to Figma Mate because they've sort of solved that. Figma have solved the sharing the link problem. And I just thought, sharing the link problem isn't a big problem. It doesn't feel like, you know, frontier technology. It feels solved. It just so happens that Figma bothered to do it, and it's where the design work's already happening. So, so I thought, "Actually, I've built... I've actually built it already, all of the stuff required to [00:38:00] solve that," it's just a sort of side project within Ooda so I just flipped the sort of narrative to m- in my own head. I was like, "Oh, well, how about I tell the story of Ooda being instead of build, run your dev server in the cloud," which is what it was. Instead, just publish stuff to the cloud, but it's not public, it's only for inside your organization.

Like really easy. So I changed the H1 In fact, I, I could probably pull it up. I changed like three words in the, the-- on the website from, let your, dev environments fly away" or something to, "Don't let your team's websites fly away." And I changed a few bits lower down the page, like Publish in Seconds, Sites are Secure by Default was, was sort of a new section to really land the idea of like prototypes, tools, and documents. So just to like really try and sort of be really [00:39:00] simple on the proposition. I shared it with a or three people last week, and like every single person was like, "Oh yeah, I'm trying to solve this at the moment." And w-- just one of those, you know as, as a, of being a founder before and you, you sort of know it when you hear it, people going, at least get this."

You know? They might not all be customers, but they understand what the problem is immediately, and they wanna talk to you about it a little bit more, or they have a take on it, or they have like a here's the nuance for their organization. And so like one person booked in straight away for a call to see if they could get access.

Another person listed out a ton of use cases and was like, "How many of these can you solve?" and probably about half of them at the moment, but no reason why I couldn't solve more of them. So I was like, "Okay. Well, now I feel excited about this again, having been not sure what to do with it for a couple of months."

And then thought, "Oh, the perfect way to force myself to m-- get [00:40:00] it working is to use it in our fifteen-minute tool creation, And so I spent a couple of hours on Friday just making sure it worked really well. So now basically all you need to do is after you've logged into Ooda, because it's only for teams, so you have to have a, an Ooda account, all you do is ask Claude to publish this site on Ooda, and it knows exactly what to do, and within about a minute, probably thirty seconds, you'll have your full website prototype, whatever it is, live on a link that is only accessible to people who are within your Ooda organization.

So b-- you know, i-it's a login by default. You have to log in to, to see it, which just sort of solves a little pain point that I think is interesting, and I'm not sure many other people are doing. So really early days and, right now I'm [00:41:00] just like excited about What it could be. The-- I think the most interesting thing for this conversation though is that almost none of it has been designing a UI. Like, the UI looks pretty awful. Claude has done most of it. To get the whole thing working and to prove the value of it, almost all of the value is felt in terminal because it's... Once you've lo- created your account and logged in once, you sort of never have to go back to it. You ask it to give you a link, it gives you a link back in terminal, you share it. You can change sharing settings and do everything else from terminal, AKA via your agent. You just ask your agent to do it. So, it's like a UI-less platform, basically. The, the UI is only for managing other team members and things really at the moment.

So, it's been quite-- It's very like a product of, I suppose, a product of 2026, and, and how we're doing things. But, yeah, so the, been fun [00:42:00] to, fun to play with. I'm excited to see where it goes. Don't know how to charge for it or anything like that yet. I don't know what the plan is. But, we'll see what...

Yeah, see, see what happens from here

Tom: Super cool. Hearing you talk about it like that makes me, I don't know about this reference, but it reminds me of the early days of, of Dropbox, where it was just like a folder on your computer and you're just like, "Oh, I want to share this file with somebody, so I'll just drop it into this folder," and then making the, the journey, the round trip to share something as short as possible and, and independent of an existing tech stack that someone's a part of, but with like really secure permissions.

so I think that feels like a really valuable-

Jonny: of it is the thing that excites me the most. Yesterday, was working on it actually, and, and Claude wrote me a big plan for a piece of, that I was thinking about building, and I was like, "I don't wanna read this inside terminal. It's so, so hard to read." So I was like, "I'll just put it on [00:43:00] Ooda, send me the link," and I read it in a nice HTML page. the fact that, A, that was very quick to do, and B, it sort of means that these-- that websites are ephemeral now. You know, it doesn't-- You don't need to, like, create a new site on Netlify and, like, click a, a billion things or, you know, Vercel or all of this kind of stuff. just feels so simple to do that, it sort of-- It, it's that, again, the tools, the symbiotic relationship between a person and their tools.

Like, because I have that tool now, I might do things in a different way, which may lead to other consequences and interesting thoughts and ways of doing things. So yeah, I think it sort of all comes back to that. But yeah, I'm ex-I'm excited about it. I'll probably talk about it more on future episodes, so,

Tom: Thanks for sharing.

Jonny: Pleasure

Tom: I,

Building a Calculator with a 5-Year-Old
---

Tom: I know we're [00:44:00] coming up on our, time, but I wanted to share one, one, fun thing about over the weekend, and I would love to hear anything else fun that you wanna share too, Jonny But, but after Friday

Jonny: You haven't told me about this yet, intentionally, so I, I'm excited

Tom: so on, on Friday, so my five-year-old son, Waltie, was, home from school on Thursday and Friday 'cause he was, he was sick. and so I was looking after him for a bit, and we were talking about what I was gonna be doing on Friday night, and I told him I'd be building this tool, with a group of people and, and that we could maybe do a test run together, like a practice run.

and I told him that we'd be b- building a calculator. And he's five, so I was asking him, "Do you know what a calculator is?" "Yeah." "Have you encountered one?" And he, he knows basic arithmetic, but I don't think he's ever used a calculator before. anyway, he, he... Well, he claimed that he knew what a calculator was, but the, but the [00:45:00] thing I forgot about was I was about to leave on Friday, and he was a bit poorly, so he was not maybe his, his usual self during the day.

but just as I was about to leave, he said, "Oh, but, but Daddy, you said you were gonna build a calculator with me, and we haven't done it yet." And so I felt really guilty leaving on Friday night going to this event. And then one of the talks there, Ini, o- one of the speakers was talking about her seven-year-old daughter and how they have, been playing with AI together and what that has taught her daughter about, these amazing new tools that, what they're capable of.

It was a very inspiring talk, and it was focused on, finding moments of play and imagination, throughout the weekend. And so that directly led me to thinking, "Oh, I need to make sure I do this over the weekend." So on Saturday, I sat down with Waltie, and we, we built a, a calculator together, on my, my laptop.

I was a bit nervous about how do you... [00:46:00] like how will he understand, like, AI on my, my laptop? Like, he doesn't... I don't, don't know if he has a really good foundation of what software even is to begin with, so it's kind of interesting viewing it through his eyes of is... Will he just assume this is how calculators are built generally, or d- will he know that there's something different about how this is coming together?

So we built the calculator. it was cool. He, he got a lot of enjoyment out of just doing five plus five equals 10. And he's like, "I, I know that. That's a, that's a, that's great." but then we started adding features, and that's where it got kind of fun and cute and interesting, and I asked him what feature he wanted to add.

And he said he wanted to add a car. So we first added a, a car button. We added it to the calculator. and then what did it do? The car drove across the screen. It was like an emoji car- And then his eyes lit up. He was really excited about this car button and what it could do. but it didn't do any math, so we were talking about, like, what could it do?

And he said, "Oh, it should multiply by [00:47:00] 100." So we actually did- held down the voice button, and he spoke... He g- he gave the feature request to Claude, and Claude added the, multiply by 100 when you press the car button. and then we added a horse button, and I think his brother came over at that point and was asking, like...

He's just learning to talk. I think he just said, "Horse." so we added a horse button. Walt- Waltie was loving, loving that. and, then the final thing, I was like, "Okay, well, we've got, got a few good features here," but Waltie's really interested in Lego at the moment. He's really obsessed with Lego. So then I gave this complex request of now take every...

Like, imagine the calculator as a, a Lego board. I want you to remove all the numbers and operators and everything, and then in order to make a calculation, you have to put the Lego pieces together on, on the board, and then it can do your calculation. So it took five minutes for Claude to do that. It was pretty complex.

But then Waltie was, just going through the process of not even using it as a [00:48:00] calculator, but he was just so fascinated with, like, oh, there are these Lego pieces on the screen, and we've just turned it from this calculator. And he was like, "I just wanna line up all the numbers so that it goes one, two, three, four, five, six, seven, eight, nine, ten."

and it was incredibly cute and heartwarming. but also at the same time, I was wondering, like, going through my mind like how, how is he conceptualizing all of this? And

Jonny: Yeah

Tom: he, he doesn't have, like, a huge amount of screen time generally, and, and so his base knowledge of how to use these tools is, I, I imagine quite minimal, but just sort of fascinating to think about someone who is growing up today.

Like, and they will, like, they will grow up, they'll be teenagers, they'll be in their... They'll be adults when this technology is much more mature. And, like, how the scene we're setting today, how that informs how they think about this stuff, I think is... It's kind of wild to think about, like, how, how AI will be at that stage.

But, but for now, it's extremely cute, [00:49:00] in a way that feels inspiring and exciting and magical. I just need to make sure that I don't create some kind of addict in him wanting to play with AI tools, constantly. But, but it's been, yeah, fascinating to, to play it with him and just to see what's going through his mind as he's sort of processing what he's seeing on the screen and how he's, how he's interacting with it.

So, yeah, a fun...

Jonny: Ex-extremely cute. I, I, yeah, you, you do sort of think, "Well, how much should we be encouraging kids to be using AI at the moment, especially young ones?" But at the same time, it's gonna be inevitable for them. Like, it's gonna be everywhere, so maybe it's better to know about it. it also made me think, are we due, like, a touchscreen but times 10, where, you know, Joe, my four-year-old, will, like, poke the telly, and not really understand why that's not a touchscreen because mo- [00:50:00] much more used to a touchscreen than a, than a non-touchscreen.

Or at least, that's what he sees sort of the direct manipulation happening, and he's like, "That makes more sense to me than having to use a controller. Like, why would I do that?" And, so you know, the next, the next thing they're gonna grow up with is direct manipulation of UI maybe, or direct manipulation of the information they get at any point in time that completely changes, so might then get frustrated by, like, having to find a button in a menu to do something rather than just shout at the thing. like what, what are the new interactions that become so normal

Tom: Yeah

Jonny: it seems crazy that you would ever have done anything else?

Tom: Yeah

So true. So true. And Waltie had very similar instincts when we were looking at things on the laptop screen. He went, first went to touch the calculator, and then he remembered he needed to touch the keyboard, and then he was struggling. It took him a really long time to move the mouse to the right place.

Jonny: Yeah

Tom: yeah. And it's, yeah, introducing this [00:51:00] technology, and this is definitely not a parenting podcast for anyone who's listening and wondering about what I'm about to say next. but I do think there's, I do think there's a difference between technology as a pacifier, like putting someone in front... a kid in front of a screen versus using it as a way to, like, engage directly with them and, and learn something together or explore something together.

And, for me, that

Jonny: Yeah

Tom: felt like a good, good use of that, this, this moment to try and, like, explore what technology makes possible in that instance. also reminded me actually of a fun, fun fact I've never, never told you, Jonny or anyone who's listening to this, but when, when I was a, when I was a kid, so my dad was a software engineer, and when I was pr- I don't know, maybe eight or nine, I remember him, like us writing a computer program together with a BBC Basic computer, and it

Jonny: Amazing

Tom: allowed you, allowed you to type in letters, and it translated them to this different code, so we could like write secret messages.

[00:52:00] and, yeah, we wrote it in code together, and it was kind of a formative experience, but also looking back now, it's quite a sort of nostalgic, bonding experience with my dad, and kind of a, kind of cool that that was possible at that time. and so now I'm wondering about the equivalent of that in the AI era.

Like, is that something... Like obviously not needing to teach kids to code anymore because of, of AI, but what's the equivalent of that? And maybe it is a version of building your own tools or building your own personalized software to help with specific homework or a specific thing you're, you're looking to tackle.

Jonny: I feel like that experience for you is about creativity and creating something new and magical, and that feeling of magical is not going anywhere, I'm sure. It just might be that the way you do it is different.

Tom: Sure. right

Jonny: On that, I think we should wrap, episode three in the can. Thanks, Tom.[00:53:00]

Tom: Thank you, Jonny

Jonny: get back to it next week, I suppose.

Tom: Yeah. Yeah. thank you for listening if you've been listening to this, and, and thank you, Jonny This has been fun. As always, excited for this week. See you soon.

Jonny: Bye​