Near Future Podcast

Tom and Jonny catch up after the summer with a roundup of August’s four guests, from Monzo’s fake app to Poem/1’s self-learning fulfilment agent. They ask whether designers are giving away their leverage to agents, answer a listener who wants the show to branch out, and share what they built, including Labyrinth rebuilt in two hours.
Key takeaways
  1. Being deliberate about what you give away is most of the job now. Every August guest drew the line somewhere different, and every one of them drew it. Jonny’s version: even when it is slower, and even when an AI would do it better, make sure the decisions being made are still yours to make.
  2. Make one decision with an agent and fifty get made for you. They arrive interlinked, so working out which ones landed and why costs more than piling on another prompt. You end up with work that is technically yours and nothing in your head connecting it.
  3. A learning loop is an agent that remembers its own mistakes. Matt Webb’s Poem/1 fulfilment agent handled queries from 700 units by turning each new exception into a standing rule. Jonny’s caveat is that the abstract version convinces nobody, which is why “everyone should have agents” lands so flat.
  4. Agents accumulate cruft, so prune them. Jonny moved his main agent from NanoClaw version one to version two this week and deleted about half the notes it had written to itself over six months. What survived reads as concise instructions instead of six months of contradictions.
  5. When everyone can make the impressive thing, impressive moves. GPT-6 Astra put one-shot 3D and cloned games within reach of anyone with a prompt. Jonny’s yardstick for where the value went: his dad, a vicar, could now get further on a demo video than Jonny could have managed five years ago.
  6. Understand the grain of the medium you are designing against. Randy Hunt’s framing was the old-school graphic designer collaborating with the printer to get intent onto the page. Having an agent write your code is a different skill from understanding what that code does to your design.
  7. AI writing feels thin because the thinking happened somewhere else. Matt Webb can feed a model everything he has ever published and still fail to stand behind what comes back. It may convince everyone else, and it is still not something you can point at and call yours.
  8. Build the silly project yourself or you will never learn the tool. Demos online tell you nothing about where the limits sit. Tom is close to downloading Blender for exactly this reason, and Jonny built a full 3D Labyrinth in two hours partly to find out what two hours now buys.
  9. The impressionists are the closest analogue to this moment. Photography arrived, exact landscapes stopped being the point, and painters who already had mastery chose what to break. Whoever does best next will understand the technology underneath well enough to bend it.
  10. Skills get better when you rewrite their output by hand. Tom burned a month of Claude Fable 5.1 credits in half an hour, then rewrote both research skills’ results manually and fed those back as the template. A 10,000-word essay he would never read became a 400-word list of links.
  11. The craft question matters more than the job title. Tim, a listener, asked whether the show could branch out beyond senior design folks. Yes, with a lens: creative technologists, engineers, product people and founders are all fair game, as long as the question stays how AI is shifting the role and craft of design.
Chapters
  • 0:04 — Intro: back to school, and a cosy September episode
  • 0:59 — Rounding up August’s four guests
  • 1:38 — James Storer: Monzo’s fake app and its biggest token spender
  • 3:29 — Randy Hunt: Notion, SVA, and two full-time roles
  • 4:36 — The printing press, and the grain of the medium
  • 6:30 — Patrick Lewis: four times a founding designer
  • 7:57 — T-Rex arms, and shaping a role around what you enjoy
  • 8:51 — Matt Webb and an inventor’s practice
  • 9:49 — Poem/1’s self-learning fulfilment loop
  • 11:31 — The Gmail triage loop Tom hasn’t built yet
  • 12:43 — OpenClaw, and agents that rewrite their own memory
  • 13:36 — Lobotomising an agent onto NanoClaw version two
  • 16:59 — Where every guest drew their own line
  • 17:22 — Are we losing our leverage to agents?
  • 18:30 — GPT-6 Astra, and what happens when everyone can do it
  • 22:20 — Make one decision, get fifty
  • 24:10 — Is design an art or a science?
  • 26:47 — Inventors, not technologists
  • 28:44 — Why you have to build the silly project yourself
  • 30:20 — The impressionists, photography, and our toddler phase
  • 31:50 — Listener feedback: branch out beyond senior design folks?
  • 33:45 — Grading the podcast, and 36 YouTube subscribers
  • 36:29 — Design at the core, but a wider cast
  • 39:40 — Labyrinth, one-shotted in two hours
  • 42:45 — Burning a month of Claude Fable credits on skills
  • 45:40 — Consolidating everything into one Dia browser
  • 47:39 — Design leadership circles
  • 49:12 — Hatch Berlin, Brighton, and save the date for 22 October
  • 53:16 — Wrap up

Mentioned in the episode

The August episodes we round up
  • Ep #12: James Storer (Monzo) — share.transistor.fm/s/ab6c1806
  • Ep #13: Randy Hunt (Notion & SVA) — share.transistor.fm/s/29c8593b
  • Ep #14: Patrick Lewis (Klay) — share.transistor.fm/s/8d030297
  • Ep #15: Matt Webb (Inanimate) — share.transistor.fm/s/c39062f5

People
  • James Storer, Senior Product Design Lead at Monzo — linkedin.com/in/james-storer-5b543625
  • Randy Hunt, Head of Design at Notion and chair of MFA Design at SVA — randyjhunt.com
  • Patrick Lewis, founding designer at Klay — patricknlewis.com
  • Matt Webb, writing at Interconnected — interconnected.org
  • Christine Røde, product designer at The Browser Company — linkedin.com/in/christinerode
  • Lenny Rachitsky, host of Lenny’s Podcast — lennysnewsletter.com/podcast
  • Tim, the listener who wrote in with the feedback

Companies and past roles
  • Monzo — monzo.com
  • Notion — notion.com
  • SVA MFA Design — design.sva.edu
  • Klay — klay.vision
  • Inanimate, Matt Webb’s current company — inanimate.tech
  • The Browser Company — thebrowser.company

Tools, models and products discussed
  • GPT-6 Astra — openai.com/index/gpt-6-astra
  • Claude Fable 5.1 — anthropic.com/claude/fable
  • OpenClaw — openclaw.ai
  • NanoClaw — nanoclaw.dev
  • Grok Bot — x.ai
  • Cursor — cursor.com
  • Dia, the browser by The Browser Company — diabrowser.com
  • Blender — blender.org
  • Poem/1, Matt Webb’s poetry clock — poem.town
  • Labyrinth, the Ravensburger board game — ravensburger.co.uk

Jonny’s own projects
  • Labyrinth online, built in two hours — labyrinth.jonny.ooda.run
  • Ooda, secure collaborative hosting for teams — ooda.run
  • designtools.fyi, run by an agent that maintains itself — designtools.fyi

Events
  • Mostly Working, the London design mornings Tom runs — lu.ma/mostly-working
  • Hatch, Berlin, 16–18 September, where Jonny is speaking on the 18th — hatchconference.com
  • design + AI, Brighton, 24–25 September, where Tom and Jonny run a workshop — designplusaisummit.com

About the pod

  • Podcast feedback: podcast@nearfuture.works
  • 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.

Tom: Hello, welcome to the Near
Future Podcast with, Tom and Jonny.

This is a solo episode or a duo
episode, I guess is the correct term.

It's September, we're catching
up after the summer holidays.

It's back to school week.

It's raining here in London.

but we're back to share a roundup of
what we've been up to, who we've been

speaking with, and some other overall
themes, that are happening right now.

So, Jonny, any quick thoughts following
your, your back to school week?

How's it going?

Jonny: So after a sweltering hot
summer, I'm looking out the window,

it's raining, feeling a certain
amount of hygge, and, feeling cozy.

Let's have a cozy episode, Tom.

That's what I think.

Tom: Yeah.

I'm excited to see what the,
automated transcript does with

that, pronunciation there.

We'll see.

Jonny: I'm sure it'll get it right.

Yeah.

Tom: You, you, you nailed it, I think.

Jonny: Thank you.

Thank you

Tom: yeah, I, I agree with that.

So, yeah, so I figured we could start off
by doing a little bit of a roundup, of

some of the guests we spoke to throughout
August, 'cause we-- there were four of

them, I think, and then both of us were
on holiday at different points and,

sort of dropped off the social media
landscape, so we didn't really talk

about any of these posts going live.

so if you don't subscribe to this
podcast, you might have missed them.

but yeah, I thought we could just talk
a- talk about, some any-- some high-level

takeaways from, from them first.

How about, our first episode at the early
part of August was James Storer of Monzo.

any takeaways from that, Jonny?

Jonny: Feels like a decade ago, being
the other side of summer holidays.

And yeah, we're, we're really showing
our, age by being terrible at clipping and

getting our clip- our interesting tidbits
out into the world and talking about them.

So, please everyone do go
and look at back issues.

some really, good stuff in there.

yeah, James.

Oh man, I've, I've known
James for a long time.

It was a really fun chat for me.

think in terms of what he's been doing,
he is-- he's always designed fast.

He's always had, had 100 ideas an hour.

we're quite similar in that way, and he's
always, to just create and create and

create and create, since way before AI.

And hearing him now at Monzo just
being unleashed, he is the biggest

token spender in the company.

and him talking about their of fake
version of the Monzo app that they use

for prototyping and for getting confident
about things that they're gonna ship.

You know, the, the design process
or the discovery process resulting

in higher level of, of confidence
when you go and put something out,

by building this extraordinarily
deep, prototype, basically.

Including data, including different
state, that they can, manipulate.

Anyone can manipulate and
build anything they want.

It's a really good example, I think,
of a design team on a really ambitious

project around vibe coding, which
is basically what this is, and using

it for real business value rather
than just fun stuff at the side.

So yeah, it was really great to hear
about that alongside his silly experiments

at home, forcing the rest of his family
to use his harebrained apps and taking

over the telly and stuff like that.

So, yeah, a really fun episode.

next up we had Randy, Notion legend, SVA.

Tom, what were your thoughts?

Tom: Yeah, I really enjoyed the chat
with, with Randy and how he balances

everything, how he manages to fit in
what appears to be a full-time role

as head of design at Notion, alongside
a full-time role of chair of the MFA

in design program at SVA in New York.

That's a lot of tasks on his plate.

but it seemed like they were very
complementary, like the, the practical

aspects of understanding the intricacies
of working on building tooling around

agentic workflows in Notion directly
contributed to how he set up the

operational aspects of SVA and having
SVA become a, a Notion customer.

I think the, the one thought I've been
left thinking about is around … He,

he gave this reference to thinking like
a, a, an old school graphic designer

going to a printing press and speaking
to the printer around, or the, the

person who is in charge of the printer,
around how to get the best translation

of a design from a graphic designer's
intent into the actual output of the

work, and really involving them early
in the process and collaborating with

the printer to get that great execution.

And he made a reference to this as it
re- relates to an engineering perspective

for digital product design today.

And I really like that, that framing
because I think there's a lot of

talk at the moment around how AI is
turning everyone into a design engineer

because every designer is writing code.

But there is quite a big difference
between having an AI agent write that code

versus understanding the, the constraint
of the medium and the grain of the

medium on which you're designing against.

And so yeah, I like that example
from Randy, but I think it's also

just a, a theme for us to, to take
into how we're pushing back against

the constraints of engineering with
respect and understanding for the

medium, rather than trying to overcome
it through AI as this magical tool.

Jonny: Yeah.

Yeah, I just also love nerding
out about Notion with, with him.

Tom: Yeah.

So true

Jonny: yeah, as, as Notion fans,
to be able to talk Notion and hear

how they their agents inside Notion
in the open and, yeah, really cool.

to, to-- Yeah, that-- It's like an
iconic-- Especially as designers,

I feel like every designer signs
up for Notion personally, even

if they don't use it at work.

So have someone talking about how they use
it internally is, is really fascinating.

up, we had Patrick, from Klay, who
is a four times founding designer.

so, you know, glutton for punishment.

you know, if, if you wanna join a company
as, as the first designer, then you

gotta, you gotta know that you're gonna
be subject to the full force of, of

founders being like, "It doesn't feel
right," basically over and over again

for however many years you're there.

and, variants of that, you know,
get-getting into something, building

something beautiful, and then it hitting
the cutting room floor immediately.

there's so many challenges building
with, really no walls to lean against.

So, know, you could-- I, I feel
like you could feel the scar

tissue coming out through the, the
microphone, as Patrick was talking.

but yeah, how he, how he thinks about his
work, I think how he conscious of what he

keeps to himself versus what he's willing
to give away, and seeing the role of

the human in the design process, I think
it was j- really fascinating to hear.

and, yeah, Tom, Tom, what did you
think of, of the Patrick conversation?

Tom: I thought it was great.

I thought it was really interesting.

And you mentioned the framing there of a
glutton for punishment, and he gave this,

story about, throwing discus and doing
hammer throw in school, and his body not

necessarily being the right optimal shape
for being naturally gifted in that area.

he described himself as having
T-Rex arms and throwing, throwing

the discus, which, funny imagery.

but I think there's something about
the, the persistence and discipline and,

even if you're not naturally well-suited
to something, like knowing what you're

passionate about, knowing where you
get energy, knowing what you really

enjoy doing and leaning into that, I
think the role of a founding designer

can really be shaped around that if
you're-- if there's enough alignment

between what that is for you and what
a company needs or a founder needs.

so I think that really, really
resonated And then the last guest,

the most recent guest, which we, the
episode went out last week, was Matt

Webb, of Inanimate, or currently of
Inanimate, but of various, exciting

legendary design practices in London.

And there were so many good nuggets
in that episode, everything from, how

he built a book vending machine that
was toured around London and what he

learned about refrigeration or lack of
refrigeration in having labels stick on

each book, through to agentic workflows
with AI and various other things around

how he's reignited his passion for hi-
the, the practice, the sort of invention

practice he goes through as a, as a
designer, which I thought was really

interesting to hear his perspective on.

but the, the one thing that I still think
about today from that conversation is

the way he used agentic loops to create
a self-learning system for fulfillment,

of his Kickstarter art project, Poem/1.

And if you haven't yet listened to
the episode, he describes, sending out

these 700 units, and then customers
would write in with, questions or things

that weren't quite working as expected.

And he essentially built this agent
that, or built a set of standards

and princip- or standard operating
procedures to resolve each of

these questions as they came in.

And so when a new one would come
in that hadn't been read before, he

would manually respond to it himself.

But then over time, this body of, this
source of truth of how these decisions

are made would be established, and
it would keep learning and growing

so that over time, the agent would
be able to respond to customers and

resolve their queries without him
needing to be in the loop at all.

And that's something that I've been
really fascinated with as an area

generally, and finding the opportunities
for where are loops appropriate in

these self-healing, use cases that can
have human input to train and improve

and refine them, but Ultimately there,
the human input is about improving

the output of that system to create
greater value or speed or efficiency

or whatever it is for the customers
who are on the receiving end of that.

And I've been thinking about building,
I haven't yet built a email triage

learning, self-learning loop, built on
top of Gmail, but it's been on my to-do

list to do, and maybe I'll get to it this
week, based on those same principles of

how can I create this nice UI that feels
like it's productive for me, but it's

actually about teaching an agent how
to do these things and, and learn more

about my tone of voice and how I make
decisions so that the small stuff can

be taken over or outsourced a bit more,
giving me more time and head space to,

to work on the more meaningful things.

Anyway,

Jonny: Mm-hmm.

Tom: I'm sure you have some thoughts
about Matt's, Matt's conversation as well.

It was a really good one.

any takeaways for you?

Jonny: Yeah.

I mean, just, just on, on that,
Yeah, I'm, I'm a big Matt fanboy.

If-- You'll know if you listen
to my intro to the Matt episode,

which went on for about 15 minutes.

yeah, on, on the, the, the learning
loops, yes, he-- I think he gave a

really practical example of how an agent,
I think is basically what a learning

loop, a sort of self-learning and
self-healing loop is, is, is, is an agent.

You know, if, if you think of, something
like OpenClaw back in February,

all of that stuff was coming out.

really at the heart of what they were
doing, what that technology was doing

was it would take some input, it would
perform an action, and it would then

do it repeatedly, and each time it did
it, it would get a little bet- better

because it would update its markdown
files, its soul or, and its memory to,

support, a better execution next time.

that's sort of at the heart of this
technology, but it's so hard to-- I

found it very difficult to describe
in the abstract unless there's a

very clear use case to do it, and
I think Matt's was really good.

this week I lobotomized my main
agent that I run, a- and sort of

rebooted it on a slightly more
modern version of the technology

it's built on, which is NanoClaw.

It was on NanoClaw version one, and I
decided to put it on NanoClaw version two.

And,

Tom: in

Jonny: doing so, I was working away
with, with Claude, who was SSH'd into

my VPS, for the people that know what--
love their three-letter acronyms.

and we were looking through
the guts of what it had built.

And, you know, when I first set this
thing up in March, and, all I've

really been doing is talking to it.

And we looked through this week all of
the different files it had created, all

the routines, all of the, notes it had
basically made to itself, like, "Don't do

this this way because that didn't work.

Do it that way."

And, there was a lot of cruft, so
we really managed to slim it down.

You know, deleted half of that stuff
to just make it concise instructions

rather than of conflicting things
that were built up over time.

So, you know, you have to prune these
things, but, it was just fascinating

to see built up institutional
knowledge over six months that had

just been happening of, some of it
through me, some of it through itself.

you know, it tries an API endpoint
and is like, "Oh, that's changed.

I'm gonna go and look, and now
I'm gonna update myself so that I

don't make that mistake next time."

and, and that one is, is,
an agent that mainly runs a

website for me, designtools.fyi.

and, I decided after a summer
of ignoring it that I was going

to pay it some attention again.

So now, he's happy and
up and running and, and,

Tom: continuing

Jonny: with, with his work.

so but, yeah, always l- looking for r-
real examples of this kind of thing in

the wild because in the abstract everyone
should have agents isn't very useful.

the other thing I got from Matt's episode,
I think, that has really stuck with me is

his-- is how he-- We were talking about,
writing with AI and our experimentation

with doing that, description of why it
doesn't work, I think is just the best

that I've heard, which is you can give
it all of your writing, and Matt's got a

lot on- online, so, the models have been
trained on his writing already, I'm sure.

But-- And you can tell it to write
something as you, but as you read

it, it doesn't mean anything.

It feels thin.

It feels like you can't it.

You know, you couldn't
an argument around it.

It's not your thinking.

It's thinking that has been
done by something else.

So even if for everyone else it looks
convincing, for Matt, and I think why

this resonated is I feel the same way.

I just don't feel like
I can stand behind it.

Like I can't point at
it and say it was mine.

so yeah, just a g- quite a sort of
nuanced take on, on creating stuff

with AI, that has stuck with me.

And I feel like there's, there was
a bit of a theme throughout everyone

actually this fall, but probably every
interview we've done, which is what

do I keep and what do I let go of?

What's okay for me to give away, and
what's important to me and, and part of my

craft and part of-- part of what I enjoy,
but also what is my value, my ultimate

value in this, that I want to, I want to
or believe that I should keep hold of?

yeah, so as a, as a theme that sort of

Something that we've talked about
before in various contexts, I think too

Tom: Yeah.

Yeah, so true.

And I feel like every guest had a
different flavor of that answer, or

there was a different, a different
line, or they drew a different

line on what was the thing that was
required or they need to keep hold of.

And I think there was, But
yeah, definitely a, a question

of how blurry is that line?

How much does it vary based on your
environment versus the sort of innate

human quality of m- making decisions
and feeling ownership over those

decisions, and perhaps that extends to
feeling ownership of the, the clarity

of thought produced through writing.

Yeah.

Yeah, very interesting.

there is actually a related topic to
that which we're gonna g- jump to a bit

later, but are we losing our leverage
by giving our design work to agents?

And it feels very much in this vein.

Curious if you have any thoughts
on that as a topic, Jonny.

Jonny: it was a good tee up because
you know I have thoughts on this.

but what I think, the-- Tom, you
organized this Mostly Working,

and there is a WhatsApp group which
is fairly active that people that

have come to the meetup can, can join.

And people should, by the way.

If you've, if you've been to a Mostly
Working event and you want to join the

conversation outside the room, then, you
should definitely check that out, and

maybe we can put that in the show notes.

But, a conversation started around
how cool, this new-- these new models

are that, you know, GPT-6 Astra, which
is OpenAI's new model that's just

come out, and the internet is abuzz
with how amazing it is at 3D, and,

people are one-shotting games and,
or cloning games that already exist.

And, it started a conversation around how
magical it is to be able to just create

this stuff and how much it would've
cost to have created it in the past

versus how quickly you can do it now.

And I think the conversation was
interesting in the sense that

on one level it is amazing, and on
another level, as soon as everyone can

do something, is it amazing anymore?

I think that w-where sort of I was
arguing, that you look at these

incredible tools that are, that are
being created, these amazing visuals,

whatever it is, you know, whether it's
interesting shaders or 3D effects or,

motion, all of this kind of stuff,
feels magical because five years

ago none of us could do this very
well or definitely very quickly.

And now suddenly we can all do it, and
the, I don't know if it's an Overton

window or equivalent, but it sort
of shifted towards this being kind

of normal or less impressive, and so
therefore what is the new impressive?

You know, what is, what is now a designer
can deliver that no one else can?

Now everyone can deliver
extraordinary technical ability,

and code is cheap to free.

what, what is our value anymore?

And where do we have leverage?

You know, let's assume that the technology
continues to get more powerful and easier.

So the trend for really anything you
can create technically is, is to zero.

You know, the cost of it is to zero.

You need-- The, the amount of ability
you need to create it trends to zero.

my dad, who is a vicar, I keep picking on
him, I think, on this podcast, but, you

know, I could, I could text him now and
say, "Go on this website create a video,

a demo video for product," he probably,
using natural language, would be able

to get further than I could have got
five years ago, technically, you know?

so what do we keep?

What, what is the magical thing?

People talk about taste.

is it taste?

Is it constraint?

Is it just the buildup of,
both domain expertise and,

knowledge of the goals of the project
that we're working on or the, the,

the task that we have at hand that is
unique to us and can't be replicated?

know.

but one thing that I feel quite
strongly is, and I think has been

reflected in these, these conversations,
is being deliberate about what

you give away and what you keep.

and even if it's slower and even if
it feels like you're worse at it than

an AI could be, sure that decisions
that are made, that you're making are,

are yours to make still.

thing that I find, as I design
with AI is I will want to make one

decision and then 50 will be made,

Tom: Mm-hmm

Jonny: and that they're all interlinked.

You know, t- too much stuff has happened.

The prompt, the-- it ran away with it.

Too much was added or too many different
things were changed, and at that point,

it's far harder to unpick it and work
out what worked well and what didn't

and why decisions were made than it
is to just crack on and add another

prompt and pile, pile more stuff on top.

I think that's true in design,
true in really any, any vertical.

And that then means there's this buildup,
a bit like Matt was talking about.

There's this buildup of work that has been
done that is technically done by you, but

that you have no synapses connecting in
your brain as to why it was done really.

You can sort of paint a broad
brushstroke, but individually,

you know, w- we used to perfect
every pixel, and that was too slow.

And I think, you know, we're, we're
not going back there, but you at least

knew why you were doing everything.

Tom: Mm-hmm.

Jonny: there's this sort of balance.

and I, I think I don't-- I haven't
been convinced by any of the

technology, even GPT-6 Astra, that

the models without any human
support can beat humans at

soon.

And so I feel confident that if we keep
some of that leverage to ourselves, then,

we will survive all of this and maintain
our value and all of that kind of stuff.

But, you know, I'm

Tom: Yeah

Jonny: talking my book, but

Tom: Well, I think that raises a
few interesting questions of the,

is design an art or a science?

And how

predictable is the outcome and how
testable is the outcome versus how

much is it shaped by opinion and
individual voice or the voice of a team.

And, I remember from various teams
I've-- design teams I've been a, a

part of in the past, there has been
this consistent theme of however

large the team gets, the experience
a customer or user should experience

the product as if it was created from
a, a single individual or a single

hand that is crafting that experience.

And I think that speaks to design
systems and consistency and making

the same types of UI choices across
and throughout an experience.

But I do think that raises, yeah,
there are, there are aspects of that

that I could imagine being-- once
the direction and standards are set,

I could imagine AI being a really
valuable tool in executing against that.

But there is this, human perspective,
this op- strong opinion, this

originality that I think is really
a core part of the equation of a

great product experience that I don't
think is possible to achieve through,

purely through prompting, at least not
right now in the state of the models.

And, and to me, that's
less about technical use.

I think we're on this interesting
wave of these micro trends that, keep

appearing every time a new model drops
and, and yeah, amazingly impressive

for a month, and then it feels like,
oh, that sort of style is everywhere.

And I don't think we should steer away
from that, but I think there's probably

an even greater need for originality and
human creativity that should be tapped

into and applied to all the creative
possibilities that are now out there.

But it's really hard to create distance
from the trend versus what is the

originality that is maybe tapping into the
technology that's now available but is--

sort of sits adjacent to the trend piece.

I don't know.

Jonny: Yeah, I think what the, what
the, the trends show is a, a sort of

technology first approach where we
can now do this stuff and everyone's

doing it, so now I'm gonna do it too.

And a problem centric approach is
what designers should be focused on.

So I will invent-- I like the idea
of us being inventors, you know?

Or I will invent based on the
problem that I'm trying to solve or

the context that I operate in, and
maybe I should pull in some of these

extraordinary technologies or, I should
leverage AI as part of the production

process to get to the final result.

But the-- I am the inventor.

I'm Caractacus Potts, is Ch-Chitty
Chitty Bang Bang reference for

Tom: It's cool.

Jonny: as a deep cut.

w- I, I am the inventor and, only I
really know exactly why something needs

to exist and exactly how it should feel,
and I can't even really describe it in

a way that I could write in a prompt.

And so maybe

Yeah, that's the art, right?

That's the art, and then the science might
be we've-- we have a robust design system

now in component lib- components in, in
code, and we have our brand guidelines

and our colors and et cetera, et cetera,
and all of the things that it's hard

because it's zero to one stuff to get to
the point where that's-- happy with it.

But then once we're happy with it, that
production process moves a lot faster.

that maybe feels like a good balance.

But

Tom: Ja

Jonny: I mean, as always, I would say I
reserve my right to change my opinion.

I think I've changed my opinion multiple
times while we've been recording this

podcast over the last three months odd.

So,

Tom: Yeah.

I, I also do want to give a, a shout-out
for pushing the technology, and

designers experimenting, and, that might
not be the right long-term solution

for a everyday product application.

But getting your hands dirty,
experimenting, trying these new models

out as soon as they're released, trying
all of these fancy, shiny demos, I think

is really important to understand where
the limits are and what's possible here.

So I would embrace anyone and
encourage anyone to try, try

these things, myself included.

I'm close to downloading Blender
and, and trying, trying myself.

because I think without that,
it's really hard to make…

And I think that's something we've come
back to time and time again, throughout

this podcast and through our, our work
together, has been you can't really

understand what these models are capable
of from just seeing demos online.

You need to, like, build something
yourself, and maybe there's a project

on your to-do list that might lend
itself well to playing with this stuff.

that doesn't necessarily mean that
doesn't change who you are or how you

approach design problems generally.

It's more just having a deeper
understanding of the technical landscape

that you could be working with and
having a, a broader, deeper set of

tools to draw upon in your everyday
design work, which I think is a really

useful thing to, to have access to.

as long as it doesn't necessarily change
you into only chasing the latest trends.

It's more of a s- complement to
the core foundations you have

as a, as a strong designer.

Jonny: Yeah.

Goes back to what you were saying
about Randy, I think, or, or that

Randy was talking about, which is
understanding the grain of the paper,

understanding the printer works.

you have to understand th-this
technology, and the only way to

really understand it is to use it.

So building silly projects is
essential, I would say, just to sort

of get a feel and a bit of confidence.

And then, you know, all the impressionists
came out of photography turning

up and suddenly, you know, exact
landscapes weren't needed anymore.

And so the impressionists used their
ability with paint and their, what they

already knew and their mastery of, of
that technique, and then chose to break

it and chose to invent within that.

It feels like we're in that sort
of moment again, where the best

people at the next phase will really
understand the technology underpinning

it and, we are, and be able to sort
of master that, but also find ways to

adapt it a-and use it appropriately.

we're in our toddler phase, I think,
with working out how to use, use

this, AI in a way that is meaningful
and actually matters rather than

just slap, slap some more motion or
some more, interesting effects on it.

Tom: Yeah, absolutely.

I th- I think this is, actually
potentially an interesting segue

to, to, if you're happy to move
on, to some feedback we've gotten.

Jonny: I was gonna suggest the same thing.

Tom: Okay.

Okay, perfect.

Well, we have, we, we are open
to feedback on the podcast,

podcast at nearfuture.works,

and we have received some f- some
feedback, which has been, really great,

and I think it asks a, a specific question
that I think would be a good one for us

to, us to actually talk about together.

it says various complimentary things,
and then it says, "Can you branch

out beyond the senior design folks?"

And I'm curious, Jonny, what do you think?

Should we branch out beyond
sen- senior design folks?

Jonny: Mm-hmm.

Yeah.

Thanks, Tim, longtime listener, first
time caller for, giving us this feedback.

And please everyone else have your say.

This is an opportunity as we have
this conversation, and react to it,

you can also tell us what you think
and it will help us to decide which

guests to get in, et cetera, et cetera.

yeah, what's this podcast about, Tom?

We,

we started it and, and you know,
we've called it sort of design and

AI, a sort of overarching theme, and
it's sort of by talking to people and

working out what we're most excited
about and where the combination of

our interests and, abilities probably,
combines with our network and, you

know, who we get excited about bringing
on into the, what the podcast is.

But that feels-- I suppose that could
feel a little bit like haphazard and

or

Tom: Yeah

Jonny: it, it does feel like a
nice moment to just reflect on

have, have the topics we've covered
so far felt directionally right?

Like is there anything
else we can jump into?

and personally I would give
us a 8 out of 10 so far

Tom: Nice.

On, on what, what scale is this?

Jonny: seven and a half.

Seven and a half.

Tom: What is the scale?

What is the scale of,

Jonny: Scale.

Tom: of, quality?

Jonny: putting me on the spot here.

think a mixture of, I mean, well, I
will duck, maybe I should go lower

because I'm gonna dock myself points
for both mic quality at various points

and also a range of extraordinary
backdrops in strange, strange places

and, camera angles looking at my nose.

So, know, that's one issue

Tom: a feature.

That's a feature, not a bug

Jonny: No, maybe it is, yeah.

and for our video watchers, video
audience, 36 YouTube subscribers.

I think that is about

Tom: Whoop, whoop

Jonny: yep.

guys, stop subscribing.

It's too many.

anyway, yes.

well, I think, I, I mean this is-- And
maybe we've been just been allows our--

allowed ourselves to be selfish, about who
we, what we talk about and who we get on

and hope that we're interested in is what
enough other people are interested in.

But I suppose I would give, to answer
your question, what is the scale?

The scale is

conversations that have
personally interested me.

All killer, no filler in terms of content
that I would personally want to listen to.

but, but I think opportunity-- I
think there is an opportunity to

branch mildly out of design because--
or designers, I should say, because

design is not just done by designers,
and increasingly this is true.

And so, you know, creative technologists
and engineers and product folks and,

founders, we've got a guest coming
up who is building a venture-backed

design tool, as a founder.

I think, you know, they all add to
the conversation and add a perspective

that I think will be valuable.

But personally, I think we
should have design at the core,

because that keeps it
clear what we're about.

What do you think?

Tom: Yeah.

I, I agree.

I think, I think if it-- the only
guests we had on had design in their

job title or were head of design
at X, Y, Z, I think it could get a

bit repetitive, a bit, a bit samey.

But I do…

I think the core reason I am most enjoying
doing this is the exploration of what

design, is evolving into as a craft.

And as you say, a lot of what might
be considered design output isn't

necessarily created by designers.

but I think designers are obviously
probably the, the majority of the

guests we'd have on this podcast.

But w- if, if there is someone who
brings an interesting perspective on

what that, design-- what the shape of
design is becoming and what that craft

is evolving into, then I think that
would be fair game from my perspective.

And yeah.

And I think that could be the sort of
practical executional level of, like

as you say, engineering and design
engineering or brand engineering even.

but then there's the sort of macro
view of what shifts are happening in

the overall landscape and what is the
role of a designer evolving into and

the role design plays within companies.

So yeah, I would strongly agree with that.

So, so Tim, thank you so
much for your feedback.

Y-yes, I think we answered your
question that, we, we, we may branch

out a bit beyond people with design
in their job title, but the lens

will still be how does AI shift the
role and evolving craft of design?

Jonny: Yeah.

I think-- I don't think
we're gonna turn into…

I mean, o-one example just to sort of
round this off is Lenny's Podcast, which

is overtly like a product management.

He was a PM, product.

but he gets designers on, he
gets engineers on, he gets

founders on, he gets all sorts.

And it's more-- It feels like it's
more content for product managers or,

senior product people that-- to enjoy.

And of course, senior product
people don't just want to

hear about product management.

My God, that sounds deathly dull.

but actually, they're m- the product
managers and product folks love thinking

about strategy more broadly around, you
know, how, how companies are built and,

macro trends and all that kind of stuff.

I would say we do sometimes talk about
macro trends, but, mm, I suspect we--

it'll be a little bit more tactical
than that, would be my, my gut.

But just like what would
designers be interested in?

And that will probably capture
people adjacent as well.

so

Tom: Yeah.

So if you're listening to this,

Jonny: you think.

Yeah

Tom: yeah, if you have an opinion,
we'd love to hear from you,

Jonny: Yeah

Tom: podcast@nearfuture.works.

And now I think we're coming to
the end of this conversation.

let's check in.

What have we built this month?

Jonny, you've built a few things.

Tell us about them

Jonny: Yeah, I, my, my proudest
achievement, I think, over

the last month was built very
quickly in, like, two hours.

When I was a kid, we used to play this
board game Labyrinth a lot, it's, game

where you have tiles on a board, and
you basically have to go and collect all

your treasures and then go back to your
home tile faster than everyone else.

And you push the tiles along the board
so it changes shape of the labyrinth,

and the maze changes, and you can block
other people, or you can open up new

corridors for yourself to get through.

Anyway, childhood game.

This summer went camping with,

There were about 35 of us, I
think, They're all-- Huge family

camping trip, with a dozen odd
children under the age of 10.

And I brought along our, our
Labyrinth game second year in a

row, and once again, huge hit.

the next generation are getting into it.

It is timeless.

It's a, it's a complete classic,
and anything that gets people

away from screens and, and whatnot
and, is, is good in my book.

Ironically, what I built was on a screen,
but, I thought, "Wouldn't it be cool

if you could play Labyrinth online?"

And there's no way to do that.

So, I sort of, as they say,
one-shotted, Labyrinth online.

and it work- It, it built a full sort of
3D board that you could navigate around.

and a couple more prompts got
me, you know, online multiplayer.

So Tom and I, we could have
a, have a game remotely.

it's got leaderboards for solo mode,
so you can pick how many gems you

wanna-- how many treasures you wanna
find, and then see if you can beat the

high score in terms of number of turns.

Anyway, it's just a complete, jolly,

com- you know, n-no value
really beyond my own enjoyment.

played a few test games with some people,
and it worked really nicely, very smooth.

And so, yeah.

I'll, I'll-- It's
labyrinth.jonny.ooda.run,

but we can drop it in the,
in the show notes as well.

If you play a game of Labyrinth,
please let me know because, you know,

the leaderboard is full of my name
at the moment, and I need some other

people to go and, go and beat me.

and, that's-- I mean, I, been,
been working hard on, on production

stuff for, for Ooda and getting
that ready for a launch actually.

So, that's the other thing I've been
working on, including attempting

to create a demo video entirely
with AI which is comically bad.

and you know, the-- It turns out
all of those guys one-shotting

demo videos are full of shit.

Or at least I'm doing it wrong.

Maybe it's a skill issue.

but yeah, that's me.

How about you?

Tom: well, well, I didn't think
I'd really built, built much

this month because of the summer
holidays, being away, various things.

but I did manage to burn
through m-my Claude Fable 5.1

credits in about half an hour yesterday
morning, building, building some skills.

Very inefficient use for Fable, and then
I had to pay for some extra credits just

to get some basic stuff done after that.

but I think I've probably learnt
a valuable lesson about building

skills, which is before I was just,
like, almost, like, one-shotting.

Like, "Oh, I have this
idea for this thing.

I need to complete this task.

Why don't I just do it as a skill?"

quickly realizing or generally realizing
I sort of never really got great results

from that approach, and so now I'm
trying to double down a bit on if I have

a, an idea for a skill, I will try and
try my best to not just one-shot it but

actually get the, get the resulting…

And these two were sort of
research-oriented skills, so I'd get

the result back, and then I'd basically
completely rewrite the result by hand

based on what the AI produced, and then I
could feed that back into the, the model.

So it's sort of a form of eval, I
guess, but just, but using the actual

content that was found through this
research from the AI and manually

getting it, massaging it into the
right format and right template.

the key value for me is it's taken the
result that comes back from a skill from

this 10,000-word essay that I will never
read that repeats itself 20 times, into

a 400-word bullet point list of links.

It feels much closer to a Wikipedia
entry, a sort of concise Wikipedia

entry, rather than a sprawling mess
of words that probably make a lot

of sense but I don't have time to
read or the interest in reading.

and also there's just such a
strong bias from the sources that

information is out there around.

So, one of these research skills I
was making was around understanding A

company's health in terms of financial
team, team growth, various aspects

like this, but all the information
around that is released in PR

statements or press releases, from VC
companies, mostly in investment firms.

So this report was coming
back with like heavily talking

about all these VC companies.

I'm like, "Not interested.

I don't care."

Like, I don't care about that stuff.

or at least for the purposes of the
skill, I don't care about that stuff.

so sort of interesting taking this sort of
heavily biased data set and then trying to

correct it into something that is actually
the use that was more relevant for me.

So anyway, that was the thing I built.

also coming back from holiday,
I decided to finally give up…

Jonny, you and I have been talking for a
while now about having so many different

domain names, different Gmail accounts,
different browser identities that

are attached to each one of them, and
I've basically consolidated everything

into a single Dia browser instance.

We had a Mostly Working event in the
middle of August with Christine Røde from

The Browser Company talking about the
Daily Brief, which is a feature in Dia.

It's like the most loved feature
there, which kind of gives you

a summary each morning of what's
happening across your day.

I was hooked, and it made me like
decide to, okay, I'm gonna solve

this identity problem and just do
everything in one browser instance.

And then that has meant I've, I've been
redirecting my email into a single inbox.

I've been like having one-
a single calendar view set

up so I can view everything.

Yeah, so it's, it's probably not the
most efficient long-term solution,

but my brain is significantly calmer
by just having a single place where

I can see everything in one place.

Jonny: forwarding emails from different
places into one central inbox?

Tom: Yes

Jonny: that I would, thinking
about it, that's the bit I'd be

like, "How would I even do that?"

But

Tom: It's actually very easy to do.

I mean, there's, there's probably a
separate, privacy and security question

to resolve before doing it, but, but
technically it's quite easy to do.

Jonny: Right.

Tom: If you use Gmail, at least,
you can, you can set it up to

automatically forward, and then
you can set up your email to reply

as that email address as well.

So in theory

Jonny: 200 pounds to Cursor, and
then you get a Grok Bot, and bot

can have multiple Gmail accounts.

That's the big thing I've been reading.

Maybe, maybe next month we'll
have tried out Grok Bot.

Tom: Yeah, maybe, or maybe Claude will
add that capability or, or I don't know,

Jonny: come on Claude.

Tom: ChatGPT.

Jonny: Ja.

Tom: w- w- one other quick, quick
shout-out I wanted to give out when I

said I hadn't really built anything with
AI, I did launch a design leadership

circles, which I've been-- I started
running the first one last week.

this is completely unrelated to AI.

It's just running a series of, peer
group community of practice sessions

with design leaders who are, navigating
this current moment, some with big

challenges they're facing, others
just curious to meet and connect with

other design leaders in the space.

But it's been really insightful to
understand, like, what the temperature

is across design leadership right now.

And although it's not focused on
AI, AI is the backdrop for a lot

of these conversations and the
change that that's encouraging.

So I've been really enjoying those
conversations, and that's sort of

quickly come about, and I'm running
those sessions throughout September,

and I'm hoping to do some more exciting
things with that, in October, which I'll

talk about in our next episode as well.

Jonny: Awesome.

Yeah, they sound mega.

get-- I, I've got FOMO, you know.

I know what, what's being said.

What's, what, what are we coming--
What are we-- What are the conclusions?

Maybe you can share.

Maybe you probably can't share.

It's all,

Tom: I, I,

Jonny: coming from that?

Tom: I obviously can't share any
names or companies, but I'd say

I've, I've not prepared any themes
to, to play back at this time.

Maybe I won't answer that.

Yeah, next month.

Jonny: it next month.

Yeah,

Tom: Yeah, yeah, definitely

Jonny: Right.

Let's wrap up.

So, we got coming up, we've got, we've got
a few, few things happening, haven't we?

Tom: you're, you're heading to Berlin

Jonny: I'm going to Berlin on, what is it?

A couple of weeks' time.

On the 18th, I'm speaking on main stage
at Hatch, so if anyone listening to

this is gonna be there, please do, yeah,
drop me an email or, or like DM me on

LinkedIn or something "Hey, I'm gonna
be there," and then we can hang out.

That would be nice.

yeah,

Tom: I'm excited for your talk

Jonny: my talk.

I s- I feel like it's been like nailing
jelly to the wall trying to get this talk

clear because the technology moves so
fast, and I feel like I've got this

enormous wealth of things I could say.

And how do you say one thing concisely
in 20 minutes when your brain is so

full of all of the things you could say?

But luckily, that's why we have
a podcast that I can just blah,

blah, blah, just get it all out.

so but yes, final prep for that.

And then the week after, we are running
a workshop, which is actually sold

out, I believe, at, in Brighton at
the Design and AI Summit on the 24th.

really excited about that one as well,
and planning our travel and whatnot,

so that's starting to feel real.

and then in October, is shrouded in
mystery, but all we would say is save the

date on the 22nd of October, especially if
you're a Londoner or adjacent, because we

will be running an event, a bigger thing.

and we're w- it doesn't have a
name yet, and it doesn't have,

yet for us to fully announce it right now,
but we do know it's gonna be on the 22nd.

So, Tom, is there anything else
that we can say about that?

Tom: I think you have sufficiently
disclosed the details we can say.

and when Jonny says we haven't
got a name, there are a small

shortlist of possible names.

It's not like we're starting
completely from scratch.

Like, we know generally
what this is going to be.

And if you're a designer or a
design leader who is-- Well, if

you've made it this par- far in the
podcast, then you, you're, you're the

kind of person who should come to this.

but it is an in-person event.

It will be lots of fun.

It will be potentially the, the
biggest thing we've done to date

in terms of people in a room.

Jonny: Ja

Tom: yeah, keep your eyes peeled.

We will reveal more in
an episode coming soon.

But yeah, very excited for, for this
to come together and, it's gonna be fun

Jonny: Yep.

Yep, yep, yep.

And more big announcements about the Near
Future podcast, I think coming soon too.

That one's even more mysterious.

Tom: Yeah.

Jonny: so unsatisfying
this, this news bulletin.

We can't really say
anything about anything.

But, yeah, that I, I think, watch
this space in the next month, I

think that there's gonna be some
evolution and things happening, for us.

but the podcast's not going anywhere.

Really excited to get to…

What's the milestone?

Maybe, I think 52 episodes c- if
we do it every week, 52 episodes,

that means we've done a year.

What are we up to?

Like

Tom: This is probably 16 or 17.

Jonny: Yeah.

Yeah,

Tom: And we, and we have a few,
a few guests booked in who we're

speaking to soon, so yeah, I think
we're on track for one a week for

the rest of the year at least.

Jonny: Yeah, and we will endeavor to
tell people about episodes in the future.

We have not done a good job over summer of
telling people episodes exist, so please

do go back through the back catalog.

there's really good stuff in there.

Right, that's it.

Tom: Let's wrap up.

Jonny: shall we skedaddle?

Tom: Yes.

Thank you for listening
and making it this far.

Jonny, great to chat.

Let's wrap up

Jonny: You too Tom.