Ep#5: Are middle managers actually finished? Plus: personal software, Ideogram & Midjourney goes medical

Tom and Jonny are back for another wide-ranging episode.

Jonny starts by recapping his golf trip to Mallorca — where a non-technical friend used Lovable to vibe-code a fully functional scoring app for their annual 16-man Ryder Cup-style tournament, the Stryder Cup. It's a perfect case study in personal software.

From there, Tom brings a meaty topic: the "death of the middle manager" narrative gaining steam in the press. Drawing on his five years as Director of Product Design at Monzo, Tom and Jonny debate whether this is a real structural shift or a post-ZIRP overcorrection — and what managers should actually do to stay relevant.

In the quick-fire round, Jonny introduces Ideogram, a small but opinionated image model that excels at visual design and can be self-hosted and fine-tuned on your own work. Tom follows up with Midjourney's unexpected pivot into health and wellness medical devices — and what it says about where AI companies might be headed.

They close by testing themselves on "In the Weights" — a site that checks if your name appears in AI training data. Jonny is in (top 25% British product designer). Tom is not.

Finally, they tease an upcoming half-day AI design workshop open to individuals — coming late July/early August. Email [e-mail] with feedback, ideas, or complaints.

Chapters
0:00 — Welcome & podcast updates
New email address: [e-mail]

1:23 — Guests & format chat
Guests are coming — but this pod isn't going guest-only

2:00 — Jonny's golf trip to Mallorca
16-man Ryder Cup-style tournament, Mallorca, 35-degree heat

4:06 — The Strider Cup app — vibe coded on Lovable
How a non-technical friend built a live scoring app with AI — and why personal software is the perfect use case

8:59 — Can AI improve your golf game?
Spoiler: Jonny is a lost cause

10:49 — Are middle managers finished?
Tom introduces the "Middle Managers in the Firing Line" article and his Monzo perspective

18:36 — Tom's take: the functions matter, not the title
Is Jack Dorsey right? Tom's nuanced answer on direction, psychological safety, and player-coaches

22:02 — Jonny's take: I don't buy it
Manager-to-report ratios, the pastoral role, swanning around looking busy, and the market doesn't care what you enjoy

32:56 — Quick fire: Ideogram
A small, opinionated image model built for visual design — and why it points toward personal, self-hostable AI tools

38:27 — Quick fire: Midjourney's wild pivot
From sci-fi image generation to… health and wellness medical devices

43:50 — In the Weights — are you in the training data?
Tom and Jonny test themselves live. Results are humbling for one of them.

47:55 — Upcoming AI design workshop
Half-day, ticketed, late July/early August — build your own AI design tools

---

## Links:
- https://www.linkedin.com/news/story/middle-managers-in-the-firing-line-8966722/
- https://ideogram.ai
- https://www.midjourney.com/medical
- https://intheweights.com

Find us at nearfuture.works or email the pod at [e-mail] with suggestions and ideas for content or guests.

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 back
to the Near Future Podcast.

we're Tom and Jonny, and, we're
here bringing you on the journey

of our explorations, learnings,
what we're building with AI, and

sharing where we've got to so far.

today is extremely hot and
sunny in London and Manchester,

where we both join you from.

you can see, if you're
watching this on YouTube, Jonny

sunglasses for the occasion.

I'm not wearing sunglasses, but, but I…

Maybe I have better shade in,
the space I'm working from today.

so thank you so much for coming back,
especially to all the listeners who've

been, tuning in since the first episode.

There must be at least two
of you or three of you.

enough, in fact, to demand that we
create a, dedicated email address,

which is now [e-mail]

Jonny: Yes.

Email us your content
ideas, your complaints.

there is an ombudsman.

and, any feedback.

Honestly, we're just getting started.

We're trying to work out what this
format is and what's most interesting.

We do have some guests coming up.

So, far Tom and I have been doing this
on our own, and love you man, but,

let's get some fresh faces on here
and bring some tension, onto the pod.

Tom: Absolutely.

Yeah.

slash just, fresh perspectives,
exciting new things.

excited for that.

E-emailed, messaged, messaging a few
people, a few special guests to come and,

join us for the conversation, and we have,
I think, three or four on the horizon.

Jonny: Yeah

Tom: to introduce, our huge,
massive global audience

to, these fantastic people.

Jonny: But there'll be
plenty with just us as well.

We're not--

Tom: yes.

Jonny: have guests every week.

that's too much admin
for any of us, I think.

Tom: I, will segue us into
this next section, which is

gonna be about, your, trip.

You went on a trip late last week, Jonny.

You just got back, and, were
telling me about something cool

that happened with a Vibe Coded app.

So

Jonny: Yes.

Yeah,

Tom: us

Jonny: yeah, I've, I have, been
out of the world of AI and design

effectively since our last recording.

I think we recorded last Wednesday, and
then I went on a golf trip on Thursday

morning and came back on Sunday.

We- it's now Tuesday, so I had one day o-
of desperately trying to recover yesterday

from three rounds of golf in 35-degree
heat in Mallorca and three big nights out.

so I go on this trip every year.

it's 16 of us, and we
play Ryder Cup rules.

That won't mean much to a, lot of people,
but, it's basically a team format where

we all, have t- two teams of eight, and
we play the same teams every year, and

we dress up in red, and blue, and, it's
a kind of classic boys need competition

in order to make a holiday work.

And, we're now into our 16th
year, so there's lots of lore and,

memories of big, wins and putts to
win it and all that kind of stuff.

but one of my friends, Chris, has been
keeping track of all of the stats across

all 16 tours in a spreadsheet, and he
started doing it in year one or very

s- close to the start of the tour.

And through, I think, 46 different
players, in our rotating, list of golfers

and across 16 years, so that's three
rounds of golf a year, a lot of rounds of

golf, he has, got this sort of wealth of
data And so every year we have a bit of

a data roundup and, how did people play
off their handicaps, and are the teams

balanced, and all of that kind of thing.

and we use on the course this
app called VPar, which has

been around since we started.

It's a little bit old and
rubbish, and very complex to use.

And this year, Chris, who is not a,
designer or an engineer or, anything

of the sort, thought, "I'm gonna use
AI to build my own one of these, and

it's gonna be simple, and it's gonna
be just for us," a personal one for

our, tour, which we call the Strider
Cup, named after Tinchy Strider.

dates the tour somewhat.

Yeah.

Tom: Great reference.

Jonny: so the, Strider Cup app,
came to life on Lovable, and Chris

hadn't heard of Lovable before.

He Googled it and found it
and started building, and it

was an unbelievable success.

this-- I think it was just the perfect
example of this sort of personal

software or small-scale software
that you could never have dreamed

of building, or you wouldn't have
paid a software developer to, to

build something for a 16-person tour.

for a long time, I actually
owned stridercup.com

but didn't do anything with it really.

so finally this app has appeared, and
it has, it allowed us to track each

other's score live on the, course.

It provided us with stats and, even
he built betting odds into the app, so

that we could live bet, on each other.

and, everyone was so excited about it
that then I took Chris to one side.

I couldn't help myself.

I took Chris to one side, and
I was like, "That could be…

This could be a thing, If you wanna…"

he works in golf at the moment.

he works for the DP World Tour.

and I was like, "I feel like if you built
this a little bit more, then there'd

be an opportunity for other groups like
us to use it and maybe pay you for it."

But it was-- I found it very, inspiring.

It was a genuinely good bit of software
that he built by asking natural language

questions, Rather than we need a
database, he would ask something like,

"We need it so that people can see
each other's scores live and, up-update

each other's scores or update their
own scores so other people can see it."

And then it-- Lovable would
know what he meant by that and

build something appropriate.

so I just, I loved seeing that because I
think it was about 18 months ago, a year,

18 months ago that we started playing
with Lovable, and it really was the first

tool of its kind, I think, that broke out.

Maybe that and Replit really broke
out as like now normal people can

actually build software that is
meaningful to themselves and to

others, and is the perfect use case.

I-- people have built big businesses
on Lovable, but I think the perfect use

case is this, small thing that, you can
have super high context over what you

need from it, and you just build it for
yourself, and then all your mates use

it, and you can put your in-jokes in it.

Like we-- the, app was sponsored by Bo
Deans for historical reasons, and, had

a Rolex clock in it and stuff like that.

So there's, you can put gags in your
software that only are meaningful to

16 people, and justify it entirely.

So yeah, it's, a, good reminder
because I haven't really opened

Lovable for a year probably, that
these, bits of software are really so

easy for people to just get started.

it's made me think I should dig back in.

I think I've got-- still got a
ton of credit to use, maybe I

should, dig back into Lovable.

But yeah, so it's-- that, that was
really my only, the only th-thing I

could think of to bring to the pod in
terms of experience with AI this week.

Tom: That sounds, that sounds good.

That sounds great.

Have you, you tempted to try and
use some, prompt injection, dark

magic to hack your score so you
were doing better in this app?

Or were you ch- keeping

Jonny: w- it, th- I think one, one of
the most interest-- Yeah, it, That's

a really good question because I
think s- trust was implicit in, this.

There's, no auth, anyone can go
on this, app, which is why I'm- I

won't link to it in the show notes.

and change each other's scores, but,
you don't because you're all friends.

So there were various technical
things that Chris didn't have

to worry about, tracking, events
and working out who did what.

who changed the score?

Was it actually the person that was
playing the match, or was it someone else

who was messing around with the format?

and, yeah, so to, to, productionize
it, you probably need to do a little

bit more, but just, beautiful.

And it, because you didn't have to log in,
everyone was like, "It's so much easier

than anything else we've ever used."

yeah.

W- we've known each other for long
enough that you, wouldn't last long

if you started trying to cheat.

Tom: Yeah, definitely.

I guess that's no different
from a spreadsheet if you're

giving everyone access

to a spreadsheet and they're
trusted to update it with,

Very cool.

big question though for you, Jonny.

Have, has it led you to thinking about
how AI can improve your golf game?

I've n- I know nothing about golf.

I've barely ever played it, so I don't
know how much, how, AI could, feed in,

but, has it left you with any ideas
of how to improve your score for next

time using AI to help in some way?

Jonny: one thing I think I probably
did tell you ahead of this is

I did not play well this year,
But I'm not sure AI can help me.

Unfortunately, I think it's just
completely reconstructing my swing

from the ground up, and spending
more time playing golf than I do.

that said, I did-- I was captain
this year, so I had to make

decisions about who goes out first
and how to structure the team.

And I did throw up a little,
mini website with some AI in it

to help me decide the pairings.

But,

but for me personally, no,
I'm, a lost cause, I'm afraid.

I won't be troubling the
PGA Tour anytime soon.

Tom: Okay.

Okay.

That's, it's good to know there
are, there's some limits to where

AI can help and where it can't.

so maybe when robots are part of
the picture, maybe it'll become

more complicated for the, golf
team of 2050 that's gonna happen.

okay.

Let's, let- let's move on.

Let's move on.

I think we were, going
to talk next about…

there's a couple of topics
we were, thinking of sharing,

on the pod today.

how about I go next and talk

Jonny: Yeah, I think I, I'm intrigued.

you shared a link that I haven't
clicked on, and I think it's gonna

be pretty relevant to us and our
audience, especially given your

last role and, yeah, I'll hand
over to you, Tom, to introduce it.

But yeah, we're, going a little
bit more serious than golf now.

Tom: so the headline of this, article
was, Middle Managers in the Firing Line.

and so I'm just-- I have it up in front
of me here, and it says, "CEOs at some

of the world's biggest companies are
rethinking their org charts, pivoting

away from middle management toward
the idea of an intelligence layer."

And some other senior-sounding
person says middle managers

probably shouldn't exist anymore.

I, the reason I thought it would be
worth, sharing today is I think this

has been, a narrative that's been
around the last year, two years,

even before AI really took hold.

I think it was the sort of post-ZERP era
of, the macroeconomic climate where it

was harder to raise money and there was
more, talent around and, fewer headcounts

than previous or there were before.

and so I think it's-- I think starting
at that point, there was a conversation

around do we really need people managers
in the same way or middle managers as this

article talks about, and that seems
to be heating up, with this latest

piece that I saw recently on LinkedIn.

but I wanted to bring it up because
I think there's a few things

here that, I've been definitely
thinking about a lot, recently.

And you mentioned my last role, Jonny,
and for any, new listener who has zero

context of who I am, previous role,
full-time role before this was as

director of product design at Monzo.

I was there for five years, and a big
part of my role was, hiring product

designers, managing product designers and
managing managers of product designers.

so definitely got to see firsthand,
how this shift was at least starting

to take shape, And I think it's just
a-- it's definitely a moment to…

there's a lot of thoughts that this
has led me to, and Jonny, feel free

to ask me follow-up questions in, this

area.

Jonny: Yeah, Oh, yeah, I have some.

Tom: th- But, but, a couple things
just to anchor the conversation

that I, I'm thinking about.

The first is, shift toward
the intelligence layer

and what that, that means.

Jack Dorsey and, his, Cash App and
various other companies he's connected

to, have made some quite radical changes
in their organizational structures and

moving much more toward player-coach
model managers who are mixing IC

work with management contributions.

that's one example of this.

there are various other flavors of
trying to the-- or automate some

of the information coordination
and gathering across teams

and, positioning, managers in the
traditional sense to focus more on sort

of strategic impact or getting closer to

individual contributor work itself.

I think, yeah, so Depending on who is
listening to this, I think there is

the sort of version of this that there
is the element of, truth to, you are

in a middle manager role right now, I
think that's a difficult place to be.

difficult in the sense that you
have pressure coming on you like

most roles that are somewhat
connected to AI right now.

This is true of

Jonny: Yeah

Tom: it's true of everyone in,
working in these companies.

But I think people managers are, in
my experience, absorb a lot of the

emotional labor of how, the anxiety
and fear that surrounds teams, and I

think a lot of that falls on, leaders
and managers to, set clarity and

direction in a time where it's…

there's a lot of uncertainty, and that's a

really difficult place to be.

I think combined with that, you're
going from managing six people to n-

now managing people or 18 people or
even more, which I think the traditional

systems and structures just aren't really
set up to scale very easily to that.

it

really requires re- redesigning
from the ground up, what does that

even look like in this new era,
be able to operate effectively.

But I think most fundamentally, it
is what is the, outcome, are the

outcomes that you're prioritizing
as a manager to, justify validate

the areas that you're focusing on.

And I think that is something
that a lot of companies are

still figuring out to an extent.

But the old rule book of primarily
being about hiring and growing teams,

developing, getting more from individuals
and teams and helping coordinate

them, I think is getting thinner
and thinner as the sole priority.

Jonny: Yeah

Tom: I think it's really about
how do you Scale your impact

and, drive that through others.

But the one tactical thing I wanna add
before jumping to any questions you

might have, Jonny, or thoughts you…

i'm curious to hear your thoughts
too, is around, this, this tool that

I think I would find very useful
if I was, in this position right

now and I was a manager of people
anywhere, in any company at the moment.

it's something that I learnt about during
my time at Monzo, but I, don't know how

long it's been established beyond this.

the concept of servi-
manager service levels.

I forget the original engineering
leader who introduced this at Monzo,

but the basic concept is, you have this
sort of different degrees, of what you

can expect from your manager during
different times, at work, and there

is the, everything is good, every…

This, the manager has lots of
capacity and attention, and this is

what you can expect a direct report.

And then on the other end of the
spectrum, there is, is broken.

We're hiring a million people.

everyone's on and sabbatical and,

p- that's, that- that's what the
expectations look like when, that

is the sort of picture of, things.

Jonny: Yeah

Tom: and I really love the, not
necessarily those examples, but the idea

of having this spectrum of, like, where,
as a manager, where do you w- how do you

expectations for your team, on, in the
good times and in the difficult times?

I think it, it introduces this idea that
are all construct, constructs, artificial

constructs of how we allocate and set
expectations with the teams we, work with

and the people we're responsible for.

but I think giving managers, and teams
a level of control over defining what

that should look like and then helping
people understand, what that range can

feel like, I think that expectation
aspect of being a manager much more

of an exercise in enabling managers
to, to pick and choose what they

think is most impactful to the team.

hopefully that involves actually
speaking to the team too and really

Jonny: Yeah

Tom: needs are and how they can be helped.

but I just wanted to throw that out
there as anyone who is struggling

to navigate this transition as
a manager, I think you have…

You're in a lot more…

you have more, control over the managing
expectations with those you manage

and the rest of the leadership team
that you work with, to help reasonable

expectations that position you to
actually do a good job and lead to good

outcomes for your company, not just

feel like you have to do with more with
less, which is, focusing a bit too much on

the, what is being taken away rather than
what you're being given and how that can

actually, unlock these new possibilities.

Jonny: Yeah.

Tom: Jonny, I've talked a lot.

What, are your

Jonny: no.

I suppose I've got one question
for you first, and then I

have a couple of thoughts.

But given all of that, and given
the sort of just the headline of

we don't need managers anymore,
what's your take on that?

what's your gut on-- it's you have been
a manager, you have managed managers.

Do you think that is true?

Or is that like a sort of reactive,
like post-Zurp, falsehood that

we will stumble back out of and
everything will just be the struct-

the same structure will remain?

is Jack Dorsey correct, basically?

Tom: I don't think he's,
wildly off actually.

I, think of this less in titles and
more in what are the outcomes that are

necessary for a team of people in a
company, and I think teams need direction.

They need clarity.

They need psychological safety.

They need,

around what a company
is trying to achieve.

I think all these things traditionally
fell on a manager to help shape and guide.

I don't necessarily think as a standalone
role is the most thing to preserve

in the future we're moving toward.

but I think making sure that there
is-- there are those elements

in some kind of leadership
capacity in a team is important.

I suspect what we'll see happening is
teams will continue to have a head of

design or a sort of leadership function,
because I think that person is most or

best positioned to focus on the direction
and the way that design's impact can

ladder up to company-wide outcomes.

But beyond that individual person,
I think it might vary based on how a

company is uniquely operating itself
and how it coordinates across different

functions, and whether that takes a
slightly more design operations role or

people management roles or, even a more
sort of chief of staff-shaped support for

that head of function, I think varies.

I could definitely see a world
as well where there's more…

the player coach, concept, which on paper
sounds like, oh, you can be a manager

and an IC, get your cake and eat it.

Or I think

Jonny: Yeah.

Tom: that and think, I can just get,

person to do the job of two people.

Great."

I think in reality, from my experience
when I've seen that, that setup, I

think it feels like 80% of the work
is as an individual contributor with

ment- with mentorship and some light
accountability around maybe some sort of

light performance performance management
and light performance review stuff.

but which I think works potentially
quite well if the s- the team

skews very senior everyone is
high-functioning, highly capable, highly

aligned with what the company needs.

that is not true of every single company.

so I think there's a risk that
the player coach who stumbles

into a more challenging,

more challenging management challenges
might struggle in that kind of

setting, unless they have a background
and experience to, to dial up.

but yeah, an interesting quirk
of the sort of player coach era.

Jonny: I, suppose I just…

I think I haven't read this, article,
but I read a month or two ago, yeah,

Jack Dorsey's, a bit of Jack Dorsey's
very long piece about management

generally, and I don't buy it.

First of all, I think that, the
market doesn't owe anyone anything.

So if anyone is a manager right now
and, really likes, aspects of it

that may become less important over
time, which will-- I think is true,

then that is, that's gonna be tough.

But I think I don't buy…

I think, manager to report ratios
is a function of how many people

you have on, your team primarily.

if you think about the pastoral role of
a manager, and if you think about the,

taste bar setter role, and if you think
about managing up and across, all of

that kind of thing, actually a lot of
those ratios are set by the, world around

the person rather than anything else.

And I do think, like a tastemaker,
maybe a design team of 100 could just

have one tastemaker who sits in these
great big crits and, and actually,

having one person deciding what good
looks like is more efficient as long

as they can get round all the work.

So that separates off from-- That becomes
more like a super senior IC or something,

and then you have managers that are

communicating expectations up and
across and, some of that last part

feels like it might be less necessary.

But at the same time, everyone's
team is about to go through this mad

change where they do need, as you
say, more psychological safety and a

little bit more support possibly than
they did over the last few years, or

a different kind of support anyway.

So I think that it would take a particular
type of culture to make what he describes

work, and maybe, Block, which is Jack
Dorsey's company is, that culture.

But I'd be surprised if that's
really broadly applicable.

But, may- maybe I'm wrong, but from
my time as a manager, which admittedly

was-- or a design manager, which was
a while ago, but then a manager of a

whole small organization, admittedly.

I don't think we can just, automate
away how we communicate with each

other and how we transmit information.

I do think your whole job being
sitting in meetings all day has

always been ridiculous, and it's only
a function of the ZIRP era that it

was ever allowed in the first place.

but people removing management layers
because of AI, I think I don't buy.

And as you say, the post-ZIRP was
where the actual change happened,

where people went, "Yeah, you
can't just exist here, so everyone

feels a little bit more supported.

You need to show ROI."

and I think a lot of this change now
is look at how fast the ICs are moving.

Look how many lines of code and
how many PRs the engineers are,

merging, so therefore everyone else
should be as efficient as them.

But actually, other parts of the business
don't move as fast as that, and things

that involve humans who are messy and
communicating in rooms with people are,

like, not the same thing as writing code.

seems like a really obvious thing to say.

I think there might be a little
bit of a rejection of that concept

more broadly, and most companies
will keep more managers on.

Or at least if they're removing them now,
they might think, "Oh, it didn't work

very well," and bring, some of them back.

That said, I think every manager needs
to w- look at where they are adding value

today and work out if that's the places
they're really focused on are what their

organization needs in the future, and try
and predict the future a little bit here.

'cause you don't wanna be-- What might
happen is managers are split into

managers that drive real business value,
which might be, like, better products

and higher quality design or whatever
it is the department they're in, versus

those that swan around looking busy.

And, if that really is you, swanning
around looking busy and making sure

your team have, a really fun breakfast
every quarter, then that, that becomes

less, value, in this world compared
to your peers, and then you might

be in a little bit more trouble.

But I don't know for sure.

I don't know.

But,

it's an interesting time.

I think that there's, this,
narrative which is like all

managers need to start designing
again, and maybe some of them do.

But they're, the ones, ironically,
that are probably sitting most

comfortably in a management role
right now where they're not actually

involved in the work very much

Tom: Do you think they sh-
do you think they should?

Do you think anyone who is in a purely
manager role right now should get hands-on

and design things or build things?

Jonny: I think to me that's
where it goes back to the market

doesn't care what you enjoy.

The market cares about what, is valuable.

And you may be forced into being one
of the best managers or not being

a manager anymore by that fact.

And so if you're not a manager
anymore, you need to work out what

you are, and that might be an IC
or it might be something different,

like a potter or something.

But, but I think-- I th-
and, that's not about design.

That's just, that I think that
might be what th- this, technology

does to whole industries and
all knowledge work ultimately.

Tom: Yeah.

Jonny: Yeah.

Tom: yeah.

I, I, that makes sense.

And, I think to build on your point,
I think really understanding the

mechanics of how the business you're
in makes money and, how your function

ladders up into that, I think is to
have significant-- a significant bearing

on the role that you play in your
team add value toward that top line.

Jonny: Completely

Tom: I think that will probably look
quite different based on the organization.

I think-- So I think, yeah, given the
shape of your team and the role also

design plays for your business, or,
product plays for your business, I

think could have quite a big bearing on
the importance of or what, is uniquely

important about design leadership within
that organization and how much, how much

that is about taste versus vision versus
direction, which are all overlapping

areas, but I think they vary quite a bit
based on what a company needs right now.

thing to mention on this, I saw a post,
post by, Tom Mora, who's head of design

operations at Fin, from earlier this
week, which didn't directly talk about any

of this stuff that we're talking about.

but there were some themes in what
he mentioned, I think do overlap.

There was one that was around design
being uniquely positioned for the

first and last mile, so like setting

direction and deciding
what to work on and then,

being the guardian of overall
quality of what ships.

then another one that was around, into
AI and automation for communication

alignment and, the information
layer of, how a team operates.

I think those for me, definitely make
sense in a design operations capacity.

I think are also areas that historically
have been more in in the realm of the

people manager role within design.

Jonny: Yeah

Tom: think those roles have already always
had some overlap, but maybe that overlap

is, increasing and almost abstracting away
what the overall design function needs in

your company into, yeah, what are the sort
of operational aspects versus what are

the taste and direction setting aspects

versus what are the sort
of business impact aspects

Jonny: Yeah

Tom: fall out obviously as, oh, we
need five people managers, but we

need someone who's really good at
operations, someone who's really

good at taste, someone who's
really good at this thing.

I, think that's a bit
disconnected from what those

titles will be, but the functions

Jonny: yeah.

Tom: I think

Jonny: Maybe the thing that needs
to, maybe the thing that needs to

die or move on is actually just the
title and because it comes with all

this baggage of what is expected

Tom: Yeah,

Jonny: historically.

And in this new world, a lot of that
stuff may either be divided up across

different people or across different
functions, and some of it can be done

by ICs or be done by product managers.

and so therefore, what is a design
manager anymore is probably not, might

not be a thing, even though all of the
functions of a design manager exist.

And so therefore, yes, you might
need fewer design managers, but all

of the same capacity is required of
what the design managers used to do.

Which then sounds a little bit more
like player coaches and other stuff.

But, or you have these sort of guilds
of, I don't know, wizards and orcs,

and wizards are really good at taste,
and orcs are really good at, once ones.

I don't know.

Tom: being good at one, one-to-ones,
that's probably not the most

part of, this whole strategy.

but, yeah, I also think there's
probably a, really exciting world

of the AI-enabled manager who is,

really using AI to scale a
lot of what they're doing in

the space and actually has

Jonny: Yeah.

Tom: impact.

Jonny: Yeah

Tom: but the label for that is
perhaps a sec- secondary thing

to, to explore or consider.

won't even un- open the can of worms
around, cross-functional management

and if, a design-- if, is, may-
maybe a design manager doesn't

exist, but, a people manager or
people who ship things to customers

something that sort of spans disciplines.

Jonny: Yeah, interesting.

Yeah.

Tom: let's not go down that path.

It'll have a whole other podcast episode.

We're already a long way into this, but,
but really, yeah, fascinating, to think

about this and the sort of implications.

There was a whole other side of this
around like talent retention and how that

has shifted the dynamic right now as well.

But

again, I don't wanna go
down that path either.

Jonny: save it for next
week or another week

Tom: Yeah, so these are just
teasers of what, what's to come.

we did-- was another thing we
essentially wanted to talk about.

do you think we have time, Jonny?

Do you wanna talk about Ideogram?

Jonny: let's, let's do
it as part of quick fire.

So what--

Tom: Okay

Jonny: we have a new quick fire section.

I don't know if we're calling it…

we now are calling it quick
fire section outwardly.

so welcome to the quick fire section, But
I can kick off with a cool tool that I

first heard about on a podcast and then
have played around with a bit this week.

It's called Ideogram, ideogram.ai.

it's-- I think, one thing that

these models, even the state-of-the-art,
frontier models really struggle

with is design and taste.

The models out of the box just don't
have-- or they have a taste, but because

they are trained in specific opinionated
ways, the taste is always the same.

So if you ask Anthropic or if you
a-ask Claude to design you a website

and 1,000 other people do, all of
those websites will look the same.

And we've seen that, first it was like
the blue-purple gradients, and then

it was the greige sort of milk soup,
milk, toast, with italic, serif fonts.

And, we see these like phases
where the new model is trained on

a new set of data, and so therefore
that's what all design looks like.

And I think a lot about this kind of
stuff and, how do you move away from that.

And the harness, which is basically
all of the prompts and everything

around the model can change that.

You can tell the model to come up
with five variants that all look

different, that kind of thing.

but ultimately, ideally, you would
have a model that s- looks for and

celebrates divergence over convergence.

So it is forcing itself to
come up with routes even if

some of the routes are wrong.

And, there's some bad ideas and some
good ideas, but at least you've got

some ideas, and you can make it look
different, with different prompts.

Anyway, thinking about it, and
then, I hadn't heard of Ideogram,

and last week I listened to,
their founder talk on a podcast.

It's basically a s- a, a whole new model,
that is quite small in terms of, size.

It's nine, nine billion, parameter
model, so way smaller than

the, big state-of-the-art ones.

but it's trained in a certain
way, that, effectively makes it

very good for, visual design.

And so it can do things like come
up with variants, and it can even do

things like add editable text to images,
and you can give it a layout and draw

boxes over how you want that layout to
look, and it will fill that layout s-

with a mixture of text and images and,

And it's not good at anything else.

It only takes JSON as a, an input.

it's very opinionated, but it does
this one thing well, and I th- and

it, m- just I've been playing around
with it, and it's pretty good.

I'm not e-excellent at it yet, so
I'm not producing incredible things.

But,

I think it points to a future where you
start to rather than have these great

big AGI-esque models that are meant to
be good at everything, you start to have

smaller models that are good at one thing.

The most exciting thing about this,
because image models is tricky,

there's a lot of training data in
there that is effectively stolen.

and one thing that he said on this
podcast is, you can take this model,

which is open weight, and you can
train it yourself with your own work,

and then it can become your own model.

and because it's only nine billion,
in size, it's-- you can run it on

small hardware and run it yourself.

And so you can start to have what
looks like a little bit more of, a

personal model as a tool, which I
think has been hard to imagine so far.

So it's an exciting direction.

I've only just started
playing around with it.

I think the pricing looks fair.

and we do actually have an opinionated,
art direction for our podcast artwork and

our artwork for our s- site generally.

So I'm thinking we could set up a
little workflow where we can generate

consistent artwork with, this tool.

I will-- I'll basically report back
when I've put something together

that I think might work and, we could
even host it ourselves somewhere

and then, it really is just ours.

Train it with our own, favorite
images or our own images.

so check out, ideogram.ai.

and, I haven't played around with
image models for ages, really.

Not in, in any sort of way
where I'm evaluating new ones.

I think we've played a little bit with,
Flora and some of these kind of more

visual tools, but it's exciting to see
innovation in the w- in image models.

And actually, this is, this is such
a good transition into, an image

model tool or company that has
built something completely mad or

unexpected, which is Midjourney.

So

I spotted that this week and
had to double check it wasn't a

prank or it wasn't April the 1st.

it's so unusual.

But Tom, you wrote it down,
so you, tell us about it

Tom: yeah, very interesting.

Thank you first of all for that.

That was a very long quick-fire round.

Jonny: know, sorry.

Tom: r- but it

Jonny: It's a

Tom: very

Jonny: mid-fire, mid,

Tom: Midfire.

Yeah.

Jonny: speed.

Yeah.

Tom: we can work on this.

we can work on this.

We can shrink our speed,
for these quick-fire things.

yeah, so the Midjourney pivot, Midjourney,
people might know it, know them from their

image generation, technology capability
that's been there for a long time.

I feel like they've been one of the,

the image

AI companies for the last few years now.

Jonny: Smith eating spaghetti.

I'm sure it wasn't that
then back in the day.

Tom: Oh,

Jonny: May-

Tom: Yeah.

Yeah.

Jonny: know.

Tom: that was the great
challenge of image generation,

Jonny: Yeah

Tom: managing to overcome that.

anyway, last week they pivoted
to, running a wellness, health

and wellness-focused spa company.

I don't know if that's a fair
description of what they do now, but

they have an exciting website which
has a very cinematic video of, a woman

lower- lowering herself into this
water-type container and scanning

lots of vital signs and giving a
health report on various things,

making health checkup more
like a spa experience.

So yeah, big change in
direction, but, interesting.

I've been to Nico Health in London,

I don't know if you're familiar
with that, Jonny, but they

Jonny: I've been to the one in Manchester.

Tom: Oh,

Jonny: Yeah

Tom: Yeah.

They've, I feel like it's in the kind of
space of imagining what is now possible

with the level of, health data that it's
easy to capture now and creating a whole,

an experience around making that
far easier to get that data and

level of insight and, commentary
from medical professionals in that

Jonny: Yeah.

Tom: So yeah, really interesting

Jonny: But

Tom: But yeah, a wild pivot

Jonny: The, stats around
this are mad though.

It's like way cheaper,
significantly cheaper than an MRI

to do-- to produce better data.

So like I think, Neko Health
is interesting and great.

I, come from a family with history of skin
cancers and things like that, and what

Neko Health does amazingly is it does like
a full body photo, and it tracks every

single mole and everything on your body.

and they do a blood test
and other stuff as well.

This I think from reading…

I can't find the bit in the article
now, but describes, being able to take

internal body scans to war zones and
all of these things that you just can't

do easily, or cheaply anywhere else.

So it like dramatically reduces the
cost and the complexity of, finding out

what's going on in the inside of us.

so it could have really
profound impact on medicine.

and like I think if you h- to hear that
bit of news and it was from a medical

device company or some enormous I don't
know, postdoc at Cambridge or something

like that, you'd be like, "Yeah, cool.

Good.

Advances."

And just f-for it to come from a,
company that you'd associate with

like slightly edgy sort of sci-fi
image generation is a real curveball.

So it'll be interesting to see
how they balance Midjourney

Medical with Midjourney.

the only, the only-- the company
it reminds me of actually is,

Aerobie, which is like a Frisbee, an
AeroPress invented by the same guy.

Tom: Yeah.

Jonny: Just like he's a-- That
is a true inventor, and it feels

like Midjourney are inventors.

they're, like, "What else can we use
this bit of-- this algorithm for?

If we put it in a camera,
what could it do?"

And then they didn't really think of
anything as outside their purview,

Tom: Yeah.

Jonny: cool.

Tom: That is true.

That is true.

Yeah.

Jonny: It ma- makes you think about what
these companies are going to end up being.

This-- the, An- Anthropic setting
up wet labs to get into bio and,

Midjourney building medical devices
and, these companies that have this

IP around intelligence don't have to
stop at consumer hardware, which is

what the last generation of companies
did, and, actually can say, "What

problems should we solve in the world?"

This is an optimistic, alternative to
this like dystopian, world run by tech

bros who are entirely disinterested
in anyone other than themselves, which

I think you can see the sort of Thiel
and Musk e-end of things is that.

But then there are also people
saying, obviously they, wanna make

some money out of It, it will be a
valuable company, but it's also just

something that is great for the world.

yeah.

What-- It'd be interesting to
see where some of these companies

end up and what they, what other
lines of business they have, but

Tom: Yeah.

Yeah.

So true.

Jonny: It's a good spot

Tom: speaking…

Yeah, so true.

And, speaking of lines of business, I
think there's one final thing that I

wanted to make sure we covered today,
in this podcast, which is a website

called In the Weights, which has been, I
think, doing the rounds on social media.

I'm gonna share, my screen actually
with, with In the Weights on it so

you can see what I'm referring to
if you're watching this on YouTube.

it is a website that helps people
understand if their name is

specifically used in the training
data for each of these AI models.

And I thought we could close today's
podcast by, typing our own names into

this website and seeing who comes back
as being more in the weights or less in

the weights or not in the weights at all.

I don't know.

Jonny: Yeah.

We can be mysterious.

Yeah

Tom: we have not done this before, so
I'm gonna, I'm gonna try this live.

I'm not gonna add my credit card
information, which just popped up there.

That be a terrible idea.

Don't do that, please.

Okay.

I'm gonna add…

I'm gonna search for you.

Is Jonny Birch in the weights?

You have a 22…

228 strength, top 25% British
product designer and entrepreneur.

Oh,

Jonny: Oh, I'm in,

Tom: and

Jonny: OpenAI knows a lot about me, and
Anthropic know a little bit less about me.

And then who was the other one?

If you just scroll up again.

Yeah,

Tom: Kimmy

Jonny: a cyclist and a songwriter.

Tom: Oh, oh, so this is not you

Jonny: Ki- Kimi K too.

No, the top one is me.

Top one is me.

Tom: Also, you are

Jonny: And then there's two other…

I'm the top Jonny Birch of three
Jonny Birches, which is exciting.

Always wanted to be the
best Jonny Birch I can be

Tom: Oh, you do a great job.

Jonny: Yeah, Kimmy.

I don't know.

I th- yeah, d- don't know much about Kimmy

Tom: Okay, clearly

Jonny: you, Tom?

Tom: in the training data.

so you got 228.

Do you think I'm gonna
get higher or lower?

That's … Or just n-
I'm … Maybe I won't exist.

We'll see.

Oh, this is thinking harder.

probably a sign that
I'm, not in the weights.

I have not actually tried this before
doing this live, so we will see

Jonny: Oh, he's really thinking

I can cut out some of the silence

Tom: Yeah, or just so we can

Jonny: Exactly.

And

Tom: speed up this wait

Jonny: 21 seconds

Tom: clearly I'm the
first person to enter this

Jonny: Yeah

Still thinking

Tom: So I'm not-- this

Jonny: There we go.

California Republican poli- no.

Tom: Yeah.

Jonny: Australian politician.

Tom Harmon I feel is a bit
more of a common name, though

Tom: Yeah, maybe.

Maybe.

British

Jonny: he's very-- he's top 7%.

Tom: Which one?

Oh, yeah.

Yeah, this is, Actually, I
think he might own the original

domain name, tomharman.com.

so

Jonny: would make sense.

Tom: been my nemesis for many
years with that domain name.

Jonny: How do we topple him, Tom?

This podcast needs to go really well.

Tom: but yeah, not Australian
politician, not a British former

pr- football defender, not a British
social work professor, not an

Australian rules football player,
not a 16th century, English writer.

So I'm, I'm basically not in there,
I think is the answer to this.

Showing four possible hallucinations.

Nope, definitely not a conservative MEP.

think I'm that either.

A British journalist, Australian rugby
league player, an English cricketer.

So no, I am not in the weights
is the answer to this, question.

Jonny: Wow.

I'm so sorry, Tom

Tom: I think in the weights is a
proxy for being, and, by that measure,

Jonny, are more famous than I am.

I think we've also learned that there
are a lot of people with my name who

Jonny: Yeah.

Tom: fam- more

Jonny: you've got tough,
you got a tough competition

Tom: So maybe my name is more famous than

Jonny: Maybe we can make you infamous.

Tom: we, shall see about that.

Um, anyway, that was fun.

In the Weights, try it out, see if your
name is in the dataset, or probably

just don't give them that information.

Maybe we don't need to put our
names into the internet anymore.

before we wrap up though, one last
thing I want to mention, and maybe

this is a teaser for a future is that
we have been cooking up an exciting

design workshop, which is open to--
will be open to independent people

who want to sign up and join us for a
live online half-day workshop focused

on building your own design tools.

so we'll share more about
that in the next few weeks.

But now, know that is coming soon.

It will be ticketed,
late July, early August.

So if you are working at a company that
you're interested in, spending more time

building with AI and you, would really l-
like an excuse or you would like structure

or you would like support and expertise
around you to help do that, we will be

launching that in the next few weeks.

So keep your eyes peeled and, be a
great use of that personal learning

budget if you have one as well.

So

Jonny: yeah.

Tom: keep that in mind.

Jonny: We did some of these this time
last year, and we made them four days

long, and the thing we learned was we
should not make them four days long.

So half a day, only
half a day of your time.

Tom: Today.

But, we have also run, a similar workshop
to this with, a, few companies and

the, specific format has been tried
and tested and feels, very effective

for a cross-disciplinary design group.

So we have run it with a group of up
to 60 people, so who knows if we'll get

to that number with this, but it, it
scales quite nicely to, to large numbers.

check it out.

Come along if you fancy it,
and, we would love to have you.

But more details with actual information
and a URL, we'll send you that soon.

But I think for now we should probably
wrap up there, unless there's any final

words you wanted to mention, Jonny?

Jonny: No, please do tell us
what to put in our podcast.

we would really appreciate
guidance on this.

If you're invested now, if you got to
the end of this episode, then you are

invested to the tune of 50-odd minutes.

So spend one extra minute and
send us a message, and that

is [e-mail],

or if you know us, then just
use whatever communication

channel you, you know us through.

and that's it, yeah.

Until next week.

Thanks, Tom, and let's go and enjoy
the 39-degree balmy heat of, of

London and Manchester this week

Tom: Yeah, stay safe everyone.

Jonny: See you next week

Tom: we'll see you next week

Jonny: Bye