Dustin Senos - Designing Dia (the AI-native web browser)

Our software can be very different now

and it's like we finally as designers

have this new putty that lets us build

very very different tools. It's cool

because so much of the design process in

that world of personalization is

completely removed from the pixels.

Don't look at as like vibe coding. Look

at it as like vibe designing. We all

have the tools to express ourselves at a

high level of fidelity. The only thing

that matters then is having an

understanding of what is possible. Your

motivation shouldn't be like likes and

stuff on the Twitter. It should be

understanding how rewarding it is to see

a problem, a unique problem that no one

has solved before and solving it. It's

like it's it's the dream. No one is

going to give you a curriculum of the

things you need to learn, especially if

you're in a role. You got to figure out

how to do that. Welcome to Dive Club. My

name is Rid and this is where designers

never stop learning. This week's episode

is with Dustin Cenos, who's the head of

design at the browser company. So, we're

going to do a deep dive into how design

operates, what collaboration looks like,

and get a little behind the scenes of

what it's like designing and creating

the strategy for the all-new DIA

browser. So, let's dive right in and

learn the origin story for DIA. We as a

design team are really, really heavy on

prototyping and exploring, especially

early things, and we don't have like a

separate product team from design team,

at least not at that point. So when we

kind of realized ARC wasn't going to get

us to kind of the bigger dream we had

that we realized we wanted to pursue, we

just immediately started prototyping and

we kind of checked the boxes on the

obvious things first. It's like, okay,

what if we had a horizontal tab bar that

could switch into a vertical tab bar?

What if we could do that? What if uh

rather than having all the profiles or

spaces in one window, we made them uh

separate windows to be more familiar

like other browsers? And we just spun up

so many prototypes and tried so many

different approaches to that because

what we had realized with ARC, a lot of

the feedback we were seeing is people

who love Arc love Arc and they're

committed to it and they're like, "This

has changed how I work. This has changed

like the internet for me. This is

fantastic. Thank you." But for those

people who never crossed that bridge and

got that far, they just felt like Arc

was too hard to pick up and they

couldn't find value soon enough. So we

were really as a design team trying to

like deconstruct that and help ourselves

and the you know future users find value

earlier and I think that was that was

our primary role was just like really

really chasing down that value for

folks. Talk to me a little bit about the

early prototyping like how wide did you

go on that spectrum? What were some of

the things that you were exploring? We

we tried so many different things. I

think a lot of folks even almost had

their own browsers running. I have like

a Swift browser that I just hack on all

the time. Nate Parrot had an Archia

browser that he hacked on all the time.

Just like whatever code base and like

setup felt appropriate for us. A lot of

the crew also were like heavy in

origami. So we're pro prototyping many

things in origami. And the shape of some

of those prototypes looked like you

clicked the maximize button and now you

were actually just looking at like a

desktop inside of a browser that you can

like doubleclick on and open up like the

GitHub uh window inside of that to this

browser is just a chat thread. That's

all it is. Like we really tried all

kinds of different things and let our

minds really wander quite far. Real

quick message and then we can jump back

into it. So you know how I've been

talking about how I use Genway to do

research with AI? Well, something

surprising is happening. And just for

context, I use Genway in two different

ways. One is contextual interviews. So,

I prompt the AI with what I'm hoping to

learn and it has a dynamic conversation

with each person. And the second is

usability testing. So, I upload a Figma

prototype and their AI agent helps me

test them with people all over the

world. I mean, you should see the

quality of the follow-up questions. It's

pretty crazy. But here's the surprising

part. At the end of the interview, most

people say they're more comfortable

opening up to an AI agent than a real

human. And it's just another reason why

I'm hooked on the product. If you want

to try it out, there's even a secret

landing page just for dive club

listeners, which gets you 2 months free

and 10 credits to recruit people. Just

head to dive.club/genway.

That's gnw a y. For a long time now,

raycast has been my portal to AI. I

mean, all it takes is a quick keyboard

shortcut and all of the models are

instantly available at my fingertips.

But the problem was you had to be on a

pro account. But that ends today because

they just announced that Raycast AI is

available to everyone. No subscription,

no account needed. You get 50 messages

to try and you even get access to all of

their extensions. I mean, I'll put it

this way. If I'm going to get a new

computer tomorrow, the very first thing

that I'm going to install is Raycast. It

is that good. And you can try Raycast AI

for free today. Just head to

dive.club/racast.

That's r a y c a s t. Okay, now on to

the episode. You talked about this idea

of making sure that dia was easier to

pick up. Can we go a little bit deeper

there? like how did this north star of

what good design looks like evolve as

you made this shift from art to Dia? I

think one thing to touch on is like art

kind of came from from nothing. We

looked at all the existing browsers and

we're like what if there's like a

different way of doing this and we just

kind of let our curiosity lead us and

our minds wander and out came Arc which

was a realization that you know you

don't actually switch between that many

apps and you don't actually go to that

many websites potentially if you're

working there's a few places that you

could anchor and build a product around

a few different contexts which we pulled

up in spaces for DIA the realization was

that novelty inside of ARC was great for

some people who are like early adopters

are willing to pay the price of novelty

and was actually a detriment to many

many other people and that the future of

LMS and AI is actually less about

helping someone say organize their tabs

in their spaces and more about just

helping people get more things done

easier faster so then they can just get

off their computers and like go live

their life. Building a browser that is

more familiar means that we are more

interesting to more people and easier to

pick up for more people. I like when

people use this phrase of like a novelty

budget and you know you're now catering

towards pretty kind of a different set

of audience. You know you have to make

something that can break out of the tech

proficient bubble. So then given a

smaller budget, how did you think about

the right places to spend it with India?

We spoke a lot about um as you say the

novelty budget and we spoke a lot about

the balance of innovation or familiarity

and I kind of I kind of like to stretch

it further where anything innovative I

really think should strive to feel

familiar so that it doesn't it doesn't

have to feel totally different to be

innovative. I think the most innovative

things once you see them you're they

just look obvious. you're like, "Oh,

yeah, that that should have existed from

the beginning or why did this not exist

before?" So, we really really chose to

invest in our like innovation budget or

our novelty budget in the things that we

wanted to be uniquely different about

DIA so that they didn't get in the way

of someone learning or just like waking

up on Tuesday and deciding to switch to

DIA. And you know, we get a lot of

feedback online from I just want chat

inside of Arc. It's like, yeah, I

understand that. But that is the

audience of folks who have already

committed to ARC, have already brought

everything over to Arc, already have

invested everything in ARC, and then of

course they want ARC to let them go

further. But what we realized is the

novelty of Arc without chat was too

much. It was too much. It was a

detractor for so many people that if we

were just to add chat to Arc, it's like

we're we've just added even more things

to Arc that people would feel like they

have to learn. So it didn't feel

appropriate to say peacemeal things out

of ARC to whittle it down into something

we felt was was more approachable or

like brought people value sooner. It

felt like we did really need to start

start from scratch and really really pay

the novelty budget in places where we

thought it could make the biggest

difference. When you open up DIA and you

open up the AI chat, especially when

it's in like the sidebar, it feels

obvious, you know, it's like, well, of

course it's there. Yeah, I would imagine

it probably wasn't immediately obvious.

Like what was that process like to even

arrive at something where we can open it

and be like, "Well, well, yeah, of

course it's there." You know, it's

funny. We we spent so much time thinking

about where the chat button goes and

what that chat button looks like because

it can be at the what we call scrim,

like the platter level of DIA, or it can

be at the tab level. It can be treated

as like a sidebar button to be native to

Mac OS where it's like, oh yeah, just

show or hide the sidebar or it could be

like we have it now is labeled and we

had early prototypes that kind of pushed

the URL bar to the background and

brought the always present chat bar to

the surface more closer to the person.

It was at the bottom of the web browser.

So if you imagine like Dia looks like

this and at the very bottom there's just

like a big chat. Hey, how can I help

you? And what we thought was like it's

such a natural place to have chat at the

bottom of that screen because then you

could just chat the web page would just

zoom up and I would just be inside of a

chat thread. But then we realized well

now as soon as someone's looking at this

browser they go okay which input do I

type into if I want to go to you know my

my cooking website. Do I click in that

chat one and say hey cookooking.com or

do I click on the URL bar? Even simple

things like that. And we tried that for

a long time and we're like this this

makes it really clear you can chat with

the website. It makes it really clear

it's a browser. But even that like the

thought that someone would ever have to

think and stop and think about which URL

bar or what input they're going to type

into in a web browser was just like no

that's just going to uh encourage too

much tax on someone. So we we pushed

ourselves to be like it has to be one

box and it has to be smart enough to

route to where the person wants to go.

And that's hard because like muscle

memory from other browsers is so so

strong. You can type in Chrome and never

look down at any of the autocompletes,

press enter, and it's going to take you

where you expect it to go. So then we

had to figure out a way to insert. You

could could be going to the website or

you could be asking a question to our

LLM. And we had to figure out how to how

to do that. It's interesting because

like one of the first things that I

noticed when I was playing with the beta

was the fact that you're still making

floating and sidebar available. Maybe

you could talk a little bit about that

decision. Totally. We had three. We had

actually we have so we have full

response page, we have sidebar. Uh full

response page is if it's you're never

starting from a website. You're just

starting straight from chat. So but if

I'm on a website and I chat, we had

sidebar. We have the floating window

which we had now. And then we actually

had one that never brought you anywhere.

So if you clicked at the URL bar at the

top and asked a question, you know, who

won the last baseball game, it would

just answer right there. So that it

would actually never take you anywhere

cuz we're like, well, that why would you

have to open a tab for that if we could

just answer you in line and then you

could optionally expand that into a full

page or a sidebar. And we left that one

at the drawing at the cutting board um

pretty quickly, but we we we built it

out for the sidebar or what we call like

kind of the pip modality.

We built the pip modality for people

that would have smaller screens so that

if you don't want content reflowing or

you're used to seeing your website or

say you're working on an Excel

spreadsheet and you want all of that

content all the time, the pip works

really well to kind of live on top of

that whereas the sidebar does take up a

lot of real estate. What are some of the

other interaction patterns that you all

were trying to figure out or maybe where

there were a lot of different types of

prototypes when you were figuring out

like how the user should interface with

AI inside of the product? I think some

of the more curious ones we tried were

so let's say I'm on a website and I open

up the chat sidebar and then I ask for

like related things or things to

question like maybe the stance on this

article and then you show links inside

of that. We tried some prototypes where

it's like, well, what if we just let you

click on that link, the chat pushes

over, the website pushes over, and now a

whole new web view opens to the right of

that, and then you can just keep

following down this rabbit hole. So

rather than just constantly opening new

tabs, we just kind of let you thread

down in the conversation. I feel like

that navigation is always compelling cuz

it's like of course it's like this like

tree you can traverse down and then

always in practice you just end up just

getting lost in it. So it makes for a

really good Twitter demo though. It

makes Yeah, it makes and you when you

see it you're like of course of course

that's how these things should work. My

browsing is rabbit holes. Why don't we

just let the person like make a physical

rabbit hole? But at least for us we

landed on something that was like we

were just incurring novelty tax. It's

like, is this the problem we're trying

to solve right now? Or if you click on

something that looks like a link, should

it either open a new tab or navigate in

place? And then the decision was, well,

which one do we handle when? And then

had to kind of work through that problem

versus the other problems. I want to

keep coming at these early explorations

from different angles. And there's

another tension that my assumption is

you are probably dealing with, which is

okay, we want to have something that is

much easier to pick up. There's a

simplicity that doesn't exist in Arc

that we have to achieve and yet you're

also laying the foundation for this new

type of AI native browser that has to

support functionality and use cases and

even modalities potentially that are

almost impossible to predict. You know,

this thing has to be just as valuable

today as it is 3 years from now. And at

the rate of change, who freaking knows

what that's going to need to account

for? So, how did you think at like a

system level in terms of where does the

right levels of flexibility to exist so

that we can grow into these potential

futures? Totally. So I think that goes

back to like where the chat button gets

placed because the things we realized

were the spatial metaphors that we pull

on like it's very intentional that Dia

has a scrim has a kind of a tab wrapper

that sits on top of that platter. The

chat button is inside is on top of that

platter attached to the web contents and

that our URL bar and tab actually match

the color of the web contents. So then

at that level that gives us a container

that we can play with as far as we want

to explore things with bringing context

to the person assuming that the if we've

done our job right they realize that the

context of the LLM's bringing them is

using that container and everything they

see inside of that container. So that

kind of gives us flexibility to play

play with inside of that. And then at

the grim level, things like bookmarks in

DIA are very similar to other browsers

and other browsers that people are

familiar with and very different than

Arc. You know, we tried many different

explorations on like, well, how could

bookmarks be different inside of DIA and

well, what if they're, you know,

autocategorized or what if all these

things that LLM let us now do and make

we had all these little stacks of, okay,

you're looking at sim racing stuff,

here's a stack of that. We intentionally

didn't chase some of that stuff down

solely to prove out the pitch of Dia

that like having this chat or this

assistant living with you is going to be

valuable. If we don't make everything

else novel, it gives us future runway to

then introduce novelty in different

places. Things like bookmarks, things

like tab organization without

introducing novelty everywhere. And

then, you know, LLMs get better. we

identify usage patterns that we want to

chase down, we'd have to back out of

this novel system like we did with ARC

and have to start over again. So by

choosing the novelty and choosing to

keep it simple in specific places or

familiar in specific places, I think

gives us the best runway for, you know,

the shifting sands of what AI is right

now. I really like that reframing

because you're right, there's a lot of

lowhanging fruit that AI would be pretty

good at and ARC incorporates a lot of

that, but you were kind of saying either

we can make this horizontal rapper layer

of AI work or nothing really matters.

Yeah, totally. You know, if we if we

can't get people to find value out of

the core pitch of Dia right now, we've

either buried it or the value is not

there yet. So everything else novel

would just be like a detraction from

that. And we really look at this launch

of DIA as like the starting line for us

is kind of how we look at it internally.

It's like cool, we're at the starting

line now. We have a product back out.

It's in public. We're iterating on this

thing in public again. Where are we

going to take it? Versus like this is a

complete product. It's like no one

thinks this thing is complete. Like

there are so many rough edges in it that

like you know we still stress about as a

design team. So I think that's that's

exciting to me. It's like cool. We're

back in the race. Where are we going to

go? How'd you think about the release

line then? like when was this ready to

go out the door? I think that that adage

of like if you're not embarrassed about

a couple things when you release you've

waited too long is really true. We

picked a date and we're like okay this

is this this is the date and we're like

sometimes that's the best way to do it.

Exactly. picked the date and then some

things came in and I think this is a a

testament to the team at browser co

which like poundforpound are the most

hands down talented across the board

engineers designers everyone people I've

worked with like the skills inside of

dia the right skill code skill the

custom skills we we've we built all of

that in like less than a couple months

from design to code to production to out

and it was just like came the idea

appeared and we're like, "Yeah, that

makes sense. That's a good way of

framing it." In like less than two

months, more closer to one month, we

built that whole thing. So, we were like

very very close to the edge when we were

when we actually released this thing.

Let's talk about skills then a little

bit because I think the way that I

interpret skills is kind of going back

to the earlier question where you have

to lay the right foundation for where

the heck this is going to go. It was

very important that you picked the right

building blocks that can create a

pathway towards potentially even a more

agentic future. So how did you think

about that opportunity and ultimately

why did you land where you did? I and

and this includes DIA. So I'm not

throwing any other products under the

bus. I'm throwing Dia under the bus with

this as well. I feel like a chat input

with like an attach icon and a send

button and some like, "Hey, do you want

to do this or that that turns into a

chat stream is like the best we've got

right now and not the end state or not

the ideal way of communicating with like

LLMs. I feel like if I hand you a piece

of paper and say, "Draw me something."

You're like, "Uh, I don't know." But if

I hand you a piece of paper like draw me

a cat, you can just like immediately

pick it up and start doing something. So

you see as all these different companies

chipping at that same problem of like an

empty text box is terrifying unless you

know the power of LLMs. You're not going

to know what to put in that. So circling

back on that skills are hopefully

intended as a way of showing you some of

the things that DIA can do. DIA is

particularly good at helping you write

or particularly good at helping you

code. And by kind of branding those

things, making them feel like a tangible

unit, we hope that that helps people

understand the output of what DIA can

help you with. It still is not an ideal

entry point. The UI of just this like

chat input is like I think that's the

sweet spot that I'm still really really

intrigued to iterate on more of like how

do you clearly show someone what they

could type in this thing at the same

time without making them feel like they

have to make a decision? But which one

of these seven skills am I trying to do

right now is is kind of like the big

boss I think across the industry for

design and like AI and LLMs. I know it's

one of the hardest questions in the

industry, but like pull on that a little

bit more like where do you at least want

to point the ship and start exploring

and maybe any ideas rattling around in

your brain for what what could this look

like in the future? D is in an

interesting place because as a browser

striving for familiarity, I don't want

anyone to hit command T and have to stop

and think about what they should be

clicking on or typing in that box. Yeah.

So if we were say working on a writing

tool, the input could be very very

whittleled down to just before writing.

And maybe it's not even an input. Maybe

it's, you know, buttons or, you know,

like a fig gem like toolbar at the

bottom of some canvas helping you like

get started that like an AI assistant is

like holding your hand and guiding yours

like another cursor on the interface.

For DIA, we need to still strike that

balance of the input should still work

like a URL bar on a typical web browser,

but it should feel like you can do more

with it. Things we've done in Dia right

now is like when you switch to a

different tool, we colorize the input.

kind of poof it and you could see like

now I'm in a different mode. I think we

could pull on that thread further where

it could feel even more like you're

asking the assistant to do something

different for you. I think we could do a

better job that when we detect that what

you're typing in is likely going to be a

coding question. We could do a better

job flipping that to show you that we're

asking coding questions. Maybe there's

not one input, maybe there's multiple

inputs when you switch into those

different kind of uh forks in the road

you're going down. I think there's a lot

of a lot of different opportunities to

explore in that. Maybe we can go a

little bit deeper on the skills piece

specifically then like when you imagine

what this might unlock in the future,

what are some of the things that you

think about? Very curious about like the

agentic stuff like can I send one of

these agents off one day to like, you

know, do my grocery shopping for me

based on the things I've bought in the

past? Like it seems like that would make

sense. I'm very deep in the like MCP

server land of like how those do now let

these LLM speak speak to each other kind

of in in human language which is really

really powerful. I can see skills

continuing to get deeper and deeper

based on DIA knowing more about you. So

you can imagine if I am writing and it

sees that I am using DIA all the time to

write and I write on this website in

this way or that website in this way DIA

could continue to like learn that

context and understand that about you.

So it's given you very turnkey but very

contextually appropriate suggestions to

you and hopefully the skills stay at the

surface level pretty familiar easy to

pick up easy to like wrap your arms

around but the depth of what we can do

continues to deepen and deepen and

deepen. To me, the depth is kind of tied

to this idea of personalization. And at

the end of the day, whatever you can do

uniquely is almost always tied to like

what context can you give these models.

So, you're really deep in that rabbit

hole compared to probably the majority

of people that are listening. So, what

are some of the design considerations

that you have to start thinking through

when so much of a product's value

proposition is tied to this depth of

personalization? The memory stuff is off

by default because we want to make sure

people know that that they can opt into

this and that DIA will uh learn from

their history if they opt into that. We

also from the get-go we do not store

anything from like banking websites or

sensitive pages like we from the very

beginning we were like security

security. How do we make this give

people more power but not open up a lot

of security or open up hopefully as few

security issues as possible. So like the

data is stored locally, it's encrypted,

it's per your browser profile. So like

having that foundation then kind of put

a lot of us at ease being like okay if

it's secure, it's private by default,

that is the stance this company is

taking. Then the things you can go

really deep on in how you design for are

not overfitting I think is one thing

where it's like if I tell Dia once to

not use M dashes then I don't want Dia

every single time it responds to me to

be like because you told me to not use M

dashes. It's like it doesn't always need

to remember things forever and like it

can I think memories can fade I think is

an interesting thing and really thinking

about it as like do you should feel like

a friend that knows you really well and

a friend that knows you really well

doesn't always bring up like oh yeah

that's cuz we went to high school

together in every conversation so like

that that level interesting thinking

about which memory should fade is a very

interesting concept that I hadn't spent

a lot of time considering before and it

it's cool because so much of the design

process in that world of personalization

is completely removed from the pixels

call vibe coding what you will but it's

like we're in a very different place now

of how to build and think about products

and I think designing very much from

like an AI first lens is I think the

correct a correct path to be at least

curious about and going down because our

software can be very different now and

it's like we finally as designers have

this new putty that lets us build very

very different tools, which is really

fun. Let's pull on that a little bit

then. And hypothetically, let's say that

you were to hire somebody listening to

this who is not working at a product

that's making a lot of use of AI.

They're a little bit more green. What

are some of the AI first practices or

mental models or ways of even

approaching the craft that you would

expect them to learn on the job while

working at browser co across the board

with designers I really value just

curiosity and I think our company

browser exists because we just

repeatedly ask like what could be what

could we do and like is this possible so

I think approaching all of this towards

curiosity and also understanding that

you don't need to know everything about

AI to be able to make use of it. Like

don't get too caught up if you're just

dipping your toes in like don't get

super caught up in like all the models

and all of this and rag and all that

stuff. Like just learn it enough so that

you can know which tool to reach for

when and kind of know the bounds of what

you can do with AI from a design place.

So I'm going to make you get super

specific there for a second. Yeah. Help

someone think about where that line of

enough is. If I were to talk to someone

getting into the industry right now, I

would suggest that they through just the

pure design lens, don't look at as like

Vibe coding. Look at it as like Vibe

designing and very quickly go install

clawed code or your Vibe coding tool of

choice. Think about a problem that maybe

already exists like go redesign

Instacart. do it with AI as like your

co-pilot and see where it goes and just

stay curious with the process and ask

the AI questions while it's doing it

because I think people really really

quickly realize that the future of how

to design products I think is already

here. I think it's just not evenly

distributed yet. I think it probably is

still in the terminal for many many

people. But I think that is very very

much the right thread for people to be

pulling on. Like I spend more time in

claude code now than I do in Figma

building things and I I'm building

prototypes natively in Swift. I am an

engine I was previously an engineer by

trade. So I think I am able to like coax

it further in the directions I want. But

the time for idea to actually like be

having something to click on with the

support of AI is like compressed to

almost zero now a few minutes for many

many things. What are the types of

things that you're personally exploring

then as the design leader is spending

more time by coding? I have always been

like a very hands-on design leader where

I am I like staying close to the metal.

So I've always typically have an IC

project on the go. Right now I got

brought up to speed when Cloud Code came

up with just like an idea I had rattling

in the back of my head for a long time

to build an app to help me find a park

to go to with my kid. Uh, just because

Vancouver has open data, the data is

great, the apps are bad, Google Maps is

like this universal tool for finding

whatever you need. I just needed parks

with playgrounds and washrooms that were

close to me. So, I just chased that down

in like the evenings to figure that out.

At work, I use a lot of AI to help me

spin up interfaces that either just have

like a lot of data managing that needs

to be done, like handling streaming or

handling MCP tools. and that sort of

work which is like the code I'm getting

out of claude I would never expect to

ship to production because I'm not I'm

literally not even looking at it but

it's good enough to let me actually have

a chat stream come in that's streaming

using tools and doing that that is

streaming so I like use it to again ask

like what if what if this happened what

if when tools showed up they looked like

this but not anything going into

production right now you talked about

MCP a couple times and I'm going to

assume that probably 80 to 90% of the

people listening have an understanding

of MCP to the point of it's how LLM talk

to each other. Yeah. But it keeps coming

up and they keep seeing it when they

scroll on Twitter. So then where does

your brain go when you start thinking

about MCP and what it unlocks and what

does the nature of these explorations

with MCP even look like? I think MCP

unlocks the ability for these different

tools to talk to each other through

natural language. What I think is very

cool about that is it lets some of the

more interesting things is when like two

services can talk to each other versus

it just feels like you're talking

one-on-one with a website. It's like

well what if I can talk oneonone with

that website but that website can

actually talk to this other website in

the background. What is possible now

across these two products you use? We

don't we don't see that in many products

right now. We see a lot of that

happening at the server level, but

that's not hasn't been brought to the

client level in a really approachable or

human way yet, which I think is really

really powerful. And I think that is

obviously the agentic world loves

thinking about that. But I think it

doesn't even have to be taking actions

on your behalf. I think it could be

bringing content to you uh through this

network of connected LLMs. I'm taking a

little running tally of how many times

you're saying what is possible in this

conversation. And I think it's really

cool because it speaks to the

irrelevance of the lines between design

and code. When we all have the tools to

express ourselves at a high level of

fidelity, the only thing that matters

then is having an understanding of what

is possible. And the fact is a lot of

that is dictated by more technical

topics that do require a little bit of

an appetite for more of a traditional

designer. Yeah. I don't know what to do

with that yet. It's just like a theme

that I'm kind of noticing from this

conversation. Yeah. It's it's funny. A

few years ago, we were constantly having

that conversation on Twitter about like

should designers code etc. And like

personally speaking of from my own

personal take, I can't imagine doing

what I do if I wasn't able to at least

understand the code a bit. And like I

did work full-time as an engineer in the

past. So it's like I'm coming at this

initially as an engineer who got into

design, but there's almost no excuse now

as a designer to not code because the

tools have made the code irrelevant. So

it's like if you can't like if you're

showing wireframes nowadays, it's like

you're kind of working at the wrong

fidelity because it's so much quicker to

actually build a working prototype than

like show a wireframe and get feedback

on it. Mhm. I've been thinking a lot

about where does the line of required

technical proficiency live nowadays and

trying to get ultra specific because the

question should designers code is not

relevant anymore. No, here's my working

hypothesis and you can tell me where you

think it needs to evolve and I think it

goes back to something that you said

earlier where you talked about the

importance of asking questions while you

were trying to accomplish some kind of a

goal or outcome using you know your code

prototype tool of choice. I think it's

can you ask good enough questions to

something like cursor and then

understand the response. You don't have

to be able to read all the syntax but

you have to be able to at least at a

high level understand what is being

accomplished and how does it roughly

work and if you can do that I think you

can build almost anything that's not you

know deep on the back end 100%. And

cursor and these other tools will like

help you understand the errors that come

out of it. So even if it's spitting out

compile errors or whatnot, just have

sidecar another LLM to help you

understand the errors that are coming

out of it. So at browser co I would say

all of the designers are proficient with

prototyping. I think that goes to

different depths. Some can write native

swift and have shipped swift apps. Some

can do incredible work in origami. Some

some can do clickable prototypes in

Figma. But across the board, everyone is

always prototyping and always showing

prototypes. We are very rarely are

working in static mocks aside from brand

work. Brand work it's very different but

then as soon as you put together a

website like our deal website is like

that will that's hard to express how

that footer comes up unless

stuff. So it's like everything happens

in motion. I'm I'm very bullish on

designers learning more about coding. I

think vibe coding is a fantastic new

world for to be a designer. It's like

such a fun fun world now where you can

build anything you think of. And now the

bar is what can you think of? I'm

definitely having more fun right now as

a designer than at any point in my

career by far. 100% 100%. All right, so

let's talk about the culture a little

bit. You talked about this idea of like,

okay, design in motion, lots of

prototyping. How do those prototypes get

shared? How do you all collaborate? How

do you all even figure out which of the

many prototypes to prioritize and

explore further and ship? Can you just

share a bit about how design operates in

practice at the browser company? I think

we're around 90 people as a company.

There's seven designers included myself.

We have more recently more of a PM

function as well. And I think that has

really been a good unlock for us. A lot

of the team is remote. So we have kind

of like a gravity in New York, but like

Slack is our Slack and Notion are like

our central kind of central hub. Slack

notion in linear the culture inside of

browser co is to be always dog fooding

things. So like the dia version of dia

we have is very different than the

version of dia that is out in public

because we are always we always have

experiments running inside of that thing

through our feature flags. We've built a

culture where it's very very natural for

everyone at the company regardless of

their role to give any type of feedback.

So we have people that are not trained

designers giving feedback at a pixel

level on things just based on what they

feel about their subjective take or

their objective takes on designs and

that's happening all day long. You

scroll through dog food and it's like

you can almost like not keep up with it.

It can be kind of stressful as a

designer. I think when you first drop

into that environment because I I have

never seen or worked anywhere in my

career that gives so much constant

feedback from all kind of all angles.

And as a designer, you just need to

learn are these personal preferences?

What is kind of the motive behind the

feedback? Is it because they are using a

giant monitor and this is not working on

their big monitor? Because they're a

mouse user and everyone has been

thinking about this with trackpads.

really really distilling that and

understanding not what to listen to and

what to ignore because I think you

should listen to everything but what to

prioritize and what to not prioritize. I

imagine you have you're a designer,

you're the design team, you have lots of

prototypes being built, lots of ideas

being explored and you have this fire

hose of feedback coming at you

internally and this fire hose of

feedback coming at you externally. I

think what we do is we try to have

really strong opinions on the decisions

we make and the less product you have,

the less surface area of the product you

have, I think the stronger opinions you

need to have because it's way easier

just to keep adding stuff. I think if

anything, art was a failure of us not

having strong enough conviction in our

decisions. So, we just kept adding

things to it where the amount of stuff

we now choose to not put in needs to be

grounded in I wouldn't say like

principles. We don't have like written

down principles per se, but very very

strong conviction in the direction we

want to go. How do you put that into

practice as a leader? When I share my

opinions internally or I share my

feedback on projects, I always try to

ground it in why I'm giving that

feedback so that at least I have

internal there's internal consistency in

my feedback. So it's not saying like I

don't like that. It's like I think this

could be better because of this and I'll

share always the thinking behind that.

So as a leader at least I'm internally

consistent and then as a design team we

will like say we have someone who spikes

higher on visual design I will always

try to pull that person into the

conversations to be like hey is this

matching the visual design direction

that you wanted to push Dia in and we'll

just kind of like naturally and

organically uh pull people into the fold

for those conversations. Same with like

motion. All designers are so different

and spike in different ways that

depending on the problem in front of us,

I'll make sure the people that we're all

talking collectively and that I will

also make sure as a leader to like make

sure the right voices are getting the

chance to speak and share their

thoughts. I'm going to come back to the

visuals piece. So, I'm going to put a

pin in that. Cool. Because I want to

pull on the external feedback strand for

a second. And given the context that I

actually remember talking to Nate about

this when you all were first working on

ARC and he he brought up like hey we dog

food the heck out of this we prototype

it we're our own end users we're

designing for ourselves you know and the

fact is for DIA you can't really do that

to the same extent I'm sure right

because you have to intentionally learn

from people who are not very techsavvy

so how did you ensure that you were

getting outside of yourselves especially

early on and were there specific things

that you were trying to learn and I'm

even thinking about how you all did a go

to market strategy with college students

for instance how much did that change

the way that you learn in the early days

of the new product I still feel like

we're designing our product for

ourselves we all everyone at the company

still uses a web browser all day long

for different differing degrees if

you're a Slack user you're in the app,

you're maybe not in DIA as much

throughout the day as someone that lives

with all of their web tools, notion,

everything inside of the product. I

think the work in front of us and the

work that we've done better in DIA than

I think we did with ARC was choosing

where to hide the complexity and then

understanding and being more intentional

about the audience that's going to seek

out that complexity and find it. I'll

give a concrete example with custom

skills. So in ARC you can make a boost,

you get a remote control and you can zap

stuff off the web pages. It's like

that's cool. That's awesome. We did

everything every time we shipped

something in Arc. We were always trying

to figure out how do we help people find

this thing? How do we always bring it to

the surface of ARC and like you saved

twitter.com for me. Exactly. And there's

like there's not that much real estate

and you can't have that many top level

buttons that make sense in a browser or

that are understandable in DIA are

custom skills which I think are like I

think they're really neat and for those

who haven't played with them before you

can give dia slash commands. You can do

slashcooking you can give that when you

type that command you can then have a

canned prompt that runs. So anytime you

type slashcooking, you can have it

search your favorite websites, convert

the recipes for a family of two versus a

family of four, just all all on the go.

That's a really specific audience that

is going to understand a slash command,

understand a prompt, and then actually

be able to get value out of that empty

text box. In Arc, we would have put a

giant button on the surface that said

like make a custom skill. In DIA, you

have to hover on a message and click

save as skill. previous Ark world, we

would have been like, well, no one will

ever find that. That is not going to

bring anyone value. And but then

understanding that the curious people

that will do that, we'll we'll seek that

out and we'll find it. Then we'll learn

how to use it is kind of we've done a

better job meeting people where they

are. Okay. Let's go back to the visual

piece for a second. Yeah. Were there

intentional ways that you wanted to

evolve the either UI or the way that you

experience the brand inside of the

product as a company and for me just as

a as a individual designer as well

because my personal design aesthetic is

like very plain no color content first

that's it and then I had worked on Ark

for a long time that is this like bright

pink and yellow with gradients and

smiley faces and Every space is a

different color and you have the noise

and Exactly. And like when you could

click a gene texture, you could add

textures to this thing which is like so

far removed from my personal aesthetic

that it was always this like level of

dissonance that I was just like working

through. And I was often surprised by

when I pushed myself out of my comfort

zone with design like I designed a lot

of the theme stuff that it would

resonate with people. I'm like that's

it's wild to me that you want your web

browser to look like this. Got it. But

I'll design a tool for you that can look

like that. With DIA, we intentionally

chose an aesthetic that would get more

out of the way in order to I think not

push people away. I think a clear

example of this is like imagine you are

some top level exec at some finance

company and you and your team have

worked on all this stuff and you're

about to present present it to your boss

and all the stakeholders and you open up

arc it's bright pink there's a winking

smiley face and you're just like uh yeah

like I promise like this is like not

juvenile like what I'm about to present

to you. It's like we by going down that

design path, we narrow the audience that

is interested in the product. With DIA,

we took a much more intentional like the

product should sit in the background,

let the content come forward and let

whatever you're looking at shape what

the what the UI looks like. So, you

know, if you're looking at a bright

yellow website, more the browser is

bright yellow because that's reflective

of what you're actually looking at. We

still let you pick a theme, but it's

much more subtle in DIA. And we will

forever play with these things. We might

ramp it up more. We might give you more

options, but out of the gate, we want it

to make it feel more appropriate and

more context. I'm a big believer in the

power of video to explain my thinking as

a designer. So, when it's time to give

feedback, I'll drop a Loom link in Slack

and another link to a Figma prototype

and then feedback will be scattered

everywhere. And I mean, it's a mess. So,

I'm building the product that I've

always wanted to exist, and it's called

Inflight. You can kind of think of it

like an Async Crit. It's an easy way to

share a video walkthrough along with an

interactive prototype or whatever you're

designing, and then AI interviews the

people on your team to get you the

feedback that you need and organizes

everything for you in a beautiful

insights page. So, right now, I'm only

giving access to Dive Club listeners.

So, if you want to be one of the first

to use Inflight, head to

dive.club/inflight

to claim your spot. It's interesting to

juxtapose the two browsers as you're

talking where I guess personalization

actually always has been at the heart of

everything. It was just for Arc it was

more visually represented whereas now

everything is kind of happening behind

the scenes. Yeah, I can give a little

story of like where themes came from in

Arc. We were getting ready to ship it

and just got a lot of push that like,

hey, it's just gray. this browser is

gray. It's like, yeah, I think it should

be gray. I want a gray browser. It's

like, no, it's like we need to make this

thing have more color. It needs to stand

out. All browsers are gray. Make this

thing stand out more. Stewed on that a

lot. It's like, yeah, we are new to the

market and we do need to stand out, but

not everyone is going to want a purple

web browser. What did we do? So, where

we landed was, can we let people pick

the colors they want, but also let the

colors serve a purpose. So, in ARC, it

was like a wayfinding exercise. We're

introducing spaces. Let person pick a

different color for every space. So then

immediately you're like, I'm in my

purple space. That's work. I'm in my

blue space. That's personal. Done. But

then let you not have to pick a theme

because we're not going to let you're

not going to find a theme that fits for

everyone. I think in DIA, we turned down

the gas from like 300% to like 10%. So

now when you pick the theme, it's less

like taking over the whole browser

because it it it likely not everyone

wants that. Okay. Okay, I want to zoom

out again and talk about the design

culture as a whole. Yeah, you mentioned

a little bit about like the prototyping,

design in motion, that kind of a thing.

I think I'm going to toss a hypothetical

your way as a way to kind of get at this

topic a little bit more, which is let's

say something happens and you realize,

you know what, I'm going to go lead a

different design or in 2026. Y how has

your almost 5 years at the browser

company influenced your goals for what

you would want to instill in that new

design culture? This is going to sound

probably obvious and trit, but I think

really even more so now than ever

before,

focusing on the problems you're trying

to solve for someone or the jobs that

person is trying to do because even more

so than ever in the past, it is easier

and easier to design and build stuff. So

then you're just going to have this

forever increasing

platter of ideas in front of you that

you're going to need to choose from. So,

I think if I were to go somewhere new, I

would very intentionally have those

conversations, interrogate the process

to be like, are we actually getting to

the crux of the problem we're trying to

solve, or are we just like building

stuff? And it's fun to build stuff, but

I like building things that people use

and that help people. So, and I think in

the future where you can immediately

build anything you can dream of, it's

really really healthy to like know what

problem you're chasing down and how

you're going to know when you've solved

that problem, what signals if we put

this into the world, we think we're

solving this problem. How are we going

to know when we actually do solve that

problem or not? And being very

intentional about that cuz I think

design will forever move up the stack

further. And I think to become better

and better at design, it's like helping

helping cast design in that at that

light. That would have been an

incredibly boring answer two years ago.

It's true. Yeah. But it's really

relevant. Like you're talking and I'm

like, man, I feel that. I feel the

intention because it's not just the ease

of software creation. It's the fact that

you can just plug any model in and it

can accomplish a million things out of

the box. And so it's so easy to be like,

oh, what if it also could do this and

this and this and then it's like

goodness gracious. I mean, I've been

working on a project for 6 months and I

already feel that temptation. You go

like movie equipment. Not that many

years ago, it was too expensive and out

of many people's hands to make a movie

that looked like a feature film. But now

for like the cost of a Honda Civic, I

can go buy enough equipment to look

exactly or very very close to a feature

film. Because that has happened, you

don't see all these feature films being

made now that are exceptional films. You

still need a good idea and you still

need to have that vision. The tools have

gotten better. the tools are widely

available now, but it's still what

you're actually trying to say and how

what story you're actually going to

tell, what in our world, what product

you're going to build is all that what

is still matters. And I think we're just

seeing that in the design world now

where the tools are how anyone can

design anything they want. Now the

robots can build it for us. The problems

people have are never going to go away.

When you kind of imagine these futures,

what are some of the ways that you see

it shifting the value prop of design

within an org and even what matters when

somebody listening to this who's maybe a

few years in their career when they're

thinking about how do I invest in

myself? I go on Reddit all the time.

It's like the number one thing on the UX

design subreddit is always like what the

heck do I do? I'm motivated. I want to

do something. What do I invest in? If I

were getting into the industry now, I

would really really focus on solving

actual problems and have a portfolio of

solving problems. And I think a clear

example of that is like it is really

easy to say focus on like fun micro

interactions and tweet them out and get

lots of likes. But like that is like

such a it's called micro. Think about it

as like you've solved a micro problem.

So, it's like if you're trying to get a

job and you're showing the world you can

sign solve little tiny problems, like

that's not that's not helpful. And it's

like a hiring person, I would much

rather see someone be like, "Hey, I went

to like the local food bank and I

interviewed them for a week and figured

what and I built this software for them

and look, they used to struggle with all

this stuff and now this software has

helped them with all of that stuff." It

it's at the end of the day, we we we

have to build things that solve problems

for humans and help humans along. So I

would chase that stuff down and fill my

portfolio with like actually solving

real problems for real people and

explore the micro interactions. That

stuff is fantastic. It's cool. I love

seeing it. We we do it in DIA as well.

But like the industry is going to shift

further and further away from that, I

believe. And I think the value is going

to be higher and higher and higher. Can

you pattern match across human problems

that people are having? Choose the right

tool and choose the right design to

solve that problem. And I'd be really

curious to see people chase that stuff

down out of school. I love that answer

because it's also the best way to learn

too. Like you talked about the

importance of just getting into the

tools and immersing yourself and it's

very very difficult to do when you don't

have something that you're pointing at

and some definition of success. But when

you have like an outcome that you want

to bring into the world, man, you're way

more motivated to push past that stupid

bug that is not going away, you know?

Totally. Totally. And like your

motivation shouldn't be like likes and

stuff on the Twitter. It should be

really really understanding how

rewarding it is to see a problem a

unique problem that no one has solved

before and solving it is like it's it's

the dream and I think that is that's a

skill I think more and more people

should be should be trying to work on.

I'm sure there's at least somebody

listening who's kind of wondering how

much that type of investment would steal

from time spent, you know, refining the

quote unquote craft and getting into the

visual design. So, do you have any take

on that balance, especially as someone

who, you know, you're evaluating inbound

requests to work at the browser company?

The unfortunate answer is I think it all

matters. I think it all matters. I think

taste matters. I think the ability to

identify and you know internalize what

to you feels good or high craft is

valuable. I think I would give myself

permission if I were younger Dustin to

like understand that there will be a gap

between what I value view as good design

and my capabilities of getting there.

And I think giving myself permission to

be okay with that and know like hey I

see this stuff I love I can't for some

reason my stuff doesn't look like that.

being okay with that and identifying

that and even being able to like

articulate that because I think

reflecting on that could help you bridge

that gap. So I think there is a world

and I have made a career and again this

is just my perspective out of being a

mashup of many things. I'm not the best

engineer. I'm not the best visual

designer. I'm not the best product

designer. I can just do a medley of a

lot of it at okay different levels which

then seems to bring value to the people

I work with and the companies I work at.

So, but I've I know that and I know that

about myself. So, when you're chasing

down jobs and looking at jobs, know

where you spike and where you don't

spike and then just show up honestly and

candidly at that level because I think

it'd be unfortunate to get into a

situation where you've sold yourself as

something that like you don't actually

think you can do or sold yourself as

something that's actually not

interesting to you. So, there's many,

many ways into the design world. And I

think just knowing where you want to go

and then giving yourself permission to

be okay with your skill level, I think

is really important. I'm going to give

you permission to brag for a second

because you're describing yourself as

this mish mash of skills. Maybe you

don't spike incredibly high on any one

skill. I think I probably would say

similar about myself. And yet, you know,

you've had a lot of success. you've been

in these premier roles and people would

look at and be like, "Okay, yeah, you

know, Dustin's got something figured

out, you know." So, when you reflect on

your journey, what do you think led to

that? Like, was there a clear inflection

point or a set of skills that you have

refined or anything that you think that

other designers listening could draw

from and use as they think about their

own career journey? I don't know. And I

think a lot of it was probably luck and

I think a lot of it was just me being in

the right place at the right time and

having the opportunities that were in

front of me. You're not doing a very

good job of bragging.

I think one thing that is true if I look

back at say like different inflection

points in my career, it is that I often

times took on things that were out of my

comfort zone but I was really curious

about.

I I'll share a couple examples. I worked

at a small web shop as a web developer

way way back when, i.e. six days. And

Club Penguin was a flash game company in

my hometown that was being worked on.

And like I had built one thing in Flash

before and I wanted to be a Flash

developer. So I applied at Club Penguin

as a graphic designer and I got the

graphic designer job and I was like

cool, I want to be a Flash developer. So

then I just like found the lead flash

developer and just like talked to him a

bunch and just like put myself in that

situation and just like tried to learn

from him and then went home and tried to

build flash stuff. I think just being

curious about how to do it and putting

yourself out there and in those

situations had led me into like good

situations. Same thing happened when I

applied at Medium. I was living in up in

Canada. I wanted to work down in the

States. I was a huge fan of Twitter and

Ev Williams. He said he was tweeted out

that he was starting Obvious Corp. I was

working as a software engineer. I assume

there are a lot of software engineers in

San Francisco and maybe not as many

designers. I applied as a designer even

though I was a software engineer cuz I

was like I'm just going to try to get

out there and it turned out he liked my

work and we we had a good chat and I

ended up working at Medium. So I put

myself out there a lot. I have had a

tremendous string of mentors in my life

that I am very very thankful for that

have like helped guide me and have given

me feedback. And I think I've tried to

not be an [ __ ] I've tried to just be

a good person throughout my career and

just try to realize that like it takes

many people to build all of these things

and paying it forward is really really

good. So I think you some things I think

you kind of get what you sew. So I think

if you're trying to be a good person,

trying to be curious, I hope people will

be curious about giving you

opportunities as well. Curiosity comes

up just about as much as any word on

this show right now. And it's like it

feels like we are transitioning into a

whole other era of this industry. And in

many ways, curiosity is the almost

parachute that allows you to make the

jump between eras, you know, and doesn't

really matter if you're 15 years into

your career. That's I've been designing

products for 15 years, but if I have no

curiosity towards what's coming, I'll

get dusted by people who are a couple

years in. It's true. It's just like I

lead the design team here. I have a

young daughter. I'm trying to be a good

father and partner. And it's just like,

cool, AI is here now. It's like, oh, I

guess I got to go learn AI now. And it's

just like you just have to keep doing

that or like you said, you will get

dusted. And it's like no one is going to

give you a curriculum of the things you

need to learn. Especially if you're

already in a role, you got to figure out

how to do that or else if you want to

not continue to move forward with that

stuff, that's fine. But for me, it's

like I would just want to keep getting

better at my craft. So, I will continue

to chase this stuff down. I love it.

Well, Dustin, thanks for coming on, man.

This has been really, really enjoyable.

It's cool to hear all of the little

decisions and things that you were

wrestling with while you were bringing

Dia to life. Congratulations on getting

out the door. By the way, it's been

really amazing to just see how warm the

reception has been and you all have

definitely deserved it. Incredible.

Thanks for saying that. I also wanted to

make sure this was the work of many many

hands and very a group of very very

talented people. So, I do not deserve

all the credit by any means. So, I don't

think you're going to be the last

browser co designer that we bring on.

So, we'll do a good job of spreading it

around. Perfect. Before I let you go, I

want to take just one minute to run you

through my favorite products because I'm

constantly asked what's in my stack.

Framer is how I build websites. Genway

is how I do research. Granola is how I

take notes during crit. Jitter is how I

animate my designs. Lovable is how I

build my ideas in code. Mobin is how I

find design inspiration. Paper is how I

design like a creative. and Raycast is

my shortcut every step of the way. Now,

I've hand selected these companies so

that I can do these episodes full-time.

So, by far the number one way to support

the show is to check them out. You can

find the full list at

dive.comclub/partners.

[Music]