The Future of Linux: Age Verification, AI, Privacy, ARM Support, and More (System76 CEO Interview)

System76 creates premium Linux computers, but they also helped draft the open source exemption language for Colorado's age attestation bill. In this interview, System76 CEO Carl Richell explains how it happened, why open source is different, and what's next for Pop!_OS and COSMIC.

This episode is sponsored by SLNT. Check out their faraday bags and get 15% off + free shipping on qualifying orders: SLNT.com/TECHLORE

🔗 SOURCES & LINKS
• System76: https://system76.com

⏱️ TIMESTAMPS

• (00:00) - INTRO
• (03:52) - PRODUCT TIMELINE & EXPANSION
• (08:45) - SYSTEM76 CUSTOMERS
• (10:24) - SYSTEM76 PRODUCT LINEUP
• (12:24) - POLITICAL INVOLVEMENT
• (18:58) - SYSTEMD PULL REQUEST
• (19:52) - SPONSOR: SLNT
• (20:47) - HOW TO GET INVOLVED
• (24:44) - PROPRIETARY VS. OPEN SOURCE
• (27:51) - TOP TO BOTTOM INTEGRATION
• (30:21) - LTT LINUX CHALLENGE
• (33:47) - WHAT IS LINUX STILL LACKING?
• (35:50) - LINUX PHONES
• (38:41) - ARM SUPPORT
• (40:28) - AI
• (44:28) - MEMORY PRICES
• (45:00) - FUNDING
• (45:32) - WHAT'S NEXT?
• (47:05) - PICK A DISTRO
• (47:56) - HOW TO FOLLOW

🎥 VIDEO
Watch on YouTube

🧡 SUPPORT TECHLORE
Keep Techlore Talks independent & growing: ★ Support this podcast ★

Creators and Guests

Host

Henry Fisher

Runner, artist, musician and digital rights activist. Owner of Techlore

Guest

Carl Richell

System76

Editor

Tori

Techlore

What is Techlore Talks?

Techlore Talks brings you in-depth conversations with the experts at the forefront of privacy, security, and digital rights. Hosted by Henry Fisher, founder of Techlore and long-time digital rights educator, each episode features meaningful discussions with the people building, researching, and advocating for digital freedom.

From cybersecurity researchers and privacy tool developers to open-source advocates and digital rights activists—if they're shaping how we protect ourselves online, they're on this show.

Topics include: privacy tools and technologies, cybersecurity threats and defenses, open-source software, surveillance and digital rights, encryption, tech policy, and digital sovereignty.

New episodes released regularly. Subscribe and join the community at techlore.tech.

There are companies

that sell age verification services.

They're selling this as child protection.

But what they're really trying

to do is sell a product.

Today, I am absolutely honored

to have Carl from System76,

the Linux laptop manufacturer,

to discuss a lot of interesting things,

like, of course, the origins of System76,

what they do, how they do it,

what separates them from other options,

but also a lot of really interesting

developments, like California's

age attestation law in Colorado's,

and how System76 was actually a very big

part of the movement that was able

to get open source excluded.

So I asked Carl about what this looks

like behind the scenes, how you all

can get involved, and how we can overall

fight back against a lot of these

initiatives to keep free software free.

Without further ado,

let's get into the interview.

I'm Carl Richell, the CEO of System76.

And is a computer manufacturer

that specializes in Linux computers.

In the world of tech, we're kind of old.

We're at 20.

This year will be, November

will be 21 years.

So I was 25 when I started System76.

And I started in my basement with $1,500.

And the dream to make Linux

more accessible to people.

At the time, I think it

was, I don't remember what version

of Windows it was at that time.

I felt that Linux was a fantastic

desktop at that time.

I think we beat each other and ourselves

up quite a bit in the open source world,

but we make great software.

The bugs I run across in Windows

and iOS and Android and Mac,

and I think we have really high,

maybe we're a discerning group of people.

But the Linux desktop has been fantastic

for a really long time.

I didn't feel like it was really

well represented by hardware OEMs.

They were shipping servers.

They weren't shipping laptops.

They weren't shipping desktops.

I thought it deserved a dedicated company

that didn't ship Windows,

that shipped Linux because it was better.

And so that's why System76 was started.

Since then, we've just grown

organically over the years.

We're not really the typical Silicon

Valley, VC startup kind of company.

We build products and we support them

and we support our customers' sales needs,

helping them determine

which products they need.

We've grown just from doing that really,

really well across the company

and reinvesting in what we do up

to the point where today we have a factory

in Denver that we built from scratch.

And we bring in raw materials,

cut, shape, paint, add fasteners,

all of the work that goes into that raw

material becoming a computer

that ships out at the end of the week.

We build a Linux distribution, Pop!_OS.

We build an entire desktop

environment with Cosmic.

In a lot of ways, it's surreal

to look back on because we've become

this really unique company that designs

and manufactures hardware, that develops

an open source operating system,

that develops the entire experience that

the user has with the desktop environment.

And, you know, applied all

of that experience over 20 years

to building this experience

that we can now deliver.

And at this point,

I think it's along with what's happened

in the rest of the open source community,

we're delivering this elegant product

that is not just for, you know, us Linux

geeks and people that love computers,

but really a great product for everyone.

So that's me and System76.

Nice.

Yeah.

And I feel like to some people, especially

if they're newer to the Linux world,

they might have heard of System76

from a hardware perspective.

They might have even heard

from Pop!_OS or Cosmic.

But I don't know the percentage of people

who know they're all from the same team.

I know there's definitely a good number

of people who do know that, but I feel

like it's hard to dabble in those three

different places at the same time.

So do you want to expand a bit

on the order of operations?

Was it hardware first,

then Pop!_OS, then Cosmic?

What was the rationale

behind starting each one,

and what were kind of the timelines?

Right.

So it was starting with hardware

and we shipped Ubuntu.

We picked Ubuntu well before it kind

of started taking off.

We just believed in the ethos and the

project and the technical excellence.

And I think their focus on the desktop

at that time, in particular, over server.

That's what Mark really wanted to achieve.

Around, I think it was 2017,

Canonical was building

their Unity interface.

They started with building extensions

to GNOME to experiment

with their user experience, so the user

experience they wanted to deliver.

And then they decided to build

a desktop environment from scratch, and

that was going to be Unity 7, I believe.

What happened about 2017 is that Canonical

decided not to continue the Unity project.

I think they had really

good reasons for that.

And at that time, because Unity wasn't

coming, we decided this would be a good

time for us to respond to our customers'

needs that we know exist.

And that's the impetus between each one

of the decisions that we made over time.

So we had customers who were frustrated

with the number of notifications they

were getting from their desktop.

They were frustrated with the aesthetic.

They weren't able to do full disk

encryption out of the box from an OEM.

Because you couldn't really do

that without the OEM having the keys.

And if the OEM has the keys,

then that's not full disk encryption.

That's actually private.

So that means writing an installer

that can have that functionality.

So we built Pop!_OS

to address those things

that we knew mattered to our customers.

That resonated with our customers

and the broader community.

What's unique about System76

is that we have a huge hardware lab.

At that point, we had already been

shipping hardware for 12 years,

and we always have the newest NVIDIA GPUs.

We always have the newest

chipsets, CPUs, AMD GPUs.

So that means Pop!_OS had uniquely good

hardware support across the full stack.

So I think with the NVIDIA support

that listed a really tight feedback

loop between our tech support team

that's talking to our customers,

our sales team was talking to customers,

we were able to build a distro

that really resonated with people.

The next big step was Cosmic.

It was the exact same thing

as starting with Pop.

We knew what our customers wanted.

Bringing in pieces from different places

and trying to pull them together wasn't

really addressing our customers' needs.

And so we started designing desktops

from the ground up, understanding

where the things that we didn't

like about what we were capable

of doing before that.

Things like aluminum is very soft.

And so when a chassis manufacturer puts

threads into aluminum, you will only get

so many insertions and removals

from a screw because it's just soft

and that's what's going to happen.

So we use steel PEM inserts instead,

and those threads never wear up.

So the longevity of the product is better.

The serviceability

and upgradability is better, you know,

thanks to those kind of decisions.

We use a thicker aluminum because most

companies use one millimeter aluminum.

Your side panels feel flimsy

when you pull them off.

You can see the bend.

It doesn't feel good.

And so we use a thicker aluminum that

gives better durability and structure.

We like the aesthetic of anodization,

but it's kind of nasty

for the environment,

and so we didn't want to do that.

We use powder coating

that doesn't have any VOCs.

It's much, much cleaner

for the environment.

As we move towards laptops,

we need higher tolerances than anything,

so we're experimenting with other types

of finishes for laptops because powder

coating can be a little bit thick.

It's okay for a desktop,

but it wouldn't work for a laptop.

The case that we were purchasing,

the power button, if you pushed it

hard enough, it would pop through.

And this was a really high-end case.

And we're telling the manufacturer,

we think there's a problem here.

And they wanted like three months

of evidence and all of these things

just to make this manufacturing change.

And we're like, we can't do this anymore.

And so the power buttons

on all of our desktops,

they're these really high quality.

They're like the only thing

you touch on a desktop.

Almost.

You touch it the most often for sure.

It's got to be great.

It has this wonderful tactile feel

and click to it.

It's robust.

It's never going to pop through or break.

It's just a quality that...

Don't challenge people.

Yeah.

I would be impressed.

I would seriously be impressed.

But it's the kind

of quality we really wanted to deliver.

You know, you talk about customers and you

designing these different products.

And I want to ask you a little bit

about your lineup and then kind

of how it compares

to other Linux devices in a second here.

But who do you overall define

as your customers?

Like what has overall been

the characteristics of that?

Are they like more beginner friendly?

Do you see a wide range

of expertise levels?

Is it people's first dabble into Linux?

Yeah, I'm just curious,

like what you feel like your audience

is and what they resonate with.

I think for a lot in the open

source community, we kind of think

that Linux is a narrow group of people.

It's easy to overlay ourselves onto

who we think uses Linux, but it's actually

much broader than people think.

One of the stories I like to tell

about this is we survey

every customer for feedback.

We get the survey response back that says,

I love my new laptop.

It's fantastic.

I'm a developer, but this is too geeky.

It's never going to be mainstream.

And the very next survey that came

in was a gentleman who said,

I'm 75 years old

and I just got this desktop.

I'm retired.

I absolutely love it.

I think it'll be the last

desktop I'll ever need.

I understand the perception.

But when people actually,

it kind of goes back to what I started

with, why we started System76.

Desktop Linux is much, much better

than I think we give ourselves credit for.

That's true.

Especially now in 2026.

Yeah, and we tend to think

that it's kind of exclusive in a way.

But it really isn't.

And the customers that we have, they do,

certainly they lean more towards engineers

and developers and kind

of the first mover tech crowd.

But it really is a broad group of people

that go from all different places

and do all sorts of different things.

Very nice.

If you just had to break down your overall

product lineup in a few sentences,

what kind of overall desktop

options do you have versus laptop?

And do you have any other products

besides those two things?

So we do offer laptops,

desktops, and servers.

What we try to do is offer Linux

in a way that's very broad

for any kind of use case.

And they're incredibly powerful,

incredibly versatile.

So it's everything from the type

of product that just like a laptop

that needs really long battery life

and portability to kind of crazy things,

like a laptop that can have 192 gigs

of memory and 5090 GPU because they need

to load models into the 20,

I think it's 24 gigs on the RTX.

And they need to be able

to move that around.

We have one customer that does MRI,

builds MRI machines.

And when they're going to customers,

they need a powerful laptop

for processing that MRI in real time

in front of them as they're showing

them what their software can do.

Desktops, it's the same,

and kind of everything in between.

For some folks, you want a really

long battery life, and you want

to use a GPU sometimes for gaming.

So we have a lot of laptops

with hybrid graphics, and we put a lot

of care into the software work

to ensure hybrid graphics works well,

so the NVIDIA GPU is entirely turned off,

the longest battery life possible,

until you start a game,

and then it'll run on the NVIDIA GPU.

So those kind of things that optimize

the user's experience and the customer's

experience when they're looking

for something that's more balanced,

rather than ultra portable

or extreme performance.

Same for the desktop where we have

small desktops that look beautiful

and fit on the top of the desk.

And we have huge desktops that fit three

GPUs or 1600 watts and are designed

to go into a lab that's tuning or doing

inference or other kind of work in AI

or data modeling and those kind of fields.

Got it, yeah.

And the real middle thing I wanted

to discuss with you today

is definitely, it's weird.

I feel like maybe back

when you started System76,

you wouldn't have really guessed there

would be a huge political angle to Linux.

So, and it's not really political

in the sense of like left versus right,

like we normally see in different issues.

But this whole movement slash wave of age

attestation, age verification, et cetera,

that California proposed, I know

that things overall were toned down,

but I'd love to just overall hear

from you because on my end,

I was trying to follow the news.

I'm based out of California.

And so I really wanted

to make sure we were staying on top of it.

And I kept seeing you kind of being cited

as talking with different politicians.

And you were being attributed as a big

reason for why Linux was actually,

and open source was ultimately excluded

from these bills in some capacity.

I don't know if you feel like that's true

or not, but that's at least

how coverage was presenting some of it.

So I'd love to just hear

from you what happened behind the scenes.

How did you guys get involved?

Who did you talk to?

Of course, what you're able

to share just to kind of start the story.

I don't even know where to start

because that's literally

all I know going into this.

There's two things I don't

particularly enjoy, but that doesn't

mean they aren't important.

I don't like lobbying

and I don't like wearing a suit.

And I had to do both

of those for this one.

But it was really important.

And so I'm glad that we put the effort in.

So there's two things happening.

One is that there are a lot of, there

are companies that sell age verification

services and they are lobbying

and spending enormous amounts

to try to pass age verification,

meaning not just you say what your age

is, but your age is verified

through an ID or facial recognition.

It's all absurd, frankly, and we

can get into more detail, but they're

selling this as child protection.

But what they're really trying

to do is sell a product.

And when they think this huge installed

base of computers, legislatively, if those

operating systems are required to use

age verification, then they are required

to use these companies' products.

And they have a huge

market all of a sudden.

I've been in testimony

with those companies, or provided

testimony during hearings where those

companies are also providing testimony.

And it just could make you sick listening

to these arguments that are nonsense.

And you can see that I had the same thing

with right to repair

when we were advocating

and testifying for right to repair,

where wheelchair manufacturers and tractor

shops didn't want right to repair.

And their arguments are mostly

people are too dumb to do it.

And it's absurd.

But on age verification, when we heard

about it, we decided it was really

important for the open source community

to make the best effort we could

to avoid open source being,

or the law being applied to open source.

For lots of reasons.

One is we strongly believe the operating

system should not be a source

of demographic information.

And it's easy to combine bits

of demographic information

to determine who a person is, and then

you've identified a person,

and that has violated their privacy.

So when we first heard about it,

we wrote about it.

Senator Ball in Colorado had seen and read

what we wrote about age attestation

when I wrote the article

about our position.

And he wanted to talk about it because he

couldn't, he had trouble finding someone

in an open source community to talk to.

And, you know, an understanding challenge.

I think we in the community,

we kind of tend to think

of maybe politicians and legislative

process being very forward and far away.

But I really enjoyed working

with Senator Ball because he did listen

to what we're doing.

So he asked me to come in for a meeting.

And the moment I sat down, he said,

I have all these people coming

in and asking for carve outs.

I'm like, oh, boy, that's the start.

Because here I am to ask for a carve-out.

But then we get to talking.

The first question he asked me was, tell

me about your operating system

and what you charge for it.

And that was a wonderful opening

because I got to talk about what open

source is, how it empowers people

to use it to create their own things,

what an innovative driver it

is, how it's entirely free, how its

characteristics are the complete opposite

of something like Windows or macOS.

And by the end of the meeting,

he said, well, I think we should work

on exception for open source.

And that was fantastic.

They invited us

in for drafting the language.

We actually drafted a language

that made it into the bill.

So we had another stakeholder meeting with

more people in the open source community.

And the rest of the authors or co-authors

of the bill, we talked through it.

They were all in agreement with why open

source has different characteristics

and this type of law wouldn't apply.

And this was age attestation.

It's where you say what your age is.

You don't have to put in an ID.

It's not verified by anyone.

It is, I think, less concerning

than verification.

It's still concerning

because there are ways to take a bracket

and query that bracket and see

when that bracket changes

and then know someone's age, for instance.

So just try to avoid those as you can.

The strategy was to be successful

in Colorado, getting an exception in.

Use that as an example for California.

And much to our delight,

I was working with Red Hat and lobbyists

in California who had direct meetings

with the politicians there.

And we were able to get the exact

same language in the California bill

that exempted open source.

So it was a huge win,

but the fight isn't over.

There's a law in Illinois now

that has no exemptions whatsoever.

I've heard that there's a law being passed

around in Congress as well, but I have

not been able to see a copy of it.

And so I think if Colorado and California

could be the example nationally,

then maybe we can get something done here

that would circumvent the more abusive,

abuse was not the right word, but the, you

know, the, uh, Egregious, maybe?

Yeah, there's the laws

that would be more invasive to privacy.

That's the strategy and what happened.

Nice.

Yeah, no, I guess the coverage

is pretty fitting then.

I guess I have definitely you guys,

a big portion of you guys to thank

because I'm based out of California.

And so it's awesome to be able

to still boot up operating systems

and not have to put in an age bracket.

And actually, speaking of this age

bracketing scheme, there was the whole,

I don't know if I want to call it

a controversy with systemd and how systemd

had the pull request to add in some basic,

I don't know, would you call it metadata

or just device identifiers

that might have had to do with age?

And I saw very technical people

say it wasn't a big deal.

I saw some people say it is.

So I'm curious how that plays

into this, if at all,

or if that's something different.

And if you guys have had

to navigate that as System76.

The way I think of it,

it's a field for a birth date,

but nobody has to use it.

Pretty simple.

We have no intention

of using it in Pop!_OS.

If a distro wants to, it's fine,

but anybody can do what they'd like.

I think some of our engineers had

some technical issues with it,

but they didn't like.

But I don't recall the details.

I think they thought it might have been

able to be done in a more secure way.

But overall, again,

it's a field that they don't have to use.

Your phone and your laptop

don't stop connecting

just because you stop using them.

They can still share your location

and sync data in the background.

Those connections can help big tech

and governments track you

and give hackers a way in.

But that's why I like to use

Faraday bags like SLNT.

SLNT makes Faraday bags

that physically block wireless signals

to and from your device.

If you want a physical guarantee that

nothing can actually access your device,

just slide your phone in, close the bag,

and the connection stops.

There's no app, there's no settings,

no subscription.

It's literally a physical bag

that physically prevents

any signals from leaving.

It's one of the simplest ways to take

some control back over your privacy.

Check them out at slnt.com slash techlore.

That's slnt.com slash techlore.

And that code at checkout gets you 15% off

and free shipping

on any qualifying orders.

Now back to the video.

Something that I have issues with, and you

kind of alluded to this as you were

telling the story, is that a lot

of people, especially in the open source

world, privacy, security, et cetera,

feel quite disconnected from, I wouldn't

say just politics as a genre,

but they don't feel like they're ever

going to have their needs heard or met.

And I feel like your story is a really

good alternative to that story

that is overall told online.

So if someone's listening

to this interview and they want

to know how they can kind of get involved

or how they can make their voice heard,

they might not have

to write the actual language that ends up

in the bill like you guys did.

But what would you overall say

to people so that they

can feel a bit more empowered?

The prime mover for all

of this was the open source community

emailing the authors of the legislation.

That's the entire reason

that we were able to get involved.

It did make a difference.

It was, yeah, it just opened the door.

So writing, you know,

I can't speak for every legislator.

The three that I got to work

with, it was surprising to me even.

I've had different situations working

with legislators.

But they went in and fought

for an open source exemption for us,

along with my testimony.

And there's a lot

of strategy going on here.

It's politics.

There's weird stuff.

Part of the reason

that age attestation was brought up

was more so to avoid age verification.

It was like the lesser of two evils

because the age verification people

have this whole strategy of trying

to get it in later in Colorado.

So let's try to head it off.

It was awkward for me

because I'm testifying in favor

of a bill that the biggest part

of it isn't what I'm for.

But I understood the bigger picture I had

to keep in mind that this is better

than what it could be.

So, yes, emailing and the people

that got the most that were listened

to the most that emailed in were those

that had good technical arguments.

And the angry letters,

those they just they just kind

of get pushed to the side.

But if you really raise salient points

with the right representatives,

they got through in this case.

Nice.

So I guess to summarize

that, contacting them is good,

productively, not angry, and sticking

with as technical as they can be.

Yeah.

It was kind of summarizing that.

Did you hear from any of them?

Because I've heard some discussions online

that a lot of these people

might not, the emails don't matter

as much as a phone call, for example.

I've heard like different

takes on this online.

I don't know if you were able

to kind of get any of that info.

And in my case, it was email.

But I would guess that's per person,

per representative.

It's an issue I see a lot because anytime,

like when the California law was proposed,

I made a video about it, kind of overall

summarizing what it covers, what to do.

And I always say like, oh, hey, here's

how you find who your politicians are.

I'd say like 20% of the comments

are like, oh, you're an idiot.

You think that's going to do anything?

You might as well just watch paint dry.

And it's like, well, politics is kind

of watching paint dry sometimes.

And it's still how it works.

So it is frustrating for me.

But that's why I like your story a lot,

is it really shows that getting involved

can make a difference.

You're right.

We have one example to point to.

Maybe we can make more.

Do you have anything else to share

about that story before I kind

of dive more into some

technical aspects of you guys?

I would urge people in Illinois to look

at the age attestation law there.

A lot of the conversations we were having

about why open source

should be exempt were in person.

So I think it might be good

for us as a community

to articulate those reasons publicly

so that we can share them in Illinois

and with the representatives of Illinois

and see if it might be possible

to start a conversation about an amendment

like the one that happened in California.

Yeah.

And I guess before we move on, I should

ask, because I feel like we're speaking

about it as if it's a given, but I don't

think it's necessarily a given.

What are like, what is the very nuanced

difference between proprietary versus

open source software in this context?

What can someone cite as the reason

why open source should be exempted?

There are a lot of different angles.

Gosh, I kind of got to dig.

Yeah, I'd say just like even higher level,

just a couple higher level ones,

because if someone's going

to email a politician, they probably want

to be able to separate like, hey,

here's why it's different from iOS,

Android, etc. I guess regular Android

that most people are going to be using.

Well, I kind of recall

exactly how we discussed it.

I spent a lot of time on those arguments

to try to make them clear.

Now some of them are escaping me.

But by and large,

I think it's that the open source

community isn't one organization

that has a direct relationship

between a customer and a business.

Everyone that participates in open source

is the owner of the operating system.

The GPL does the opposite

of what proprietary license does.

Instead of restricting someone's use,

it guarantees their ability

to use it in particular ways.

And that's how the language was written.

Using that, you know, unlicensed,

that doesn't restrict the ability

to download, copy, and modify

and redistribute the source code.

Because of that characteristic,

there's no single entity like there

is with a proprietary operating system

to have the responsibility to require

age attestation or verification.

And it can just be removed

by the user anyway.

We also talked about the way

that open source enables innovation

and requiring age attestation limits

that in a lot of ways because open source

is used in ways that are simply automated,

creating user accounts

in an automated fashion where there's

an operating system without any user.

Even servers.

And servers as well.

In our factory, we have desktop

operating systems installed

throughout the imaging systems.

And not a single one

of them is a real person,

and they're all created systematically.

Those are the kind of things that you do

with open source, which you don't

really do in the proprietary world.

And those kind of functionalities would be

or could be hindered by age verification.

I guess you just call it an adult,

but it's not even a person.

And we don't even think of it as a person.

It doesn't have a username

that is equivalent to a person.

Those are a couple of the arguments.

I think it would be good

to put those together in a really easy way

for people to pull the ones

that are most resonant with them and send

to politicians or representatives to

make their arguments.

Very nice.

I had just kind of random tech questions

I grouped up here as you were talking.

I just took a quick note.

So the first one is, I feel like something

that people cite a lot with Apple,

for example, is their ability to do

like this top to bottom integration

because they control hardware,

software, everything.

And I don't think, you know,

you guys quote me if I'm wrong here,

but I don't think you're building

like your own silicon and it's not

like down to the chip necessarily,

but you still have a lot more control

over this entire user experience.

And I think that's overall

what you were speaking to earlier.

So can you kind of point to some examples

where you feel like that top

to bottom integration you have was a real

strength that maybe people can see

or if someone has one of your devices.

I'm just curious because I feel like

this is a commonly cited thing with Apple

and it's fun to see it elsewhere.

What makes Apple unique is that they

don't enable people to use their software

outside of their ecosystem.

And because they don't

enable that, they don't have

to even think about other hardware.

That's simple, but it's also limiting

in drastic ways.

And I mean, you have to buy their thing

to use their software

and their software can't go anywhere else.

It's a very closed idea of the world,

I think, and technology.

For open source, I mean, back when we

started, the lift to make hardware work

on Linux was much, much greater.

Not a single hotkey would work.

Display brightness didn't work.

Essentially, all of the functionality

and the embedded controller had

to be implemented just to get

Linux working in the way that you

would expect it to work on a laptop.

Today, it's a lot better.

Standards are better.

The activity of upstream manufacturers

like Realtek, even Qualcomm

and other companies that develop

the chips are making Linux drivers now.

That means like the Windows world,

more of the burden of bringing the product

all the way up is spread across

the chip manufacturers and the OEM.

But still today,

every product we release is a project.

And it's usually a month to two months

of engineering at the least

for a product just on the software side.

A lot of times we're doing things

like audio, uh, and sometimes it's small.

When you plug in your headphones,

does the audio mute?

Because you really don't want your audio

to continue on the airplane when you,

if you plug those in.

Does the sleep and suspend,

is, is still challenging, and so, uh,

we've got work with, uh, S0ix, um, uh,

and, you know, sometimes we ship with S3,

but we have high standards for being able

to do, you know, 150 simultaneous suspends

without any failures,

and that's what we consider passing.

So those are the kinds of things, um,

all that refinement that is the difference

between taking a product,

taking laptop, installing Linux on it,

and, like, it's okay, that thing,

you know, my touchpad lock doesn't work.

Oh well, that's, you know,

I can live with that.

But those are all

of the fine touches that we,

we put in so that it's the experience

that someone buying a Windows computer

or an Apple computer

or a Linux computer can say,

when I bought this, it had as good

or better experience than those others.

Yeah, it reminds me actually,

just crazy bad luck on your guys's end.

I remember when, and maybe you could speak

to what it was like on your guys's end

with Linus and his Linux challenge,

where I guess you might be able

to tell it better than I do.

But for people who didn't

watch this, Linus did

like a Linux challenge on his channel.

And I think Luke was the other person.

He went with Linux Mint,

Linus went with Pop!_OS.

And then Steam had like a broken package

that bricked Linus's system.

And it was just very bad luck.

But you can probably explain

a bit better what happened there.

It's funny.

I am working with a number

of folks right now because

we're releasing a new product next month.

And so a lot of review

systems are going out.

And I am not sending an email to Linus.

I decided.

I decided.

It's nothing personal.

And I don't even believe in luck.

I don't believe in that, but I think he

might be bad luck for me.

And so, no, I'm just not going to do

that because it was a 24-hour bug.

It was fixed in 24 hours.

It was bizarre because Steam didn't run

because of an issue in systemd.

They're completely unrelated,

or you would think so.

And that's what it was.

And then, you know, Linus has,

you know, quite a bit of an ego.

And I think if a computer says,

this is really dangerous,

do you really want me to do it?

He's just going to say yes.

So he did.

But, you know, we're past that one.

And there are going

to be things like that.

You know, you take your lumps

when you build stuff and you move

on and it's all right.

Yeah, I was curious what it

was like for you guys behind the scenes,

because I remember when that happened,

it was so unfortunate.

But there were a lot of, like, I think

a lot of Linux people came forward

and I think really did a good

job defending you guys.

I saw a lot of videos kind

of, like, defending Linux and kind

of explaining what happened and why it

was, like, a very unfortunate time to bug.

Because all those people it hasn't

even happened to.

And they've used it for years

and it never happened.

I had another Linus experience

that was the same.

This is the second one is when I said,

okay, this is just bad luck.

We released Cosmic,

and it's a very first release

of a platform that's going to be what we

build on for the next 10 years.

So it's early stuff, you know?

Well, he just happens to decide

to do another Linux challenge

and do it live at a gaming LAN party.

And we just released this thing.

It is as fresh as a newborn baby,

and it's like the very first release

of when macOS transitioned to their new

platform and there were some bumps along

the way to get to where it's at today.

That's where we were at.

And sure enough, we didn't decide

to do it live at a LAN party.

And man, okay.

I'm good with Linus.

Maybe if he wasn't obviously

so successful as a creator,

he'd be a great person to hire to test

to break your products internally

before they go out into the world.

One thing to kind of get back

to the tech stuff, you kind

of mentioned that Linux didn't used

to be as good about certain things.

Like standards have gotten better.

Overall, it's gotten, I don't want to say

easier, but some things are no longer

as big of problems as they used to be.

But I'm sure there still are new ones now,

where would you overall say something

in the ecosystem that still feels immature

to you, that is something you're looking

for in the next five years is going

to really upgrade people's experience

once it's kind of resolved?

I do think about this a lot

as we're building Cosmic.

In general, I think if we want

to be really successful,

for Linux desktop to be really successful,

we need to not be afraid of users.

And I mean that in the sense, in users'

ability, and build the things that we

know work really well and just make

them more accessible to people.

So when I talk to my team, this story I've

told plenty of times is about Internet

Explorer and Firefox and Opera.

Back when Internet Explorer had

a large piece of the market, Firefox

and Opera were quite a bit smaller.

Opera came out with tabs first,

and Firefox came out with tabs.

We couldn't imagine browsing

without tabs today.

Microsoft thought that would

be too confusing for their users.

And so they waited years until they

finally put tabs into Internet Explorer.

By the time they finally did it,

they had lost the entire market.

The point of the story is, don't be afraid

of being too complex for users.

They're much smarter, capable, interesting

people than, than you'd expect.

So it's okay to build advanced

interfaces, to not copy.

Instead of copying,

try completely new things.

And I think we're getting better

and better of that in Linux.

Wayland seems to have opened floodgates

to people experimenting

with different kind of compositors

with scrolling tiling and animations

and features that are unique.

And I don't know what's going to be the

thing that breaks through, but that kind

of experimentation is the opportunity.

It will never happen without it.

Good answer.

How do you feel about the mobile

space right now for Linux?

phones?

Is it something you guys are following,

keeping track of?

I feel like that's still

where Linux hasn't quite caught

up yet is in the mobile space.

That's a challenging area.

I have a lot of thoughts about this.

I wonder what an AI phone might look like.

There's going to be something

that disrupts phones.

There will be just like there's now

something disrupting Google and search.

I have a lot of thoughts

about that, about AI in general

with when it comes to privacy

and what's happening in that entire field.

I think we have a responsibility there,

open source, because these things

are undeniably very useful and powerful.

And a lot of our customers are doing

really incredible things with them

in robotics and healthcare

and material sciences.

But I think there's

probably an opportunity

to leapfrog the app ecosystem in a way.

Perhaps if there are really

well-defined APIs, the application

could be developed by an AI.

And then if you're somebody like Spotify,

your app looks exactly how it should,

whatever platform it's on.

If there's a well-defined, well-typed API

that a simple application interface

could be built upon, which is a lot

of things, except for games.

Games are unique, really unique.

It might be that because somebody like

Spotify doesn't want to make four apps

and maintain four apps across ecosystems.

And there are things like Flutter

and Electron that make

that a bit more approachable.

But perhaps if somebody like a supplier

provides an API and you can build

an app using an API for your platform,

then that platform can take

on the burden of the application itself,

but it fits in with any platform.

That's interesting.

I don't know how feasible

that is or how wild that is.

Well, you said experimentation and ideas.

I feel like it's a good example of that.

Because the thing that we have to get past

with phones is those app stores.

That's the thing

that I think really holds us up.

Otherwise, it's just too difficult.

It's just too convenient.

You get a smartwatch

while you have an app,

and that's how you interface with it.

It's the same for a bike computer.

You use an app to interface

with it and use everything.

Or a remote for your TV, or a Roku.

And so sometimes I lose the remote

and have to use the app.

But how do we replace that?

I think that's the key to competing

and creating an open source ecosystem.

How do we get past those app stores?

On the rest of the environment,

we're building Cosmic to be a responsive

app that works really well

from phones all the way up to TVs.

So I think from the open source community,

GNOME is doing it too, and KDE as well.

So I think from a desktop

environment standpoint,

we can build a great, great experience.

It's just those app stores.

Do you already have or are you planning

in the future to have any ARM devices?

Perhaps.

I have a strange relationship with ARM.

I don't feel ARM is very open.

If there are really good ACPI and EFI

implementations on a platform, then you

can install any operating system on it.

If not, you just can't.

And I think that's a terrible

characteristic of hardware.

And so we wouldn't release

hardware that's like that.

You should be able

to install any operating system.

Qualcomm didn't do that.

And we had Snapdragon hardware

that we were evaluating.

And for that reason, we decided

this was not a good route to continue.

Qualcomm is working on that now

to have sane ACPI and EFI implementations.

That could change the game

for ARM products.

But I think that that's necessary

for it to be an open enough platform

that we would be really interested.

Yeah, I just saw in my RSS feed a few days

ago, we're recording this early October,

but the new Microsoft devices,

I think they're using the new Snapdragon.

I'm not familiar.

Is it the X2?

I think they said it's going

to be Linux compatible.

Does that impact you guys at all?

Or is that a different issue?

It depends how well they do.

Okay.

I'll believe it when I see it.

Yeah, I think it's fair.

Yeah.

Microsoft is interesting

where they choose.

And maybe, yeah,

sometimes they choose to support Linux.

Sometimes they don't.

Arm makes it really easy

for hardware manufacturers to lock you in.

Yeah, that's fair.

That's not our thing.

We're going to have an opposite of that.

So I want to see Arm Silicon

that doesn't do that.

Are you looking at RISC-V

at all or anything else?

From time to time, but not in much depth.

Okay.

AI.

It sounds like you're not anti-AI,

but you might be anti how AI can be used

or the privacy implications, et cetera.

I don't want to put words in your mouth,

so please clear up anything to expand

on how you actually feel about it.

But are any of the devices that you

guys have geared towards AI usage?

Is that something you plan

to embrace more, push away from?

I know it's very polarizing in the Linux

world, so I'd just like to hear a little

bit about how you guys are juggling that.

Yeah, I'm happy to share

my thoughts broadly on it.

So we build a lot of products

that are used by companies

that are developing AI models or doing

tuning or inference or other areas

that they're working in, like the fields

that I mentioned earlier,

doing pretty cool things in robotics

or pharmaceuticals or space,

lots of neat applications.

My concern isn't the kinds of things

that they're doing, because they're doing,

we think of ChatGPT and Anthropic,

and we think that's AI.

That's not AI.

That's one corner of AI.

That's actually the corner

of AI I'm most concerned about.

But the stuff our customers are doing

is to enhance their products in ways

that they couldn't do before.

Robotics are being enabled.

The big thing that was missing from

robotics is machine learning in this kind

of type of capacity where it could.

It really started with computer vision

and all this AI stuff.

A robot that would remove packages

from a truck would go

in and grab the one at the bottom

and the rest of the boxes would fall.

It couldn't reason through how should

I actually do this, all of these pieces.

And now it's able to do

that, and it's kind of expanding the world

in really interesting ways

for manufacturing and automation,

things that just make us more efficient.

The things that concern me is that it's

so useful in a lot of ways that I think

people are going to give up things they

shouldn't give up for its usefulness.

They're going to do things

that I think are just completely crazy,

like share their tax return with an AI

to see if they miss something.

Or access to your bank account

or your retirement plan

or your user account on your computer.

You don't know if an AI,

if an LLM will reason that sharing

your private key is the right thing to do.

You can't know that.

And it's just too risky.

But because it's so useful,

I'm afraid that people are going to do it.

Look at the Muse app from Facebook and you

give your contacts, your transactions,

your emails, your messages.

It's your entire life.

And to have a personal assistant, yeah,

that's what a personal assistant kind

of has access to.

But that human has gone

through interviews and signed an NDA.

And there's all these other things

that go together to trusting that person.

And we're just handing it over.

You can hold a human liable, too.

You can't do it with a machine.

And you can't, precisely, yes.

And so it's very frightening to me.

Because I think it's so useful

that people are going to do it.

And as an open source community,

what we should be thinking

about is how can we help people get

this kind of capability responsibly?

Right.

You know, small private LLMs

and carefully sandbox and thinking through

security before jumping in the pool.

You know, all those kind of things,

because I would rather see a world

where people didn't need to give those up

to third parties to have that capability.

There is a time when I want us to think

about this because of those reasons

that people are going to, I think, give

up too much because of the usefulness.

And if other people in the open source

community are thinking about it

in this way, yeah, I think let's think

about how we can harness the good things

without all the bad characteristics.

You know, maybe we

can build something that helps

people avoid giving away so much.

Very nice.

Memory prices, is that something

that has impacted you guys

at all as a result of AI?

Is that been a problem?

Yeah, they are higher.

A lot.

I don't really talk about it much because

there's nothing I can do and there's

nothing that's going to change.

So it's one of the things,

it's just what it is.

Computers are more expensive

than they were before.

I don't like it.

I don't think anybody should like it,

but I don't know when that changes.

So it's kind of just what it is.

Got it.

I wanted to ask earlier,

you mentioned that you took profits,

put it back into the company.

This is early on your journey.

So are you guys only self-funded

with your own profits or do you have

any investors or anything like that?

We are.

Yeah, just an organically grown company.

Nice.

That's awesome.

Yeah, congrats.

I wish every company was done this way.

It would prevent a lot of issues

in the world, I think.

It's not easy.

It's a challenging way to do it.

So I don't know the right way, but I love

the freedom of us just answering

to ourselves and our customers.

Okay, congrats.

It's a big deal.

And then just kind of what's next.

I know you said you already know,

but I don't expect you

to share everything, but whatever you

feel comfortable sharing, if anyone's

listening, they're a big fan of you guys,

what can you kind of tease for them?

So we have a new desktop coming out next

month and I'm extremely excited about it.

I don't want to share too much because

there's going to be a reveal and all

of that, but it is absolutely gorgeous.

It's a product I'm really proud of.

We'll be shipping Pop!_OS 26.04 very soon.

We're on 24.04 right now.

There was this really long period where

we're building a desktop environment.

We shipped 24.04.

I wanted to ship 26.04 very soon

after that, but we decided

to get ambitious and do some more things.

And so 26.04 will

be coming out this month.

Pop is different than...

We use the numbering that Ubuntu does,

but we're quite a bit different because

the kernel, Mesa, a lot of the packages

are updated throughout its life.

So it's almost like a rolling release,

but with a more stable base.

And Cosmic continuously moves too.

So your desktop isn't stagnant.

It's always upgrading.

The other things I'm thinking

of are really, really cool.

And I can't tell you what it is because

I really want to do a reveal

when it's ready to show you.

But the things we're doing

are both getting smaller

and bigger at the same time.

And besides products coming next month,

next year, we're going to show people

what we're really, really capable

of now that we've built Cosmic.

Cosmic can do things

that are very, very unique.

We're going to start showing

people that next year.

Wow.

Congrats.

Yeah.

Congrats on everything so far.

Thank you for the teasers.

And I guess just a fun question to, you

know, let's say Pop!_OS was eradicated

from the world in some awful event

and you have to pick a different distro.

What are you picking for yourself?

Oh, I like that question.

I can't live without Cosmic today.

It's my favorite.

You can put Cosmic on any of them.

Yes, Cosmic's going on anything.

I'm thinking, okay, can I answer two?

Is that fair?

You can have a work and personal device

or however you want

to split it up or dual boot.

I would go for CachyOS and AerynOS.

I really think there are kind

of neat things happening on both of those.

And I've used Cachy

and I've used AerynOS, but not as much

as I think I would like to.

So I want to dig a bit more into AerynOS,

especially if they have some neat stuff.

Very cool.

Well, how can people follow you guys

if they're listening and they want

to learn more and keep up

with what you guys are doing?

X or Mastodon are probably the easiest.

Just at Carl Richell.

And System76 is just at System76.

And pretty easy to find us

in both those places.

Awesome.

Thank you, Carl.

Thank you.

And that, everyone,

was my interview with Carl from System76.

You can learn more about Carl

and System76 down in the description.

And of course, if you liked

this interview, please give it

a like or a rating or leave a comment

on whatever platform you're on to help

support the podcast and keep it growing.

And if you want to support what we do,

become a Techlorian down

in the description.

We have a private signal group and it

keeps interviews like this possible to do.

Thank you all for watching and I'll

see you next time on Techlore Talks.