@mux/ai

The biggest friction point

whenever I'm trying to get

something online is having

to go spin up a bunch of

different accounts and

set up billing in multiple

places and configure the

alerting right and go through

the onboarding process.

That's after I've actually

decided on the tools

that I want to use.

Sometimes I don't know

the right tool to use.

I don't know the right SaaS

solution in that vertical that

has the features that I need.

Or that integrates well

with the other parts

of my technology stack

that I'm trying to use.

Like broadly speaking, the

big picture of the marketplace

for me is the same clone and

deploy experience that people

enjoy Vercel for, which is you

can get a site online globally

distributed in under a minute,

we want to extend that to

any of the services that are

a part of the marketplace,

whether that's setting up

your video infrastructure,

setting up your database,

setting up your AI inference,

setting up really anything

that you want can all be a

part of that same workflow.

Lee Robinson,

you might have heard of him.

We're big fans.

Not only is Lee the VP of

Product at Vercel, he's a

successful content creator

himself — consistently

sharing insights on

modern web development,

frameworks, and best

practices for scaling apps.

In this chat, we're diving

into that feature he

mentioned in the intro, the

Vercel Marketplace, and how

it helps devs ship faster.

We'll also get into some

thoughts around AI and web

development, pick Lee's brain

on some tips and tools for

content creators, and more.

Something that I probably

don't talk about enough

and I kind of just take

it for granted because

it's a core part of my

developer tools ethos, but

strongly believe in free

tiers, strongly believe in.

Allowing people to sign up

with zero friction and then

kind of grow as needed.

And I think what

marketplaces do in a lot

of ways, especially for

very busy verticals, like

databases, is that it really

incentivizes people to

make compelling offers for

the things that developers

really, really want.

Which is they want to sign

up in as little friction

as possible, start storing

some data, like actually try

it out and kick the tires.

And then if the product

is good, then they will

stay and convert to a

paid user, and especially

with, you know, stickier

products that have storage.

It's like, oh, yeah, of

course, once I've already got

my data schema established,

now it makes sense, I

want to keep growing here.

I think in a lot of ways

to what Mux is doing at

the same, same level.

So what I hope is that

for every solution that

gets integrated into the

marketplace, they have that

same level of simplicity

for setting up, but then

also provide good cloud

infrastructure, good tools

on top to then grow further.

What I thought we would talk

about a little bit here is

just what value Mux being

in the marketplace brings

to a developer working

with video, or even just

like what opportunities you

might see Lee as a developer

working with video, how

they can kind of leverage an

integration like this to level

up their video integration

in their Next apps.

Yeah, well, I would say

there's kind of two or three

big main categories of video

problems that I see Next.js

and Vercel customers run

into pretty frequently.

Which is one, honestly,

they're not sure the

right way to deliver

the video to customers.

They, they know

they have video.

They know they want to

get it on their website

or web application.

They don't know the right

way to package it in a way

that's going to be performant.

And I think having

opinionated solutions is

very helpful for that.

The second one is that,

you know, they already know

they have some video, but

they want to figure out the

right way to compress it,

transform it, get it into a

format that's as digestible

as possible for the insight.

And then I would say the

third one that I think

y'all do a very good job

with is also the developer

experience of how the dev

actually writes code that

integrates into the tools that

they're using, whether that's

Next.js and React or other

frameworks, like building

tools that make it very

easy to install a component.

You know, put in an ID or an

account and just easily get

something flowing, abstracts

away a lot of the little

gotchas that people probably

haven't spent as much time

thinking about as y'all,

because just like I know

video is the same as photo.

And I know photos the

same as some other areas,

where there's so many

little tricky things that most

people don't think about to

get right, that you really

appreciate the abstraction

when you start to use a

more full featured, feature

rich component that does a

lot of that stuff for you.

You don't have to

be the video expert.

And I've said this in other

conversations too, where

like, internally, I will

still have conversations with

some of the video engineers

on our team that I don't,

I still don't understand.

And I've worked here

for three years.

Like having these

conversations about how tricky

video can ultimately get and

how that just kind of gets

abstracted away or packaged,

it's kind of the whole goal.

That's what Mux is all about.

Have you seen, with all the

projects that you spend time

with, have you seen innovative

ways that folks have leveraged

video in their applications?

And what might that, this

sort of integration enable

to make those sorts of

innovations even easier or

more potentially explorable?

Yeah, there's of course

the traditional use cases

for video and we see lots

of folks doing that, but

increasingly as more folks

are doing kind of generative

AI, both in photo, but

especially in video, they're

kind of building these AI

native interfaces that are

taking what used to be a

video editor or taking what

used to be video tools and

making it accessible to,

broadly, many different

creators who probably

don't know a lot

about video editing.

Like, I kind of think about

it as the TikTokification

of the video editor, where

it's like super easy now

for people who want TikTok.

I see that on the web now,

as well, with people building

AI-first video products, and

they can leverage Mux for

all sorts of different video

tools on top of, of course,

like the core AI workflows

that they're building out.

So I've seen that more

and more frequently now.

What, what are the, the

innovations that y'all are

pursuing for 2025 and how,

what's the, is there any tie

in to what the marketplace is

trying to accomplish with any

of the other stuff that you're

working on over at Vercel?

Yeah.

I think one of our big goals

for 2025 is supporting more

demanding, complex,

personalized, dynamic

workloads as a lot of our

customers who've been with

us for some time have grown

to a significant scale.

They're trying to run, you

know, very personalized,

dynamic experiments and

trying to build applications

that are, you know, much

more complex than what

they were previously doing.

And to do that, it

requires more compute, it

requires more integrations.

It might require working with

other services who are kind

of custom built for that.

You know, a classic example

is background jobs or queues

is something that Vercel has

a cron system, but it's for

basic, you know, workflows.

But once you start getting

into, you know, durable

background functions, there's

a lot of complexity that

goes down in that hole.

And sometimes people

want to rely on third

party integrations that

are now becoming first

party integrations.

Same thing with video,

same thing with photo, same

thing with many different

verticals, including AI

inference, which we've been

partnering with folks on, and

a very increasingly common

use case on Vercel as well.

So that's a big category is

just supporting increasing

customer demand for

highly dynamic, stateful,

personalized workloads and

just continuing making our

product portfolio there

more robust, but also still

integrating with best in class

solutions where it makes, you

know, it makes the most sense.

So one of the things that

we've played with on our end

is trying to find ways to

make v0 the most compatible

that it can be with some of

the components and products

on our side of things.

Do you see any innovation or

development in v0 land and

how that can ultimately help

bring somebody's vision with

a video or their own products

with Next.js to life quicker?

Yeah, for sure.

So for context, v0 is

an, an AI web development

agent that we build.

And I know agents are pretty

generic world today, but

it helps you understand

web technologies and build

web applications better.

So you can ask it questions,

but you can also ask

it to generate code.

And it has context of all the

Vercel and Next.js and related

tools in the ecosystem.

One really popular feature

that it has is you can drop

in screenshots of maybe a

component you want to build or

a part of a website that you

like and get some inspiration

from that kind of fork it

and build your own version

and tweak things as needed.

Just recently I saw a request

from someone to build that

same feature but for video

and the complexity there

is definitely steep, right?

Like being able to accept

in a video, process it,

understand the individual

frames, I guess try to

figure out a way to turn

the frames into individual

pieces of the app and like

what the sequencing is there.

I think something like

that would be really cool.

And then on the kind of

flip side, being able to

integrate video tooling

into the generated

apps that v0 can build.

Sometimes I think

of v0 as kind of our

general knowledge hub.

So, for it to have

really great context

of, you know, here's how

you should do a video.

Here's the things you

should look out for.

Here's the things

that you need to know.

Here's the tools that

you can integrate with.

What the marketplace does

that's really, really exciting

is, let's say you're trying

to build an app that works

with video, and you ask v0,

and it says, hey, here's

how you should build it,

and you should use Mux.

So you want to use Mux, and

you want to deploy to Vercel,

what you can say is just,

I would like to integrate

these two services together.

v0 will prompt you to enter

in your environment variable

for that service, and you

can go out to Vercel, and

in a few clicks, integrate

directly to Mux, get the

environment variable in

your project, and now you're

set up and ready to go.

And we might even make that,

you know, fewer clicks in the

future, but that's something

It allows you to kind of chat

with the AI to build your

product in real time and to

code in real time while still

storing in secure connections

to other services through

the marketplace and still

use those services kind of

on the back end, which I

think is really interesting.

Now, do you have, this is a

little bit of knowledge gap

with v0, but is there a way

to provide like documentation

or any kind of links to just

help it out with like, hey,

there's a lot of different

parts for video if you have

this link or hey, this, you

can use this as your resource

to be able to kind of scan

through and see how you might

integrate different parts.

Is that a feature

that v0 enables?

Yeah, totally.

So for common web dev

tools, frameworks, libraries

like ORMs, it already

has context of that.

But if you want to provide an

additional context, you can

attach documents into the chat

as you're working with, you

know, whatever chat you're on.

Or when you create a project

in v0, you can add additional

sources, which can be files

or, whether that's actual

code files or Markdown

files or text files, you can

adjust like a prompt inside

of there for what you want

the system to know about as

you're working on the project.

So you can provide

additional knowledge or

kind of guide the AI in the

right way that you prefer.

Man, it's such a wild world.

Like I, sometimes I sit back

and think of like, we've

come so far, it's like 20

years, 20 years ago for me to

spin up all these different

services and then write the

code to integrate like an

uploader or have like the

app deployed, if that's even

possible, like with a, without

FTP into the server, like

we've come so far that it

just kind of blows my mind.

These kids these days.

Yeah.

The it's, it's the biggest

mind bending situation for

me because I think about how

long it would have taken me

to complete and integrate

all of the different things

I'm able to do now in like

10 minutes, 15 minutes.

Like quite literally the

time to v0, hence the name,

has never been faster.

Like you can just ask for

something review the code

and kind of treat it more

like, you know, a junior

engineer that you're working

with where you're going to

trust but verify the output.

Okay, this looks right.

This looks good.

Change this thing.

Actually, you should

do this better.

But then you have a working

app and you can just

build a lot of things much

faster and focus more on

delivering product value.

Gosh.

So wild.

All right.

I want to ask you

Lee about the video

production side of things.

So you are, you're obviously

in front of the camera a lot.

You're doing tutorials

and interviews.

Why is this important to

you in terms of even just

personally, like what, what

is your personal preference

for video or being on camera

and how does that benefit you?

How does that help you grow as

a creator, as an individual?

And what is that doing

for your overall strategy?

Yeah, so me and, me and

video have a long history.

I've been making and editing

videos since I was a nerdy

kid in my early teens.

I'm still a nerdy kid.

I'm just now in my 30s.

But the passion has been

there that whole time.

And you know, when I got

started, I was just doing

things with like video games

and downloading cracked

versions of video editors.

Sorry, mom.

I wasn't going to pay for

that, but I've been doing

this for a very long time.

And when I got into

programming, when I got into

tech, that love of video

kind of never went away and

I kind of just stumbled into

doing developer education.

I didn't really set out

to do that, but it was a,

a mesh of all the things

that I really cared about.

I like programming.

I like tech in general.

I like teaching people and I

like finding the right medium

to deliver that message.

And sometimes that's writing.

Sometimes that's

an in person event.

But increasingly,

for most things, it's actually

distributing video online.

Video is the primary way

that I get my message out

to people in our community.

More than just the, you know,

local work that I could do,

or the work that I could do

when I'm in San Francisco,

I'm able to talk to people

all over the world, get the

message in all of our sub

communities of Next.js around

the world all thanks to video

So really nailing in the video

strategy, you know quality,

content has been something

that I've spent a lot of

energy on and I try to get

you know just 1 percent better

every day like I'm in the

process, not right now, but

over the next two months-ish

will be kind of redoing my

studio and better lighting

and all sorts of stuff.

So I'm looking forward

to that.

Lee, with as much time

as you spent on camera

and learning all these

cracked, or pirated software

solutions, which maybe they've

evolved over the years.

Allegedly.

Allegedly, we'll see.

What have you learned?

In sum, I mean,

there's, there's years,

years of experience.

What sort of advice might

you have to somebody that's

out there that's looking

at getting started with

video or media development

to help pursue some of the

things that you've talked

about being passionate about

pursuing with whether that's

education or just kind of

like growing their influence?

Yeah, I think, especially

in recent years, so let's

call it like the past

two or three years, The

developer creator education

space is kind of bifurcated

into two types of video.

There is educational

video, but then there's

also entertainment

video, and they are not

necessarily the same thing.

Sometimes they're the same

thing, but they do not

have to be the same thing.

So the former category

is everything you're

probably used to.

It's YouTube videos,

it's courses that

you could purchase.

It's the things that you

might link inside of a

documentation website.

It's the things that you

might have in an onboarding.

The latter is increasingly

more live streams and

actually spending time with

creators and developers one

on one versus, you know,

just to have fun versus

actually some kind of

structured learning content.

Now, you might learn as a,

you know, as a side effect

of having fun on a live

stream, but that's not

necessarily the primary goal.

It's more of a community

building thing, and

video is the medium in

which it's delivered.

So, my advice, or at least

one advice, getting started

is, understand the difference

and make sure you know which

one you're wanting to spend

time on or both, but they

require different skills and

the output of those will be

different my recommendation

would probably be to start

on the former versus the

latter because there is

an ever growing desire

for high quality developer

education and especially

in an AI-first world.

I think it's fantastic that

we have infinite knowledge

at our fingertips, yet humans

crave humans and humans crave,

you know, very specifically

structured educational

patterns and courses and a

very thoughtful curriculum.

And I think the best educators

are starting to do that find a

way to merge those two worlds.

Yes, AI is great for some

things, but here's how I

deliver a video with my

personality and my style

and the way that I like

to teach in a, in a more

human nature, even if I get

assistance on code from an AI.

So maybe that will be helpful.

Yeah.

Yeah.

I feel like that's kind of the

key here with a lot of these

AI tools or AI assistance

is, yeah, it's great.

You know, it's a great

starting point for

getting kind of the ideas

flowing, but in terms of

the actual packaging and

delivery of how we can

leverage that knowledge,

but make it relatable and

communicate it to a person.

That's an entirely

different game.

Right?

And maybe there's just some

data set that eventually

will be big enough for

the AIs are trained and

we're all just, whatever.

We can't tell the difference

between an AI and a human.

I don't know if that's ever

going to be the truth, but

right now, more important than

ever, to me anyway, is the

human side of communication.

Is being able to show up on

a live stream with somebody

and just be able to talk about

what's inspiring, where this

stuff is heading, and what's

got you most interested.

So yeah, I totally resonate

with that sentiment.

Yeah, the thing is that AI

can't replace craft and taste.

And while we always, we say

on v0 a lot that, you know,

everyone can cook now, like

anybody can be a builder,

AI is kind of like the

cookbook and the ingredients.

So it gives you some of

the raw ingredients, but

you, as the controller of

the system, you are the one

who's putting the pieces

of the cookbook together.

Like, you know what

good looks like.

You know the structure

you want even if you

have assistance along the

way of actually putting

the pieces together.

The recipes together.

In that vein of new

technology, is there a new

tech or trend that you've seen

that you haven't had time to

really dig into yet, but plan

on trying to uncover in 2025?

That's

a good question.

I haven't done as much

with voice as a modality

in some of the latest AI.

Models and applications.

And I think it's really

interesting, whether

that's like voice to voice

where you can chat with an

assistant, or voice to code.

So you just speak and you

can generate parts of your

application or voice to

refactor where you're like

talking and you're telling

it, okay, change this thing

and then change that thing.

I've started to find myself,

because I can talk faster

than I can write, speaking

a lot more and, and audio

transcription has got a

lot better, especially with

AI note taking tools have

gotten quite a bit better.

So, I would like to do more

stuff with voice next year.

This year, 2025, we're here.

So Lee, what about in your

toolbox for when you're

creating videos or anything

that's going to help you

be a more efficient video

producer or content producer?

Is there anything that stands

out as a valuable tool tool

that you can't live without?

Yeah, I think for a lot of

folks a tool like Screen

Studio can take you really

far because it does, you

know, pretty good screen

captures, and then also has

built in captioning, and I

think it also has the ability

to touch up the audio.

Those are kind of

three main pillars.

I kind of have a more

bespoke workflow just

because of what I'm used to.

I use ScreenFlow.

I find that it works well

for my screen recordings.

Sometimes I will use a tool

to enhance or fix the audio.

Adobe's podcast tool

is like pretty good.

If you haven't seen that

before, it's like does

a pretty good job of

cleaning up the audio.

In terms of other

tools for video.

I mean, I have lots of nerdy

things I can go into with

like the LUTs and the camera

settings and like all that

good stuff, but yeah, mostly

the bread and butter is

getting the right software

for recording and editing.

I, I have Final Cut,

but I don't really use

it that much anymore.

Recording, editing, audio.

And then publishing,

shout out to Typefully

for publishing to

multiple social channels.

It's a great tool.

Lee, with the new Mux

integration on the Vercel

Marketplace, would you

have any suggestions for a

developer that's interested

in getting started or what

the next step is where

they can really begin?

Yeah, we have hopefully now

made it as easy as possible

for you to get started with

Next.js or your framework

of choice and Mux and on

Vercel.com/marketplace you

can in a few clicks set

up your Mux account, get

connected, get off to the

races on integrating our

services together and get your

site online pretty quickly.

And hopefully we'll have, if

we don't already have, we'll

soon have some templates

for you to get started with.

Awesome, Lee, thank you

for your time today.

Really appreciate it.

And yeah, I'm really thrilled

about having this integration

on the marketplace as well.

Anything, like I said before,

you know, that we can do

to help make developers

lives easier and reduce the

amount of friction that it

takes in standing up a heavy

requirements for some of these

production applications is a

win in my book, whatever we

can do to reduce complexity

and just make things easier

and that's, that's all we, I

think we can ask for in terms

of the developer experience.

A hundred percent.

Yeah.

Thank you for having me.