YouTube Transcript:
The painful truth about startups (my story)

Skip watching entire videos - get the full transcript, search for keywords, and copy with one click.

Video Transcript

Video Summary

Summary

Core Theme

This content details the speaker's unconventional and "chaotic" journey from a junior engineer to a successful creator and CEO, emphasizing the importance of user focus, adaptability, and learning from mistakes.

Mind Map

Click to expand

Click to explore the full interactive mind map • Zoom, pan, and navigate

It's getting harder and harder to

describe what my job is. On one hand,

I'm obviously a creator and an

influencer. I mean, you're watching this

video, right? On the other hand, I do a

lot of software dev. I spend a ton of my

time coding, as much as I reasonably

can. But also, I run a company, a

startup. I know I've touched on my

journey on how I got into startups here

and there, but I don't feel like I've

ever properly broken down how I went

from college to working at Twitch for 5

years to running a company of my own.

and also of course what that journey of

running the company has been like too.

There's a lot to share here. My

journey's been very chaotic and I need

to be honest with you guys. I never

thought I would be an influencer, much

less a CEO. And the order that these

things happened in might also surprise

you. There's a lot to dive into here. I

hope this is useful for you as an

engineer. Whether or not you're

interested in doing a startup or maybe

you already are. The goal is to just

show how I got here so that you can make

your own good decisions too and maybe

learn something along the way. Super

excited to dive into my chaotic journey.

But as always, the bills need to be

paid. So, quick word from today's

sponsor and then we'll dive right in.

Can I talk about O for a second? You

know, the thing that everybody says is

easy but isn't. You would imagine that

really big companies are rolling their

own, right? You know, like the

snowflakes, the Cartas, the Versels, the

Cursor, and the OpenAIs of the world.

What if I told you all of those

companies were using the same O solution

and that they weren't rolling it

themselves? that they actually picked a

vendor that makes a lot of sense for

real businesses, both big and small. And

what if I told you your first million

users were free? You'd probably think

I'm going nuts. And I understand because

when I look at this page on the Work OS

site, I kind of feel like I'm going

nuts. Do you see how many real

businesses are using these guys already?

It's kind of nuts and it makes me

question why I pick other options as

often as I do because I need a lot of

the stuff that works offers me. They're

truly enterprise ready off like all of

the things you need from the compliance

side to the portal that you send to a

random IT person to get the things set

up to the integrations with your

existing infrastructure to the backend

safety. All these things that are not

fun to do, they have done for you. They

do the boring part so you can get back

to work. And I can't ask for anything

more of any provider that I work with.

If I'm being honest with you guys, I

screwed up not using them for T3 Chat.

had a lot of companies interested in

doing business plans and enterprise

deals that we just aren't equipped to do

because that layer isn't set up

properly. You bet your butts we're going

to be moving to work OS as soon as we

have the spare cycles to do it. Don't

make the same mistake I did. Check them

out today at soyv.link/workos.

When I was fresh out of college, I was

not very good. I was a practical

engineer where I like built stuff but at

the same time I just didn't deeply

understand or really care to understand

how to build good software. I just made

solutions to problems. My background in

software was making plugins for

Minecraft servers and helping make a lot

of the bucket stuff happen earlier on

when I was getting into Minecraft. I

never thought I would be an engineer. I

wanted to work with computers. I knew

that. But also I wanted to ride my

skateboard and do music too. But the

computer stuff took off quick. I decided

to go to college for it. Got decent

enough at Java by watching new Boston

tutorials. In the end, I went to

college. I got my music degree and my CS

degree, but I didn't have much to show

for myself. I had one internship that I

kind of bombed. Partially my fault,

partially not, I would argue, because my

mentor, the person who was assigned to

like manage me, took a two-month

vacation for the first two of the three

months of my internship. So, he just

wasn't there. And according to the other

interns who were in my general area,

they would have left if they were on

that team because they were yelling at

each other constantly, even harassing me

as the intern that didn't know what was

going on. I was given a 5,400 RPM hard

drive on a crappy $200 laptop, even

though it was like a Amazon internship

where I was being paid a lot of money.

Just all these little things went wrong

with that internship and I almost went

mad. As such, not a good source of

referrals. But at the end of that

summer, I did have one thing which was

my Chrome extension, Chrome Tana. The

goal of Cortana was to redirect Bing to

a better place because with Windows 10

shipping, Bing was now deeply integrated

into Windows and it was quite annoying.

So I made it so you could redirect Bing

searches elsewhere and Microsoft

responded by making it so you can only

open up Edge with the search bar at the

bottom. Long story, don't need to go too

deep into it, but that was like my main

claim to fame. It was that and a bunch

of like different watch faces for the

Pebble watches. Nothing super relevant.

I got this gig because of a couple

people at Twitch specifically that had

surprisingly high faith in me. I'll name

drop them. I don't see any harm in that.

Thank you, Bill and Wava. Bill was one

of the first hires at Twitch. He helped

build the original chat implementation

in Pearl. Absolute lunatic. Loved Bill

so much. He was the first person that I

talked to during my interview loop and I

didn't realize he was the exec for the

org that I was interviewing for because

he was just like a chill, fun dude.

Didn't ask me any technical questions.

Just wanted to chat. And our

conversation ended up going deep into

our favorite and more importantly our

least favorite da punk music. And he has

very good taste in da punk. Learned

afterwards that I was so young that I

was literally born around the same time

that Homework came out by Da Punk, which

is both of our favorites. So, he was

very impressed that I like could go all

the way back even though he was in

college at that time and I was not born

in that time. We also both hate random

access memories. That's a long rant for

another day. I promise there's a reason

I'm going into my history with Bill and

Wobba. These people have been very

helpful throughout. Wobba was the

engineering manager on this team. He was

one of the most important [ __ ] in my

life. I owe him so much for cracking the

whip and making me into the engineer I

am today. For functional program pilling

me really early in my career, for

helping me fall in love with the

beautiful language that is elixir. Waba

helped shape the engineer I am today and

I owe him forever for it. But when I

would tell people that Waba was my

manager, they would look and say they

let him manage people. Oh my god, I'm

sorry. I was so thankful to have him as

my manager. I learned so much from him,

but he hated being a manager. So after

he managed me and one other person for a

year, he moved back to IC, hopped around

a few times. Now he's at OpenAI if I

recall. He has never managed since cuz

he hates managing. But I was lucky

enough to be one of the two people he

was ever the official manager for.

Learned a ton from him. I bombed my

interview though. The only reason I got

in is because Bill really liked me and

Waba had a weird amount of faith in me.

These two decided that even if they

couldn't justify a full-time offer, they

could give me a three-month contract to

prove myself. And if I succeeded, I'd

get the full-time offer. I hadn't

succeeded after the 3 months. So, I got

another 3-month contract extension. And

by that point, I really had like locked

in, got my [ __ ] together, and started

contributing meaningfully and finally

got that full-time offer. Huge shout out

to both of them for extending the very

generous offer that I was not ready for

when I applied. and letting this all

happen. My time on this team was

interesting both because like I got to

learn a bunch. I got to build a bunch.

Like I wrote the code that ran the Bob

Ross marathon, the Power Rangers

marathon, all the marathon content on

Twitch. I helped start Clips TV, which

kind of over time became the Clips Tik

Tok product, which eventually became the

new feed. I'm sorry. I warned them

against it. All of this was done under

the creative team, which was also the

non-gaming team. Around when I joined,

Twitch was still very much a gaming

platform to the point where if you were

doing things other than gaming, you

would be banned for it. If you were to

dance between games of League of

Legends, your partner manager would hit

you up and say, "Hey, you have to stop

doing that cuz we're not a dancing

platform. We're a gaming platform." As I

had mentioned before, Bill specifically

was not a gamer. Bill was questioning

why he was still at Twitch after all of

his time there. and he convinced his

good friend who was the CEO of Twitch,

EMTT, to let him invest in a non-gaming

side of Twitch. It was the music side,

it was the IRL side, it was the fitness

side, the art side, painting, and all of

that. He also landed the contract with

the owners of the Bob Ross IP so we

could do the Bob Ross marathon, which

absolutely blew up the creative side of

Twitch. Until then, people didn't even

know there was non-gaming stuff on

Twitch. Now, there's a whole deep scene

within the music world, within the art

world, within the IRL stuff, and a lot

more, too. Shout out to my homie Skirt,

who's one of the earliest and hardest

working people in the music world on

Twitch. Always good to see you, man.

Hope you're doing well. And I was super

lucky to be on this team for a bunch of

reasons. I'm going to do one more weird

tangent about this and then we'll blast

through the rest. I had to thank a

person that I'm very lucky to call a

friend, John Wayne. Since the creative

team was the team in charge of all

non-gaming at Twitch, my role was mostly

the engineering side, helping write the

code for the creative specific stuff. We

were exploring making our own OBS clone

for Mac users doing art stuff, but

mostly my job became the premium content

side, the marathon stuff. That said, my

team was two engineers, Waba, my engine

manager, and then like seven

partnerships people. None of the

partnerships people knew much about

music and in particular, none of them

knew anything about hip-hop. John Wayne

is both deep in the gaming world and in

the hip-hop world, both as a producer

and as a rapper. He was one of my like

childhood heroes, one of my favorite

musicians of all time. You don't believe

me? Ask my godamn mom. I've been

obsessed with this guy since I was like

15. He's a one of my favorite musicians

and like a personal hero of mine. So,

when John Wayne, one of my favorite humans,

humans,

suddenly showed up at Twitch trying to

stream, having no idea who to talk to or

how to get it all set up, and I showed

up in his chat with a little wrench and

said, "Hey, can I help you out?" He was

super down. And the cool thing is, since

nobody on my team got music and in

particular the hip-hop world, the

partnerships guys were actually hyped

for me to come in and take over there.

And I ended up really deep helping music

over my entire career at Twitch. I

helped like moderate for and onboard

Kenny Beats. I helped get a lot of crazy

deals inked. We almost had a big rap

battle network on Twitch. I got to help

with a ton of stuff and it was super

cool. The reason I'm bringing up John is

he he did a really really important

thing for me. He broke my whole never

meet your heroes mindset. As great as I

was hoping John would be, he was [ __ ]

10 times better. He was just such a

good, thoughtful, well-meaning dude, and

I am so lucky to have him as a good,

close friend today. We talk all the time

about everything from like taxes to our

lives and how we think about the way the

world is working and going. I've helped

him start his new YouTube channel, which

to be fair is entirely his work. I was

just like a a YouTube coach friend that

he could talk through things with, and

he is killing it on his YouTube

channels. Now, there's two of them. And

the fact that he's found a way to, as an

independent musician, make a living

making content is just unbelievably

cool. I love so much. And he has

continued to like motivate me to push to

make all of these tools and platforms

and everything better. I love John so

much. If you're into production music

stuff, highly recommend checking out

this channel. There is so much good

content here. He did just start a second

channel that I think is really cool. He

told me that this video is based on a

diagram that I had made, which I think

is the coolest [ __ ] ever. John's the

best. The main reason I brought him up

is how important he was in teaching me

that if I reach out to people I look up

to with actual value to bring them,

there's a good chance they'll reply and

there's a good chance they'll be

awesome. And that formed the way I do

everything going forward. I need to stop

for a sec to thank the giant 50 bomb

that we just had on Twitch. Thank you so

much, Yuzu, for gifting 50 subs. Really

appreciate that, man. Thank you so much.

Yeah, there's nothing better than your

heroes being fire. So, as I hope we've

established here, a lot of important

precedent was set for me. Being on the

creative team, it meant that my role as

an engineer wasn't just write code and

shut up. It was very in the weeds of

like working with and being involved

with the people who were using the stuff

that you did. Being that in the weeds is

something I've never been able to give

up. I like talking to my users. I like

being one of the users. I like being

deeply ingrained and involved with the

people that I'm building things for.

I'll blast through the rest quickly.

Eventually, this org kind of fell apart

cuz it was a blend of premium content

contracts and helping non-gamers

succeed. So, it got split and folded

into VOD, which is the video on demand

platform, which then got folded and I

got thrown overboard and I landed in the

safety org, which was very, very

formative for me as well. The big thing

with me switching to the safety org is

that I was no longer working with Wobba,

which meant I no longer got to use

Elixir every day. I had to learn Go. I

hated it. I got into front end instead.

Started to like it a bit cuz this was

also right when we were starting the

React rewrite of the Twitch site from

Ember. And I was just so impressed with

the motivation of the team that showed

up and was like, "No, we're rewriting

the whole site from scratch." That was

the Twitch beta in 2018, if you remember

that. Learned so much from them. was so

hyped on their motivation to just come

in and do the crazy hard thing you're

not supposed to, which is at a,500

person company, rewrite everything and

succeed. It was so cool and I loved

being part of what was called the

Twilight Rebuild of Twitch. Twitch light

is the pun there, but the Twilight repo

is the still to this day the main repo

for the entire Twitch site. I got to

help a ton early on. I got to unblock

stuff, help them accelerate the launch.

Got really good at front end really

fast. Got promoted. Got to be involved

in a bunch of conversations I probably

shouldn't have been, especially as a

junior engineer. I got promoted. I got a

ton of support from my team. I helped

hire up that team from two people to

like 30. I did a shitload of interviews.

I was deep at that time. And I learned

so much. I learned how to think and

argue and be part of a like a thing

happening from this piece of my career.

But safety helped me learn all about how

to actually ship inside of like a big

moving org. How to fight for your users

and keep your platform safe. How to

deeply care about the details of the

things that you're building. How to

think full stack, not just about the

back end or the front end, but how

everything comes together. And also, I

got to build Mod View, which is one of

the things I am most proud of in my time

at Twitch. ModView is an awesome

platform. I've talked about it a bunch.

I have a whole video that's called like

the coolest thing I ever built. Still

feel strongly about it. I learned so

much about web performance and building

great stuff there. And I got to talk to

the actual moderators who were using our

tools every day. This is a theme you'll

notice. I go out of my way to build

things where I can talk to and work with

the people who are using it every day.

But here is where the first cut off I

think matters happens. 2020 In 2020,

after we shipped Mod View, I switched

teams. When I switched teams, I wanted

to take what I learned from Safety and

my other orgs that I worked in at

Twitch. I wanted to bring that to the

creator side cuz like to be totally

honest, even today, as somebody who uses

Twitch, as I'm streaming right now,

filming these videos, thank you to

everybody who's hanging out with us

there. The creator dashboard is a

significantly worse experience than Mod

View. As you can hopefully see for me

coming in here, in a lot of ways, Mod

View is a really cool, powerful

platform. The fact that I can drag any

element and drop it in the sidebar, it

stays live and active exactly where it

was. I can resize, change. There's even

a full edit view here. We went hard

making mod view as good and powerful as

possible for the creators and

specifically the moderators on Twitch.

And it was honestly a bit frustrating

that ModView had gotten so good while at

the same time the creator dashboard was

not improving and I had gotten pretty

close with the people on the creator

dashboard side. So I decided it was time

I moved to creator or I really wanted to

take the things I learned from safety

and obsessing over users and making good

user experiences and apply it in the

creator space because I'd honestly kind

of exhausted how much I could do on the

moderation side. like I'd built a great

full stack user experience from the back

to the front for moderation on Twitch

internal and external and I was out of

things to do in safety. So I decided to

move to the creator or literally the

week after mod view shipped in order to

extend the things I had learned and push

even further. This was a big mistake,

not in the sense that I shouldn't have

moved teams, but in the sense that I I

used the team move to delay a thing I'd

already been thinking about, which was

quitting. I loved the safety team. I

learned so much from them. I'm still

friends with a ton of people in that

team. I helped hire a bunch of them.

I've helped them with things since then.

I love that team. One of the reasons I

tell people to get a normal job is so

you can find a team like that that you

love cuz those connections will go

forever. Like so many of the reasons I

am successful today is the people I met

in my time at safety, including some

people deep in the Y Combinator world,

which is really helpful later on. But I

wanted to pursue solving hard fun

problems for users I cared about. And I

thought I could do that within the

company. And that was the mistake. At

that point, I should have been thinking

about quitting and I was. But I told

myself, I'll switch to this team. I'll

make some impact. I'll get promoted to

the role I should be within a year. And

if I don't, then I will finally actually

quit. This or was a bit of a disaster.

I've talked about it a bunch in other

videos. I don't want to go too deep on

the details here, but they didn't like

it when you shipped fast because it made

them look bad and they would always try

to find random holes they could poke in

the thing and it never worked. There

were a couple particular people in this

org that were terrible. There were a

couple who were great. Shout out Isaac.

Him and I, it's still crazy that we were

as close as we were cuz we were so

different. Like he was really nice and

by the books, but also cared a lot about

doing the right thing for users. So,

even though you would think that him is

like the buy the books, do it the right

way, the chillway guy, and me is the

more bombastic, here's the sledgehammer,

let's destroy the thing, that we

wouldn't get along well, but we actually

got along really well because we both

just wanted to make the right thing. So,

very thankful for him because when some

of the less good people on the team

would crash out, they would crash out on

him and it would look really bad. When

they crashed out on me, it's like, "Oh,

yeah, Theo's hard to work with or

whatever." which by the way, no one had

ever really said until that team.

But when they crashed out on Isaac, it

looked a lot worse for them. And it

showed that their motivation wasn't

keeping the codebase clean and safe.

Their motivation was looking good,

having not shipped for three months

straight. I [ __ ] you not, one of the

people who would regularly get in our

way did not merge a single PR for three

[ __ ] months.

Insane. I I've talked a lot about my

drama with that team. That's not what

I'm here for. What I'm here to say is I

was [ __ ] miserable

at this point. This is also when CO

started. So CO took away my favorite

thing, lunch and dinner. I know this

sounds like a joke, but it's not. The

thing I loved most beyond like building

real cool things for users was the lunch

and dinner conversations with engineers

who were much smarter and more

experienced than me. I benefited so much

from those conversations. I learned so

much from those developers and I would

be the fun, excited guy coming in with

the cool new things and they'd be the

realistic people showing me how it would

or more importantly wouldn't work based

on reality. I loved those conversations

and getting to just talk about things

without worrying about what level

someone is, what disclosures they have,

what they are and aren't allowed to

know, just talking about cool end stuff.

I love those conversations. And I also

got one really, really important comment

from a developer that I mostly know as

Merryweather. He was great. He helped

run the Twitch site. He told me around

this time right before co actually that

he was really thankful I was at Twitch

because it showed him that the energy

that he had joined Twitch for wasn't

gone. And I initially took that as a

massive accompliment and I still do.

What I didn't realize was that the

energy that like a startup has that an

earlystage company pushing really hard

to build the right thing for its users.

That is what he missed. That is what I

was bringing. But I was bringing it to a

company that was too big and to be frank

had lost the interest of their most

important users, the creators too much

so. And what he saw as a shining light

in a otherwise dark place was actually

me slowly getting dimmer by the day. And

it took me a while to realize that was

what was happening. But I owe him so

much both for the confidence he gave me

with that comment, but also for helping

me realize and frame the fact that I

cared more about the users than the

company did. And that's never a good

place to be in. And I lost my ability to

have lunch and dinner convos about nerdy

stuff with Merryweather, with Waba, with

all the incredible engineers on safety

and chat, moderation, and all the other

things I was involved in. And now I was

sitting at home reading code reviews,

getting blocked on things that made no

sense, and being effectively unable to

ship what my users needed. I was going

mad fast. After a couple more dramatic

things and also a rewrite of Twitch

Studio, which eventually got sunset, it

was time I accepted I couldn't do this

anymore. And early in 2021, I quit. It

was tough. One of the hardest things I

ever did. I was so unsure when I put in

my two weeks the day I finished like I

did my like goodbye call and they're so

strict with it that like my my last day

was a Friday and I had a goodbye call in

the like we use Google Meet and the

goodbye Google Meet right at 5:00 p.m.

on the dot I get disconnected because

they deactivated my account. It was

depressing but hilarious.

Very Twitch of them. And then I did

another thing that I would say was a

mistake, but it was one of those

mistakes I learned a lot from. I joined

a music startup. I don't talk about this

one too much because uh there's not too

much to say and the founder of the

company I shouldn't have joined is a bit

latigious and I don't want him to have

more reasons to come after me. Not cuz

he would win the lawsuit, but because I

don't feel like dealing with him. He,

depending on who you talk to, he

allegedly pretended to be the founder of

a different startup that shut down and

then pretended to bring it back as

though he was the original founder. This

company is allegedly on my LinkedIn. I

don't want to name them or blast this

any further. This was a really exciting

opportunity for me because I joined as a

mobile dev because I was getting deeper

into mobile dev stuff. They needed a

better Android app and a bit more

progress on the iOS app. They had an iOS

team contracting that was three people

that had shipped nothing. I built the

React Native app that worked in 3 days.

The problem was that the back end that

they also had contracted out basically

didn't work at all. And the front end

that they had for testing it worked even

worse. So I ended up building my own new

web front end, not even to be the web

client, just to let me test the mobile

app through the servers. But it very

quickly surpassed the original web

client. So it became the web client. The

Android app we had written in React

Native got better than the iOS app was.

So it became the iOS app as well. And

very quickly I owned the entire

userfacing surface area at this company

I had just joined a month before.

Learned a ton, shipped a lot, got a lot

better at running a team, got a lot

better at fighting with somebody who who

doesn't get it. This is one of the

things that was hard for me to learn

because when I was at Twitch, I was

arguing people who did get it, like Bill

and Wobba. These people were both

smarter than me. They both understood

the things we were building better than

I did, but they were also willing to

hear me out and let me challenge them in

case I could find something they had

missed. That was not the case when I was

at the music startup. My boss at the

music startup who was the founder

thought they understood engineering

better than the entire company even

though they had really not done much

engineering. It was really frustrating.

At some point the back end was getting

so untenable and the team that was

building it they were using Redux saga

in the back end. It was an actual

disaster. I did a meeting with the guy

in charge of that contracting team and

he told me that the way he thinks about

building backends is he likes to build a

giant hamster ball for his team so they

can run around and even if they hit

things they won't get hurt because the

ball will protect them. And that ball

was a bunch of bad abstractions around

Redux saga and unbelievably useless unit

tests on everything. It was like 50,000

lines of code that didn't work. I

rewrote the whole back end without much

permission. So it was 500 lines of code,

had more features, and could scale

significantly better. They blocked it

because I didn't have enough unit tests.

So I wrote 600 lines of unit tests that

had 100% code coverage and dropped that

with my twoe notice and left. So we're

now a bit into 2021. I have quit again.

So we'll do that. I quit again. And I'll

admit that at this point I'm a little

scared. I went for making at my peak at

Twitch around 475K a year, which is

still insane when I think about it, just

absurd amounts of money. But then I was

at this much smaller startup and it was

125k a year, which honestly was fine. I

wasn't spending a whole lot of money,

especially during co. But I went from

that to a much scarier number, which was

0 a year. You can see why this is

getting scary. When I quit Twitch, I was

starting at this company the Monday

after when I quit on Friday. And I was

writing code on Saturday and Sunday,

too. Like I I didn't take a break at all

between these roles. But when I quit

here, I had nothing planned. I was

starting to talk with some other

companies, but I didn't have a concrete

like what is next at all. And man,

things got interesting. I started

interviewing. I interviewed at a couple

different places. The big ones that you

guys probably know are Railway and

Retool. There's a couple other ones that

I talked to. One that's important was

Station Head, another music startup.

These interviews were essential for me.

I got an offer from Railway. It was

pretty clear I had an offer pending from

Station. I don't even remember how it

was formal or not, but like they really

wanted to hire me. Retool, I didn't, but

we had really good conversations. I

talked to the CEO. They had like 10

employees at the time. They now have

like over 500. They're a really big

company now. But I liked all of these

companies quite a bit. At the same time

though, I had a side project. Round.

The concept for Round that you're seeing

here might feel familiar. If so, there's

a reason that we'll get to. I started

Round because I was playing more with

WebRTC. I'd always been a fan of it and

was playing with it when I was at

Twitch. I wanted to see how far you

could push WebRTC to make highquality

content. And also, I was watching a lot

of collaborative content with like lots

of different streamers doing stuff

together, fighting Discord and Zoom to

make it all work. One of those creators

had said straight up in their content, I

would pay any amount of money for

something that sucks less than Discord

for doing this. I just want to be able

to drag people around in OBS and have it

all work. I thought, I think I can make

that happen. So, I built a tool to make

it easier to do those types of live

collaborations. Round.t3.gg DG was the

URL. This was a fun side project. I got

really into WebRTC building a full stack

application. This was also originally a

VIT project on Verscell with the API

directory. It became my first Nex.js

project. So this side project round got

surprisingly popular. I would like to

make sure it is clear that at this point

I was still not a content creator. I had

streamed on Twitch like two times but

that was about it. not really a creator,

but I cared a lot about creators. And I

was still burned that when I was at

Twitch, I didn't get to actually build

for creators yet. That's what I had

joined to do. I didn't get to I was

upset. I left. And now I was going to

finally do it as a side project. I just

wanted something for creators to use. I

showed a bunch of creators. One in

particular, who was actually not really

a creator, but a good friend and

co-orker of mine from when I was on the

creative team at Twitch. Gun hits me up.

Gun and I, funny enough, had almost made

a startup of our own while we were at

Twitch. We had talked about it. He

really wanted to do anime V tubing type

stuff. He was really into the anime

world. He had a lot of friends from

Japan. It was just the thing he wanted

to do. I don't give a [ __ ] about anime.

It is not my thing in the slightest. But

when I showed what I had built to Gun,

he asked for one feature. I got that

feature working and he just looks at me

and says, "What do I have to pay you to

work on this full-time because I need it

for my startup?" V Showjo V Showjo is

the org that is I, as far as I know, the

biggest VTuber organization in the West.

And Iron Mouse, who is one of the

biggest streamers in the history of

Twitch, the biggest and most successful

woman on the platform to this day. Iron

Mouse was the first ever user of Round,

otherwise known now as Ping. the

combination of Gunrun really wanting

this to go further and my talks with all

of these folks in particular with Jake

from Railway, the CEO, and Murray from

Station Head, who's become a good friend

since. Jake from Railway gave me an

awesome offer, but said, "I see where

your heart is. You need to do a

startup." Murray from Station Head said,

"I dearly want you on this team. I need

to be able to get your advice even if

you're not on this team." That all said,

"You should do a startup." And then Gun

told me he would pay whatever it takes

to get us using paying and get it ready

for what they needed. Eventually, we're

building bespoke iPad apps so he could

bring his creators out in the real world

and all sorts of other things. That

combination led to me doing the thing

never thought I would do. And in

September of 2021, I incorporated T3

Tools Inc. The plan was that we would

build a suite of creator tools focused

on live and video. Think like the Adobe

suite, but for live video. And I was so

excited to like really be diving into

the weeds and making something awesome.

Funny enough, another tangent that was

going on around here is I changed

roommates right around the same time.

I'd say like right thereish. My roommate

at the time had moved to Germany and my

friend from college who was in the city

was going through his own roommate

drama, funny enough, from people I had

introduced him to. I felt bad. I knew I

would like him a lot. I needed someone

to fill the space. So, I hit him up and

that's how Mark became my roommate. When

Mark became my roommate, I didn't know

how much money he made. And when he told

me, I felt sick because I knew how hard

he worked. I knew the stuff he was

doing. And he was making I don't know if

I should share it. I don't care. He was

making like 60k a year which is barely

livable in SF especially like living in

a high-end apartment and like the being

in the tech world. It was horrifying. So

I gave him just some examples of places

he could work and the example I gave him

was Retool. So I was like, "Oh, you do

technical support stuff for like people

who know what they're doing and you're

good at talking to engineers. Retool is

an example of a company I recently

talked to that could make sense for you.

Within two weeks of moving in and within

a week or so of me having that

conversation with him without asking for

any help or anything from me, he had

found Retool, reached out, applied,

talked to a bunch of people, and gotten

the job as one of their first ever

support engineers with no input from me

whatsoever beyond me mentioning the name

briefly in a conversation. Within two

weeks of moving in, he was working at

Retool and had more than doubled his

salary. And I realized very quickly how

insanely motivated Mark was. If you give

him an opportunity, he will run like the

[ __ ] wind to make sure he can execute

on it like I've never seen. And I have

friends of Mark in chat right now that

had no idea he was getting fleeced like

that. And also, very Mark move. He was

he killed it. So that had happened. And

also of note is that Retool was a YC

startup. So Mark went to retool. I

incorporated my company. Things ramped

up fast. Another big part of why he

incorporated. One of the most important

people I talked to to go back up here

was Bill. Bill never did a startup of

his own, but he was always close to

early stage stuff. So, I really valued

his perspective, especially as a a

father with two kids that had I don't

say took the easy path, but took the

like the family focus on your life and

the people around you path, but clearly

also cared a lot about the startup

world, the hardworking world and that

stuff, too. So, I used him as kind of

like my my ground for doing this. And

when he told me it was time, I need to

do a startup, that helped a lot. It was

the combination of having the two

startups I talked to that I like the

most tell me it's time. Gun tell me that

he needs the product and Bill telling me

this is the last chance I'll get to do

it. I should do it if I want to and he

thinks that I should. All of that

together killed my very anti- startup

bias. Like I used to make fun of startup

bros all the time. And to be fair I

still do. I just have much more ammo

when I do it. I did not think this life

was for me. I was so hesitant and you

could see my hesitation as you see the

things I did here. Instead of quitting

Twitch to build what I wanted to for

creators, I switched teams. Instead of

making my own startup, I found another

small music startup so I could build

there. Instead of committing to my

startup, I interviewed at a bunch of

different places so I could work on it

on the side. And then after everything

pointed me towards its time, I caved and

I did it. And it was one of the best

things I'd ever done for myself. I was

building way more. I was playing with

way more tech. I was making things I was

excited about. I was talking to my users

every day. We were growing like mad. The

product was hype. We had a ton of users

that were super excited. I met a bunch

of people with between September and

December. I met a bunch of people for

the first time who since then have

become like my closest friends. Like I

honestly probably met as many people

between September and December of 2021

that I'm still talking to today as

between 2017 and up to that point. When

I incorporated, I knew I needed to raise

both because I was slowly selling off my

Amazon stock just to like pay for my

apartment and pay for the infrastructure

for the product and I wanted to bring on

a designer cuz as as good as I am, this

is not a beautiful design. The image

blur is because uh the old image

component from Nex.js didn't have a good

solution for wayback machine capturing

it. So, this has there doesn't have the

high-res images from there. It's fine.

You're not missing much. Somehow is the

video, though. I'm not a designer. And

this was before I'd even like started

getting more into design. It was a bit

rough, but it worked. People were liking

it a lot. They were using it heavily,

and they would customize how it looked

in OBS directly instead. But, I needed a

designer. I knew I needed a designer. As

you guys know, I love working with good

designers. And I was referred a designer

from Jake from Railway. I really wanted

her, but she wanted a more reasonable

salary. And I couldn't pay her 150k out

of my [ __ ] pocket at the moment. As

much money as I had saved, I'd not saved

enough for that. So, I decided fine,

it's time to raise some money. So, I

raised around 300k from my rich homies.

Those rich homies included Murray from

Station Head, almost included Jake from

Railway, but he' actually given me a

stipulation of when I'm raising money,

give a hard deadline for when to pay the

safe and if they don't pay it in time,

cut them off. And then he missed the

deadline. So, I felt bad following his

own advice. Bill was one of my first

investments. Funny enough, the first

check to land was Dax. Dax had found me

in Twitter spaces just talking about dev

stuff. Found my vibe fun and

interesting. I was confident I would do

well and threw one of his first few

angel checks into what I was building

with T3 tools. Bill was check number two

and the rest were a combination of

friends from Twitch, old co-workers and

also people I knew from college who had

went other paths that had a bunch of

money from like Microsoft and Dropbox

and other companies they all worked at.

And I put together a set of like 15 to

30 people. I don't remember the exact

number honestly that threw in 10 to 20k

checks getting us up to around 300k of

money in the bank which is enough to pay

my designer for a little bit enough to

pay me enough to cover my rent. It was

clear that we needed more and it was

also clear I did not know how to talk to

investors really at all. I could talk to

my friends and convince them we were

going to do something big but I didn't

know how to explain this to an investor.

So I applied to an accelerator. I

applied to South Park Commons. I did not

get into South Park Commons. I took that

pretty much exact application and sent

it to Y Combinator. Despite the fact

that South Park Commons is very

obviously so a much less successful

accelerator, they haven't had any major

companies exit like through them. South

Park Commons denied me immediately, gave

me no feedback or anything. Y Combinator

invited me to add more information after

I passed the first set of screening and

I did. They then invited me to do a

quick screening call and they liked me.

They invited me to another interview

with Michael Cyel. And I was so lucky

that Michael Cyel took a liking to me.

Michael Cyel is one of the smartest

human beings that I have ever had the

pleasure of talking to. If you're not

familiar with him, Michael was one of

the original founders of Twitch. He was

kind of the the reasonable guy in the

team that made sure things actually

happened, that nobody killed themselves

with their overworking, that everybody

had food, that the engineers were

working on real stuff, that they weren't

going to go out of business, did the

business side of Twitch, specifically

during the Justin TV days. He was the

one who could wrangle all the chaos

between Justin Khan live streaming his

chaotic life, EMTT Sheer being a

degenerate doing nothing but playing

Starcraft all day and all of the tech

that they had to support and build to

make it all happen. Michael was the glue

that held it together. At the time he

was the CEO and managing director of Y

Combinator. He was also an active group

partner. So even though he was running

the whole thing, he also had the 15 to

20 companies of the 400ish at the time

they would do per batch that he would

take under his wing as the like ones he

would help out and focus on. I I am so

lucky that I got to learn from him and

work with him the way I did. I just

realized that there's one other

important piece I didn't cover that I

think is important context before I

explain how I actually got in. This is a

thing I've never shown before. This is

the oldest videos on my YouTube channel.

You'll see there's a couple here from

2010 and older from back when I was

posting stupid skate videos. I had one

from 2019 where I was playing with the

iPhone 11 right when it dropped. And

then in August and September, I made

these two unlisted videos that briefly

explained what I had built and what it

was for. This one was the YC pitch

because they required you submit a one

minute video that described what you

built and why you were building it. This

has never been seen by anyone other than

a couple close friends and Y Combinator

themselves. I have not seen this video

in a very long time. I will admit I am

scared to play this. I am going to

cringe so hard cuz I had no idea how to

do YouTube at the time. Howdy. I'm Theo,

founder of T3 Tools and creator of

Round. After working at Twitch for over

four years and helping build the

foundation for TTFM Labs, I'm really

excited to break off on my own in the

live content space. Right now, streaming

with your friends is way too difficult.

I want to make that easier with a simple

product that guarantees a highquality

experience, both for the creators and

for the consumers of that content. What

I have now is a simple Zoom-like web app

that allows you to embed directly into

your program of choice and include your

branding from Twitter, Twitch, YouTube,

wherever you come from as part of the

core experience. I've only shown this to

a handful of friends and the content

they've created has already reached over

half a million users. I know that we can

do something really powerful here and I

have a lot of people who are eager to

start using the product. I just don't

have the bandwidth to support them

myself. I'm a solo founder. I built

everything I have here on my own and I

only quit my job last week in order to

take this on full-time. I want to expand

the team so I can take on more users and

see where we can take this. Thank you so

much for your time. I really appreciate

it. This is the actual product demo.

Howdy. I'm Theo, founder of T3 tools and

creator of friend. After a decade of

supporting creators and over four years

of working at Twitch, I've noticed a

growing hole in the media. Oh god, I

just have to complain about that mic

placement. What was I doing? God.

Anyways, the demand for compelling live

content is exponentially growing.

YouTube and Facebook are spending

billions on content acquisition, and

Twitch has more than doubled viewership

over the past year. Despite all of this

movement, the tools creators use have

stayed largely stagnant. The number of

media creation tools focused on live

content is surprisingly small. Over 70%

of streamers on Twitch use the open-

source project OBS to stream. If they

want to add things to their stream, such

as chat, overlays, etc., They use either

Streamlabs or Stream Elements to do

such. If they want anything beyond the

simple additions, they're probably stuck

rolling their own solution for it. There

is no generic set of reliable tools for

live content creators. The gap between

the tools used by streamers and those

used by broadcast studios is massive. To

be frank, there is no Adobe suite

standard here. We think we could provide

that standard. Our first product is a

video conferencing app named Round.

We've identified multi-person podcast

style content to be one of the biggest

pain points for creators and we decided

to tackle it head-on with a Zoomstyled

app built from the ground up for

streamers. Let's take a look.

Welcome to round.t3.gg.

To start using Round, we have to sign in

with a Twitch, Twitter, or YouTube

account. This is to allow your call

experience to be identified directly by

your brand. So, when you see Ninja from

Twitch on round, you know that it's

really Ninja from Twitch. When you get

to the homepage, you'll see all of the

calls you've been invited to as well as

any people that might be in them. If you

see here, there's somebody in my room.

So, I'm going to hop in. And I know

people are going to see Prime there and

assume that was Primeen. It wasn't.

There was another streamer named Prime

Ks who was doing collaborative like

political stuff who was really

interested in doing live collabs using

Ping. Well, round by T3 tools. I did not

know who Primagen was at this point. I

was not doing content yet. I would go

live to like demo the product and like

test things and make sure it works, but

I wasn't really a streamer at all yet.

Oh, it's me. That's funny. I'll just

close my other window for now.

As you can see, the call experience is

very similar to that of Zoom or any

similar product, but there are a few

hidden nicities that make it uniquely

good for content creation. The first one

being the name plates in the corner

which are identified directly by the

account that I signed in with. So you

can know when you see Theo and the

Twitch logo in the corner that this is

really Theo from Twitch. You also have

the accent for the call which is chosen

in the settings menu. I chose this nice

light blue color. And if I shut up,

you'll see

it hides itself. So traditional mic

activity indication. The more important

features are the customization as well

as the stream focused tools, in this

case, stream mode. When I click this

button, all the UI is hidden because it

all opened up in a pop out window that

allows me to mute, unmute, hide, show, camera,

camera,

or see chat, allow people into the call,

etc. This allows me to have one specific

window. I can't do it. The the first one

wasn't that bad. Me trying to talk about

a thing without being good at talking

yet, that was a lot harder. Do I do I

have to keep playing it? Do I have to?

For context, chat is saying that I have

to let it finish

that I can capture for OBS and then

stream directly to whatever platform I

am currently live streaming on. This

platform has proven to be a very

compelling tool for many creators

already. I've seen it used for

everything from Dungeons and Dragons

streams to VTubers just playing around

to musicians that love it for its 100

millisecond or lower latency. I'm really

excited to see how far I can push Round

and what other tools I can build using

the infrastructure that I've built Round

on top of. I could see a future where

Round is the tool of choice used by

broadcast television for doing

interviews or a random streamer inviting

a person from Twitter to come on and

talk about some politics. I would love

to see all of these use cases and more

explored by Round and the technology

that we've built here and I'm ready to

expand the team to do such. Thank you so

much for your time and listening to what

we have to offer here. I hope that you

take the time to reach out. You can

reach me on Twitter at t3.gg spelled out

t3d gg or you can go to round.t3.gg

and join us on Discord or Twitter

directly from there. Thank you again so

much for your time. I really appreciate it.

it.

That hurt me. I was not a creator yet. A

lot of people seem to think I made the

company after being a YouTuber. They

just don't have the order of events

right. I was not even starting to

consider making my own content yet at

this point, but the product was real. I

had actual people using it. Between

those two videos, remember the first one

I showed was actually second. I got a

lot better at presenting myself and I

cared a lot more at that point too and

had talked to a lot of people, watched a

lot of like how to pitch all the type of

stuff and also learned how to make eye

contact with the lens. Hi. Still don't

have a teleprompter yet. I do now, but I

learned I actually hate scripting. So, I

don't script my videos. is I use the

teleprompter so I can see myself and

make sure I'm in frame. Long tangent

aside, the thing I wanted to show there

was that I had presented in a very

organic way what I was doing and why I

was building it. And you could tell that

I was much more an engineer than a

marketer. You might think otherwise now,

but I had no idea what I was doing then.

How the hell did that product get into Y

Combinator though? Specifically, how did

Michael Cybel let that product in?

Because he knew better. his phrase or

something along the lines of like we

both know it's the mouse and the cat and

mouse game, but it's what he meant is we

both know that it's the smaller part of

this industry. He clearly wanted to push

me towards like doing tools for

podcasters or something. But my answer

to this question was why I got into Y

Combinator. My answer, and I still

firmly believe this even if we didn't

build what I was thinking when I said

it. My answer was that all video is

live. It might not be live in the sense

that it is being streamed directly from

my camera to your computer when I'm

doing it. Although right now it is. As

you guys probably know, I film myself

live on Twitch. So, you see the thing

I'm filming as I film it. You can ask

questions, help me make sure I get

things right. Really fun process. But

you'll understand why I'm there in a

second because it was my opinion that

all video is live until it hits a point

where it dies. That point could be the

SD card on your camera. It's live when

I'm talking to the camera and then it

dies when it hits the SD card. And then

I have to take out the SD card, put it

in my computer, transfer the content

over. I then have to take that content

and sort through it and label it or

upload it so that my editor can have

access to it or transcode it so it's in

the right format for my editing

software. I then have to open said

editing software and start trimming it

down from there. After orgasing it and

figuring out where it needs to be, all

of these spots in the process are dead

spots. There were multiple places in the

video creation pipeline where your video

dies. And if you can make more of the

pipeline live, if you can make it so

more of the steps happen as a live

synchronous process instead of having

blocking steps along the way, you will

make video production significantly more

efficient. This is why I can run a

YouTube channel that posts videos every

day and run a company that is shipping

code every day at the same time. is

because I deeply believed that live

video production pipelines could make

video production both more efficient and

higher quality. Michael really liked

that answer and that's why I got into YC

because I had a novel insight and

thought on why this thing that he knew

was doomed from the start could actually

be valuable. And he then spent the rest

of the batch convincing me, not

necessarily directly, but giving me the

things I needed to realize that building

a product for Twitch streamers was a

lost cause, an absolutely sinking ship

before it had even taken off. He

introduced me to a lot of people who

came from both the broadcast TV world

and the content creator like business

world that were all very useful

connections. I met Emory, who was the

founder of Frame.io, which sold to Adobe

for a billion and a quarter dollars. He

was an early investor as well once we

went through YC. And all of these people

had given me what I needed to realize

that creator tools were not the move.

But I was still really committed. I

really wanted to figure it out. Despite

all of this, despite how awesome Michael

was at like staring into my soul and

asking the question that's that that

haunts me but also excites me, when he

called me and told me I was in, I wasn't

sure what to say yet. I was still so

skeptical, especially because YC would

only give you 125K and take 7% of the

company. I had just raised 300K and gave

up less than 5% of the company in the

process. That seemed like a really bad

deal and I didn't really think YC was

that great or important or that I would

get a whole lot from it and I could just

go raise the money elsewhere myself. So,

I spent a lot of time slaving over it. I

asked Michael, "How long do I have to

decide?" It turns out when he called me,

which was the night of my interview, the

batch actually started like the next

day, not the whole batch, but like the

onboarding for it started the next day.

I had no idea that the timelines were

that tight, but he told me I had until

8:00 a.m. the next morning to decide and

that he would call me at 7:50 to figure

out if I was in or not. So, I spent that

entire night. I was up far too late

researching, reaching out to everyone I

could, just like going back and forth

with Mark for hours. Mark was still at

retool which was actually very helpful

because his founder the CEO of retool

was also YC founder so he had a lot of

thoughts too. The thing that pushed me

over the line is the thing I

foreshadowed about earlier. Bill and

Waba these two people were like far

opposite sides of a spectrum. They very

rarely agreed on things. I love them

both dearly. I learned so much from both

of them. It was so rare for them to have

like a a deep set agreement on a complex

thing. They always had some weird angle

that kept them from like aligning, but I

could learn so much from how they would

disagree. I realized I needed my mentors

in this moment. I hit up Bill and Wobba

separately. I did a call with Bill. As

I've mentioned, he's one of the kindest

people I've ever gotten to work with. I

like hang out with his kids and give

them like tips in Pokemon and stuff.

Like he's a really close friend to the

point of like family. He has never

expressed anything but excitement

towards the stuff I was doing. And when

I told Bill that I was considering not

taking the offer because of the

dilution, because of the percentage they

would get for the money, without

hesitation, Bill says, "I trust you to

do what is right for your company, Theo.

I just need you to know I would be very

disappointed if you said no to Y

Combinator." and he had never expressed

that type of sentiment with me ever. So

that one hit me really hard

and that's when I hit up Wobba because I

expected them to disagree adamantly and

I wanted the other perspective. I wanted

someone especially more on like the

capitalist side to say no that's a huge

percentage for not much money. What are

you doing? So when I hit up Waba I was

very much expecting that. I got back one

sentence. You'd be [ __ ] stupid to not

take the deal Theo. And that's when I

decided I have to do it because who am I

to disagree with both Bill and Wobba and

I am so thankful I listened to them and

I took that one last piece of Wobba

advice because Y Combinator was one of

the best things I ever did for myself. I

am so thankful they convinced me I got

so much out of YC. So at 7:50 in the

morning the day that the onboarding

starts, I didn't even know that part. I

just thought he picked an arbitrary time

for the deadline. He calls me at 7:50,

says, "So, are you in?" I say, "Yes." He

says, "Awesome. You need to find

co-founders. You do not want to do this

[ __ ] yourself." By the way, here's the

Zoom link. Onboarding starts in oh 2

minutes. Hangs up. I open the Zoom.

There's 500 people in it. Michael joins.

He runs it. I still did not know enough

about Y Combinator to understand the

fact that he was the one who ran it and

he'd be my group partner and he'd be

doing all these different things and

he'd be the one to call me on my cell

phone from his goddamn phone himself.

And there was one last thing I was

really excited about. As we established

earlier, one of the reasons I quit was

that co had taken my favorite thing. I

had lost my lunch and dinner nerd out

sessions. I had been desperately looking

for a way to fill that need. I didn't

get it from this startup because when I

was doing this startup I was the most

technical person and all the other

engineers didn't really want to talk

about engineering. There were one or two

who did but we were all so busy we never

got to hang out and we were all remote

anyway so we didn't get that like casual chitchat.

chitchat.

I did get to do a lot of mentorship

there though which was very helpful for

me as I figured out what I wanted to do

longer term. I made a bunch of friends

in the space, but like Jake and I could

talk a lot about tech stuff, but we were

in such different worlds, we didn't have

much to chat about. Regel I lost touch

with as soon as I got it. And my friend

from station head, Murray, was

non-technical. He was the like finance

business guy and operations guy. He

didn't know much about code at all. So,

my poor roommate Mark was taking the

brunt of my needing to nerd out. And

also, I started participating more in

Twitter spaces. It was still a very new

thing at the time. I had like under 200

followers probably. I was just using it

as a way to like nerd out about tech. I

ended up meeting some incredible people

through those Twitter spaces. I ended up

as a straightup nobody befriending

Tanner Lindsay, the creator of React

Query, and asking him about things that

he was excited enough about to ask me to

go research TRPC and Blitz.js to tell

him which ones were good and what was

good about them. And that's how I

discovered TRPC by nerding out with

Tanner in a Twitter space when he didn't

even really know what it was. He just

knew it used react query and was full

stack type safety. So he I I got so much

from those spaces. It was incredible.

That's one of my like what one of the

like highlights of my life was this

window from September to December of

2021. As I mentioned before, I met so

many of my closest friends, the people

that I talk to every day that mean the

most to me. In that window, when I got

into YC, when I raised all this money,

when Michael took me under his wing, for

some stupid goddamn reason, he saw

something in me. And when these spaces

introduced me to a much larger set of my

long-term software development heroes

and I was so excited to be in YC because

I would finally finally get to nerd out

with a bunch of fellow devs again. And

this was the most heartbreaking thing

for me. We get to December. Well, I

guess the important point is we get to

January 2022. I was the winter 22 batch.

YC starts and I am so excited to finally

get to nerd out with a bunch of other

startups. And this was the thing that

broke me. Nobody wanted to nerd out.

Everybody wanted to talk business.

Everybody wanted to talk distribution.

They knew me as the media guy. Even

though I wasn't popular yet, they knew

me as like the person who understood

distribution and media. So, they kept

asking me for help with that side.

Nobody wanted to talk about new

frameworks and JavaScript stuff. Nobody

wanted to talk about how to build

scalable solutions and technologies.

Nobody gave a single flying [ __ ] about

TRPC. The TRPC thing in particular irked

me and it was actually irking me back

around this time. I was talking about

how cool the TRPC stuff was with Nex.js

and Tailwind and the full stack

ecosystem that I was building with and

the stack I had created for it. People

didn't get it, but I made a friend,

Jacob Evans, who I still love dearly.

He's probably hanging out in chat

helping moderate and keep things safe

here. And if not, I'm sure he'll be here

soon. Jacob hosted a Twitter space every

Wednesday, Thursday, and Friday night

that I would participate in almost

always and we became really close

friends really, really fast. And he gave

me the inspiration to go as far as I've

gone with all of this. I owe Jacob for

[ __ ] ever. Jacob was a big testing

guy. He loved unit testing, endto-end

testing, integration testing, and all of

those things. He give workshops on

testing. I am not the biggest testing

guy. I think it's useful for keeping

things from regressing when you know

risk areas and surfaces, but just

writing tests for the sake of it not my

thing. I also knew that with this new

full stack way of building, the need for

those types of tests was reducing very

quickly. He had a big open- source code

base that he had that was a big pile of

like bad decisions on Firebase. And I

offered in order to prove to him that my

stack was good to rewrite that whole

project from scratch using what became

the T3 stack. I did it in two hours. I

sent him the repo. He was absolutely

blown away with it. I then asked him

where would I put the tests and he

looked through the whole thing, thought

deeply about it, and admitted that there

wasn't really anywhere that it made

sense to stuff tests in. Then we did

another one of those Twitter spaces,

ranted a bunch about how cool my stack

was, and people just didn't get it. It

was fun to talk about in an audio only

Twitter space, but people couldn't see

it without literally seeing it. I also

wanted to do more streaming so I could

better understand what creators needed.

So, I did one last really stupid thing.

I went live near the end of November

2021 to show off what I had coined as

the T3 stack. I already owned T3.gg. It

was inspired by my friend 3v verer from

Twitch. It's this the three letters

after the T. I owned T3.gg before

Tailwind or TRPC were even thoughts.

People seem to think it stands for

Tailwind or TRPC. I'll let you reccon

it. It's wrong, but go nuts. It was

meant to just be my name. Also, a pun on

tier three, which is the highest tier of

subscription on Twitch. I was still very

Twitch pilt. So, I streamed myself on

Twitch building a full stack app from

scratch to show off my stack and my way

of thinking and building. And this got a

couple hundred plays. I was really proud

of it. I was so pumped that my little

chaotic stack and way of thinking and

building had a couple hundred people who

thought it was really cool. It was

awesome. Then out of nowhere, Dan

Abramov posts on Twitter saying that he

wants to do some mock job interviews to

show people what it's like to do a job

interview with somebody that like is

established as being a good engineer,

but also isn't someone who does

traditional interviews. Like he doesn't

do the leak code type stuff at all. He

had already done a couple interviews

with other creators. Many of them I'm

friends with to this day. They're all

great creators. But they all had a flaw.

The flaw was nothing to do with who they

are or how talented they are. It was the

fact that most of them had been doing

content as their primary thing for a

while and I don't think any of them had

ever actually conducted job interviews

before and if they had it had been a

very long time. So their like skills for

delivering an interview were not great. I reached out to Dan for a couple

I reached out to Dan for a couple reasons. One was that I really wanted to

reasons. One was that I really wanted to like get closer with him. I' already

like get closer with him. I' already like made music recommendations. We had

like made music recommendations. We had chatted a little bit. I was still an

chatted a little bit. I was still an absolute nobody. Well under a thousand

absolute nobody. Well under a thousand followers on Twitter still at this

followers on Twitter still at this point. But I wanted to show what a good

point. But I wanted to show what a good technical interview that wasn't just

technical interview that wasn't just leak code looked like. I wanted to show

leak code looked like. I wanted to show off selfishly my product paying and use

off selfishly my product paying and use that for the interview. I wanted to do

that for the interview. I wanted to do more creation stuff and maybe get some

more creation stuff and maybe get some more eyes on my things. But most

more eyes on my things. But most importantly, I just wanted to showcase a

importantly, I just wanted to showcase a really good interview with the Theo

really good interview with the Theo method, like my way of thinking about

method, like my way of thinking about and doing interviews. I went as far as

and doing interviews. I went as far as publishing my interview notes that I

publishing my interview notes that I would give to candidates. They would

would give to candidates. They would always get something like this when I

always get something like this when I was doing a more traditional formal

was doing a more traditional formal interview that would break down exactly

interview that would break down exactly how we thought about the interview and

how we thought about the interview and give them options for what path to take

give them options for what path to take for the interview. I have a bunch of

for the interview. I have a bunch of content breaking down how I think about

content breaking down how I think about interviews and why I do it the different

interviews and why I do it the different way that I do it. But that interview was

way that I do it. But that interview was meant more to showcase what it looked

meant more to showcase what it looked like in practice to do it. And I was

like in practice to do it. And I was really proud. I still think that video

really proud. I still think that video is one of the better examples of how to

is one of the better examples of how to do a good technical interview that I've

do a good technical interview that I've seen on the internet. I'm also lucky

seen on the internet. I'm also lucky that I'm good at taking like the like

that I'm good at taking like the like step back and analyze the thing I did

step back and analyze the thing I did and say like if this wasn't me, would I

and say like if this wasn't me, would I see the value in this thing? And I yeah,

see the value in this thing? And I yeah, I'm really proud of this interview. I

I'm really proud of this interview. I feel like if I had seen someone else do

feel like if I had seen someone else do this, I would have been blown away with

this, I would have been blown away with it and shared it to a ton of people. I

it and shared it to a ton of people. I was very proud of it. As cringe as my

was very proud of it. As cringe as my ability to talk and film and edit and

ability to talk and film and edit and all of that was, it showed what I wanted

all of that was, it showed what I wanted to show. And this did very well. Both

to show. And this did very well. Both because it's Dan Abramoff. He's popular.

because it's Dan Abramoff. He's popular. People were clicking for him. But also,

People were clicking for him. But also, it was the best of the interviews. Not

it was the best of the interviews. Not because I'm a better creator, like

because I'm a better creator, like visually and like structurally it was

visually and like structurally it was the worst, but it was the best example

the worst, but it was the best example of how to do an interview for somebody

of how to do an interview for somebody like Dan, and I think it showed Dan's

like Dan, and I think it showed Dan's skills the best by far. I was really

skills the best by far. I was really proud of this. People really liked this

proud of this. People really liked this video, and it kind of is the thing that

video, and it kind of is the thing that blew up my channel that at the time,

blew up my channel that at the time, remember, I was not taking this

remember, I was not taking this seriously at all. I was doing this

seriously at all. I was doing this because I had feelings about these

because I had feelings about these particular things and I wanted to share

particular things and I wanted to share those feelings with the world in a way

those feelings with the world in a way that was more useful so I could bring up

that was more useful so I could bring up the video in a Twitter space so they

the video in a Twitter space so they could go watch it and see what I

could go watch it and see what I actually meant. I had no idea where this

actually meant. I had no idea where this was going to go. I had also become

was going to go. I had also become pretty close with Ryan Carneato, the

pretty close with Ryan Carneato, the creator of SolidJS at this point and was

creator of SolidJS at this point and was starting to see more and more of what

starting to see more and more of what made Solid great. So my next video was

made Solid great. So my next video was me recreating the Roundest tutorial with

me recreating the Roundest tutorial with Solid. And I assumed because like these

Solid. And I assumed because like these two videos had blown up. They both got

two videos had blown up. They both got like a 100k plays pretty quick that this

like a 100k plays pretty quick that this one would be just as successful. I had

one would be just as successful. I had no idea what the the troughs of YouTube

no idea what the the troughs of YouTube were like in the ceilings and the

were like in the ceilings and the floors. I just didn't understand what

floors. I just didn't understand what being a creator was yet. I was still

being a creator was yet. I was still very new to all of this. And this is I'm

very new to all of this. And this is I'm assuming this was December or January. I

assuming this was December or January. I remember doing this around Thanksgiving

remember doing this around Thanksgiving or No, it was Christmas. Yeah, I put

or No, it was Christmas. Yeah, I put this video out Christmas Eve and my skin

this video out Christmas Eve and my skin was still terrible. This was very

was still terrible. This was very preacutane. God, that that [ __ ] with

preacutane. God, that that [ __ ] with my head so much at the time, too. I had

my head so much at the time, too. I had no idea what I was doing when this video

no idea what I was doing when this video bombed. It broke my [ __ ] heart. But

bombed. It broke my [ __ ] heart. But at the same time, I had just gone into Y

at the same time, I had just gone into Y Combinator. It was spinning up. That was

Combinator. It was spinning up. That was my focus. I went all in on YC from

my focus. I went all in on YC from January to March, which was the three

January to March, which was the three months of that window. I had played with

months of that window. I had played with the creation stuff a bit. I had had my

the creation stuff a bit. I had had my fun, but I needed to build tools, not

fun, but I needed to build tools, not make videos. And I posted basically no

make videos. And I posted basically no content from when the batch started to

content from when the batch started to when the batch ended. I would do a

when the batch ended. I would do a stream every Thursday to dog food the

stream every Thursday to dog food the product and I would always have a guest

product and I would always have a guest on those streams for my first three

on those streams for my first three months. Every stream had a guest and it

months. Every stream had a guest and it was really focused on like talking to

was really focused on like talking to them. It didn't matter. The content as

them. It didn't matter. The content as fun as it was for me and as cool as it

fun as it was for me and as cool as it was to demo ping interview content is

was to demo ping interview content is really hard to do in a way that is

really hard to do in a way that is compelling and it wasn't teaching me

compelling and it wasn't teaching me [ __ ] about YouTube or content creation.

[ __ ] about YouTube or content creation. It was more like learning how to use OBS

It was more like learning how to use OBS more than anything. And again, it just

more than anything. And again, it just wasn't my focus. I was still all in on

wasn't my focus. I was still all in on YC. Like, I had gotten into Y Combinator

YC. Like, I had gotten into Y Combinator before I had started making videos,

before I had started making videos, before I'd even really started

before I'd even really started streaming. So, I got into YC. I was

streaming. So, I got into YC. I was doing these Twitter spaces. I was

doing these Twitter spaces. I was nerding out. I made the YouTube channel

nerding out. I made the YouTube channel to better demonstrate things I couldn't

to better demonstrate things I couldn't do just talking in a Twitter space. I

do just talking in a Twitter space. I still thought Twitter would be my main

still thought Twitter would be my main platform at the time, as hilarious as

platform at the time, as hilarious as that was, maybe Twitch would help a lot,

that was, maybe Twitch would help a lot, too. But I never thought of YouTube as a

too. But I never thought of YouTube as a thing that would be core to me at all.

thing that would be core to me at all. YC starts, we go ham. We do everything

YC starts, we go ham. We do everything we can to make the product as good as

we can to make the product as good as possible. And I convinced Mark to quit

possible. And I convinced Mark to quit his job on December 31st and the batch

his job on December 31st and the batch starts January 1st. And we make it

starts January 1st. And we make it [ __ ] happen. So proud of him. So

[ __ ] happen. So proud of him. So proud of us. We built something awesome.

proud of us. We built something awesome. We won the Twitch streamer market. And

We won the Twitch streamer market. And our reward for that was like 8K a month.

our reward for that was like 8K a month. Winning half the streamers makes you no

Winning half the streamers makes you no money because streamers don't spend a

money because streamers don't spend a lot of money and they don't stick around

lot of money and they don't stick around for a long time. They go to what's novel

for a long time. They go to what's novel and then they move on. And if they can

and then they move on. And if they can just do the call in Discord, even if it

just do the call in Discord, even if it sucks to do, they will do the call in

sucks to do, they will do the call in Discord, screen capture it, and let

Discord, screen capture it, and let everything fall apart, complain a bunch

everything fall apart, complain a bunch about it, and then never try the actual

about it, and then never try the actual tool. I have a whole video about this. I

tool. I have a whole video about this. I think it's titled like why good

think it's titled like why good engineers burn out or something, but

engineers burn out or something, but it's really about the difference between

it's really about the difference between a creator adopting something versus a

a creator adopting something versus a engineer adopting something. That took

engineer adopting something. That took me a long time to learn, but by the end

me a long time to learn, but by the end of the YC batch, nobody wanted to nerd

of the YC batch, nobody wanted to nerd out with me. Our product was doing well.

out with me. Our product was doing well. Not as well as we wanted because just

Not as well as we wanted because just even though we were winning the space,

even though we were winning the space, we weren't making a lot of money from

we weren't making a lot of money from it. I had hired too many people. I think

it. I had hired too many people. I think we had like five or six employee pretty

we had like five or six employee pretty sure it was six. Yeah, we had six

sure it was six. Yeah, we had six employees at the time. We were going

employees at the time. We were going really hard. We raised a bunch of money

really hard. We raised a bunch of money at the end of the batch and I was going

at the end of the batch and I was going kind of nuts. I just wanted to [ __ ]

kind of nuts. I just wanted to [ __ ] nerd out so badly and I felt like I

nerd out so badly and I felt like I didn't really understand creators

didn't really understand creators because they would tell me they wanted

because they would tell me they wanted this thing and that they would use this

this thing and that they would use this thing and then they just didn't. And it

thing and then they just didn't. And it was so hard and confusing for me to to

was so hard and confusing for me to to work around all of that that I

work around all of that that I eventually got frustrated and kind of

eventually got frustrated and kind of just needed to get it out of me. So the

just needed to get it out of me. So the two main motivators when it came to

two main motivators when it came to April, which the spoiler, April 2022 is

April, which the spoiler, April 2022 is when I went from posting once every

when I went from posting once every month to once every week, sometimes even

month to once every week, sometimes even more. Going hard on YouTube. This is

more. Going hard on YouTube. This is around the time where I did the Clash

around the time where I did the Clash stream of Prime where I bought with him

stream of Prime where I bought with him over unit testing stuff, which obviously

over unit testing stuff, which obviously huge bump to my channel. The reason I

huge bump to my channel. The reason I did that, the reason I started making

did that, the reason I started making more content, the reason I did all of

more content, the reason I did all of this is I just needed to nerd out, to

this is I just needed to nerd out, to talk about the new framework that I

talk about the new framework that I thought was cool, to deep dive on

thought was cool, to deep dive on whatever patterns I was enjoying at

whatever patterns I was enjoying at work, to explain why I thought TRPC was

work, to explain why I thought TRPC was the coolest [ __ ] in the world, how I got

the coolest [ __ ] in the world, how I got into Nex.js and Verscell. I just wanted

into Nex.js and Verscell. I just wanted to talk about all of these things so

to talk about all of these things so badly. And I wanted to better understand

badly. And I wanted to better understand why creators were using but not using

why creators were using but not using our product and what it would mean for

our product and what it would mean for us to go allin on creators. What I never

us to go allin on creators. What I never ever would have imagined is I blew up

ever would have imagined is I blew up almost immediately.

almost immediately. We have had ups and downs as every

We have had ups and downs as every creator does, but I didn't have that

creator does, but I didn't have that awkward growing phase that a lot of

awkward growing phase that a lot of creators do where they have to spend

creators do where they have to spend years making content that nobody sees

years making content that nobody sees and then eventually it works. I somehow

and then eventually it works. I somehow got to skip all of that. And I'm so

got to skip all of that. And I'm so fortunate that that is the case cuz I

fortunate that that is the case cuz I did not have the wherewithal to push

did not have the wherewithal to push through the nobody watching my stuff

through the nobody watching my stuff phase like at all. I just wanted to have

phase like at all. I just wanted to have more senior dev conversations. To be

more senior dev conversations. To be honest, if the prime time existed,

honest, if the prime time existed, Prime's second channel that he like is

Prime's second channel that he like is now his main channel, the one where he

now his main channel, the one where he actually deep dives on things and just

actually deep dives on things and just records things on stream. If that

records things on stream. If that channel existed before this moment, I

channel existed before this moment, I probably would have just been one of the

probably would have just been one of the most active people in his community,

most active people in his community, helping however I could. But I needed a

helping however I could. But I needed a place where I could nerd out. I needed a

place where I could nerd out. I needed a community that was focused on talking

community that was focused on talking about more senior engineering topics.

about more senior engineering topics. And since it didn't exist, I made it.

And since it didn't exist, I made it. And it seems like a shitload of other

And it seems like a shitload of other people wanted it too. My honest bet at

people wanted it too. My honest bet at this point in time was that like best

this point in time was that like best case I'd get 10K subs. I'd get a couple

case I'd get 10K subs. I'd get a couple hundred maybe a few thousand views a

hundred maybe a few thousand views a video on average and that would be

video on average and that would be awesome. I never would have imagined I

awesome. I never would have imagined I hit 100k subs in my first year. Never

hit 100k subs in my first year. Never would have imagined that my videos would

would have imagined that my videos would be as successful as they are. That we'd

be as successful as they are. That we'd find this massive set of people who

find this massive set of people who would be interested in watching more

would be interested in watching more senior content. When I would tell people

senior content. When I would tell people that I made YouTube videos about

that I made YouTube videos about software, for the first year, the

software, for the first year, the response was always the exact same

response was always the exact same thing. Oh, so you make tutorials for

thing. Oh, so you make tutorials for beginners because that's what tech

beginners because that's what tech content meant at the time. And I was so

content meant at the time. And I was so proud to show it didn't have to be that.

proud to show it didn't have to be that. I'm not the biggest. I'm not the best.

I'm not the biggest. I'm not the best. I'm not some super genius that predicted

I'm not some super genius that predicted this is where everything was going. I

this is where everything was going. I just made the [ __ ] thing I wanted and

just made the [ __ ] thing I wanted and it worked. And that's a lesson that I

it worked. And that's a lesson that I got from YC. Make something people want.

got from YC. Make something people want. This is a poster that they throw

This is a poster that they throw everywhere. This is a I think I have a

everywhere. This is a I think I have a shirt that says this. It's a big Paul

shirt that says this. It's a big Paul Graham quote. This concept hit me deeply

Graham quote. This concept hit me deeply cuz I already somewhat operated this

cuz I already somewhat operated this way, but I felt like I was a rebel for

way, but I felt like I was a rebel for doing it, for looking at a product pitch

doing it, for looking at a product pitch and saying this is dumb. The people who

and saying this is dumb. The people who would use this have no reason to. It

would use this have no reason to. It doesn't make sense for them. In order to

doesn't make sense for them. In order to make something people want, you have to

make something people want, you have to understand those people and what they

understand those people and what they want. You have to know how to make

want. You have to know how to make something. You have to know what they

something. You have to know what they want first. And people start with the

want first. And people start with the making and then figure out the wants.

making and then figure out the wants. I'd even kind of fallen for that. I had

I'd even kind of fallen for that. I had started from a place of I see these

started from a place of I see these creators asking for this thing. I'm

creators asking for this thing. I'm going to build it for them. But it's

going to build it for them. But it's much easier to make something people

much easier to make something people want. If you are people, if you're

want. If you are people, if you're making something that you want deeply

making something that you want deeply enough, that you care about that solves

enough, that you care about that solves a problem you have. You will make much

a problem you have. You will make much better decisions because you're making

better decisions because you're making the thing that you want so that you can

the thing that you want so that you can use it. But then you also have to be

use it. But then you also have to be able to compromise and see what others

able to compromise and see what others want it for and find the balance there,

want it for and find the balance there, which is always a challenge. things like

which is always a challenge. things like I don't ever have an interest in

I don't ever have an interest in something like voice mode for T3 chat

something like voice mode for T3 chat because I just don't talk to my

because I just don't talk to my computer. I type to my computer. But a

computer. I type to my computer. But a lot of people like to talk to their

lot of people like to talk to their computers and to their phones. So I have

computers and to their phones. So I have to compromise and figure out how we're

to compromise and figure out how we're going to do voice mode because even

going to do voice mode because even though it's not my thing, it is a thing

though it's not my thing, it is a thing that people want even if I'm not people

that people want even if I'm not people in this case. D3 chat itself was a thing

in this case. D3 chat itself was a thing I deeply desperately wanted cuz I hated

I deeply desperately wanted cuz I hated using the other chat apps. It's why I

using the other chat apps. It's why I made it. All of the things I have built

made it. All of the things I have built started from being something I wanted

started from being something I wanted other than Round which had been

other than Round which had been rebranded at this point to Ping by Ping

rebranded at this point to Ping by Ping Labs. I learned my lesson over this next

Labs. I learned my lesson over this next year, especially because my channel was

year, especially because my channel was working. When I realized that by making

working. When I realized that by making something I wanted on YouTube, I could

something I wanted on YouTube, I could blow up immediately. and the the harsh

blow up immediately. and the the harsh reality of like all of the creators that

reality of like all of the creators that were early adopters of Ping. There are

were early adopters of Ping. There are so many awesome people in that group

so many awesome people in that group that I still care dearly about. I was

that I still care dearly about. I was more popular than all but like four of

more popular than all but like four of our thousandish users. And that that hit

our thousandish users. And that that hit me hard. The reality that I in a month

me hard. The reality that I in a month could surpass the majority of our users.

could surpass the majority of our users. This investment in like live being the

This investment in like live being the future just isn't working. Especially

future just isn't working. Especially when I at this point realize like live

when I at this point realize like live sucks. Live streaming isn't useful

sucks. Live streaming isn't useful beyond a method for creating content and

beyond a method for creating content and hanging out with your community, but it

hanging out with your community, but it doesn't give you any growth or any

doesn't give you any growth or any reach. It's just an engagement. And the

reach. It's just an engagement. And the number of viewers I would have in my

number of viewers I would have in my streams skyrocketed when I started

streams skyrocketed when I started posting videos. And it was clear that

posting videos. And it was clear that they came from stream. But my viewership

they came from stream. But my viewership did not come from Twitch. It didn't even

did not come from Twitch. It didn't even come from Twitter. Came from compelling

come from Twitter. Came from compelling YouTube videos that brought people to

YouTube videos that brought people to me. And I learned so much so fast about

me. And I learned so much so fast about the creator world, about how attention

the creator world, about how attention works, about how people engage with

works, about how people engage with things and ideas and conversations and

things and ideas and conversations and so much more. And most importantly, I

so much more. And most importantly, I started editing my own videos. And by

started editing my own videos. And by editing my own videos and listening to

editing my own videos and listening to my voice constantly, I got way better at

my voice constantly, I got way better at speaking. As you can probably tell with

speaking. As you can probably tell with the difference between now and the

the difference between now and the videos I played earlier, hearing myself

videos I played earlier, hearing myself talk and listening and watching for

talk and listening and watching for hours made me very aware of how I spoke

hours made me very aware of how I spoke and what did and didn't work. And that

and what did and didn't work. And that made me just better at most things. It

made me just better at most things. It made me way better at having

made me way better at having conversations in person. It made me way

conversations in person. It made me way better thinking about copy and branding

better thinking about copy and branding and the way my web pages would describe

and the way my web pages would describe the products I was building. It made me

the products I was building. It made me way better at raising money because I

way better at raising money because I could talk to investors and be really

could talk to investors and be really compelling and engaging in the

compelling and engaging in the conversation. It just made me better at

conversation. It just made me better at explaining the things that go on in my

explaining the things that go on in my head. One of the best things I ever did

head. One of the best things I ever did for myself was editing my own videos for

for myself was editing my own videos for a long time from really just April

a long time from really just April onwards. And a huge huge shout out to

onwards. And a huge huge shout out to Adz who just randomly volunteered to

Adz who just randomly volunteered to edit my stuff for me for free and

edit my stuff for me for free and carried my channel in the early days and

carried my channel in the early days and wouldn't even let me pay him. He ended

wouldn't even let me pay him. He ended up getting a job at Sentry and having to

up getting a job at Sentry and having to ramp down on the editing, which is also

ramp down on the editing, which is also thankfully when I ramped up on it and

thankfully when I ramped up on it and got way better at it. But Adz was great.

got way better at it. But Adz was great. I owe him a ton. Mir came in and helped

I owe him a ton. Mir came in and helped a bunch, especially on like the shorts

a bunch, especially on like the shorts and like shorter form side. But

and like shorter form side. But eventually, as the business was ramping

eventually, as the business was ramping up and I was getting too busy to do the

up and I was getting too busy to do the editing and most importantly, I had

editing and most importantly, I had fallen in love with video editing,

fallen in love with video editing, that's when FaZe hit me up. He had

that's when FaZe hit me up. He had actually hit me up before about offering

actually hit me up before about offering to edit for me and then months later hit

to edit for me and then months later hit me up about some skateboard stuff and I

me up about some skateboard stuff and I realized, oh, you're a skater, you're a

realized, oh, you're a skater, you're a music nerd and you want to edit for me.

music nerd and you want to edit for me. There's enough overlap in those things

There's enough overlap in those things that I have a feeling we're going to

that I have a feeling we're going to vibe. I really need the help right now

vibe. I really need the help right now down to come over and help us out a bit.

down to come over and help us out a bit. and FaZe who is here in chat who has

and FaZe who is here in chat who has edited the vast majority of the videos

edited the vast majority of the videos on my channel not just recently but like

on my channel not just recently but like in general

in general very quickly started carrying the

very quickly started carrying the channel for me and when you get to edit

channel for me and when you get to edit a video a day you skill up fast he got

a video a day you skill up fast he got way better super quick a Final Cut house

way better super quick a Final Cut house we're still all editing nerds we geek

we're still all editing nerds we geek out about the stuff all the time and I

out about the stuff all the time and I think a big part of why it worked was

think a big part of why it worked was that I wasn't hiring him so that I

that I wasn't hiring him so that I wouldn't think about the work anymore I

wouldn't think about the work anymore I hired him because I couldn't spend I'm

hired him because I couldn't spend I'm doing it and I wanted to live like

doing it and I wanted to live like vicariously through him to an extent as

vicariously through him to an extent as he did the editing and he got real good

he did the editing and he got real good real fast. Huge shout out to FaZe

real fast. Huge shout out to FaZe carrying this channel to the point where

carrying this channel to the point where most of you guys are here now. But I had

most of you guys are here now. But I had to get good at editing first. I would

to get good at editing first. I would never have hired the right editor if I

never have hired the right editor if I didn't get good first and do it for too

didn't get good first and do it for too long. I I was editing for like 6 months

long. I I was editing for like 6 months longer than I probably should have. But

longer than I probably should have. But I think that was a good thing. And I

I think that was a good thing. And I give this advice to everybody hiring.

give this advice to everybody hiring. Don't hire for the thing you want to

Don't hire for the thing you want to avoid. Hire for the thing you love so

avoid. Hire for the thing you love so you'll go get good at the thing you're

you'll go get good at the thing you're avoiding. And then hire for that once

avoiding. And then hire for that once you love that, too. Hire once you'll

you love that, too. Hire once you'll miss doing the thing. And I so deeply

miss doing the thing. And I so deeply miss editing video. It was so fun. I got

miss editing video. It was so fun. I got really into it. And I wish I could do it

really into it. And I wish I could do it more. I really do. All of this to say

more. I really do. All of this to say that by the end of 2022,

that by the end of 2022, it was clear that the thing that we were

it was clear that the thing that we were building was not working. So, by January

building was not working. So, by January 2023, I had to make the actual hardest

2023, I had to make the actual hardest decision of my life. Way harder than

decision of my life. Way harder than quitting. I know I said quitting was the

quitting. I know I said quitting was the hardest earlier. I just hadn't thought

hardest earlier. I just hadn't thought it through far enough because I

it through far enough because I literally like blocked this one out from

literally like blocked this one out from my memory. I had to accept I did the

my memory. I had to accept I did the startup thing wrong. I was cosplaying as

startup thing wrong. I was cosplaying as a big company when I wasn't. I had hired

a big company when I wasn't. I had hired too many people. I'd hired a support

too many people. I'd hired a support person when we had like a thousand

person when we had like a thousand people who had signed in and used the

people who had signed in and used the platform, but we had like a hundred

platform, but we had like a hundred recurring users that were there

recurring users that were there regularly and maybe like three support

regularly and maybe like three support questions a day. I had two engineers

questions a day. I had two engineers that were very junior but very good, but

that were very junior but very good, but nowhere near enough work for them to do.

nowhere near enough work for them to do. I had a full-time designer, which we did

I had a full-time designer, which we did not need. She did an incredible job

not need. She did an incredible job making paying look as great as it did

making paying look as great as it did and giving us a very established brand

and giving us a very established brand really early. We didn't need any of

really early. We didn't need any of that. like our conversion didn't

that. like our conversion didn't meaningfully go up from when it was the

meaningfully go up from when it was the crappy branding that I had with like

crappy branding that I had with like round by T3 tools when we moved it to

round by T3 tools when we moved it to ping and had a much better website

ping and had a much better website because the thing that was valuable

because the thing that was valuable wasn't how good it looked or how nice it

wasn't how good it looked or how nice it felt it was how well it worked and how

felt it was how well it worked and how reliable the tool was but none of that

reliable the tool was but none of that mattered because we didn't make enough

mattered because we didn't make enough money to pay one of their salaries much

money to pay one of their salaries much less all of them especially after

less all of them especially after infrastructure costs because doing the

infrastructure costs because doing the low latency video was expensive as [ __ ]

low latency video was expensive as [ __ ] I had screwed up immensely and Mark and

I had screwed up immensely and Mark and I had a lot of tough conversations

I had a lot of tough conversations and decided I had to do the hard thing.

and decided I had to do the hard thing. I laid off the team. I did everything I

I laid off the team. I did everything I could to help out. I gave them all

could to help out. I gave them all generous severance. I helped them

generous severance. I helped them network and find better job

network and find better job opportunities. This was right before the

opportunities. This was right before the layoff crisis. So, I'm very thankful I

layoff crisis. So, I'm very thankful I did it when I did. If I had waited any

did it when I did. If I had waited any longer, it would have been way harder

longer, it would have been way harder for them to get gigs. I'm still friendly

for them to get gigs. I'm still friendly or close with almost everybody that was

or close with almost everybody that was affected here. Some of them are still

affected here. Some of them are still very good friends. I did everything I

very good friends. I did everything I could and more to help them succeed post

could and more to help them succeed post my mistakes because they were very much

my mistakes because they were very much my mistakes. It sucked. I had

my mistakes. It sucked. I had effectively blown a million dollars

effectively blown a million dollars paying people to to pretend we were a

paying people to to pretend we were a big company. And this is why I'm so hard

big company. And this is why I'm so hard on the other startups that do this

on the other startups that do this because as bad as it was for us, it took

because as bad as it was for us, it took me like 9 months to recognize it. It was

me like 9 months to recognize it. It was from here to here, from April to January

from here to here, from April to January that I like recognized and accepted

that I like recognized and accepted this. and in particular the realization

this. and in particular the realization that I wouldn't try new creator tools

that I wouldn't try new creator tools even if I knew they were really useful

even if I knew they were really useful because I was so exhausted coming up

because I was so exhausted coming up with new ideas for content that the idea

with new ideas for content that the idea of bringing a new tool into the

of bringing a new tool into the structure just didn't even compute in my

structure just didn't even compute in my head and if I am not willing to try new

head and if I am not willing to try new creator tools why would I expect a

creator tools why would I expect a creator to be willing that realization

creator to be willing that realization hit me like a [ __ ] truck and we

hit me like a [ __ ] truck and we decided it was time to like to rethink

decided it was time to like to rethink everything to extend the runway we had

everything to extend the runway we had the money that we still had in the bank

the money that we still had in the bank for as long as possible as we figured

for as long as possible as we figured out what was

out what was We then started working on the next

We then started working on the next product thing. We had realized that I

product thing. We had realized that I had this reach in the developer world

had this reach in the developer world through YouTube and it would be foolish

through YouTube and it would be foolish to not take some advantage of it both

to not take some advantage of it both because selling to creators is a a

because selling to creators is a a non-starter. I fully convince that to

non-starter. I fully convince that to this day. If you're building a tool

this day. If you're building a tool that's exclusive customer base is

that's exclusive customer base is creators and influencers, you are going

creators and influencers, you are going to struggle immensely to even make

to struggle immensely to even make money, much less be a scalable business.

money, much less be a scalable business. Developer tools are much easier to get

Developer tools are much easier to get people to adopt and web hook thing was

people to adopt and web hook thing was solving a real problem we had which is

solving a real problem we had which is replaying web hooks from various

replaying web hooks from various services like Stripe like Ross Agora and

services like Stripe like Ross Agora and Aly and all these other platforms we

Aly and all these other platforms we were integrating testing them locally

were integrating testing them locally during dev sucked because playing back

during dev sucked because playing back those events was non-trivial and we

those events was non-trivial and we built web hook thing to be a much easier

built web hook thing to be a much easier way to manage and play back events in

way to manage and play back events in dev. A couple people really liked it

dev. A couple people really liked it including an employee at AWS funny

including an employee at AWS funny enough but it didn't get real traction

enough but it didn't get real traction at any point. I was also doing all of my

at any point. I was also doing all of my thumbnails myself still and really

thumbnails myself still and really succeeding on YouTube and just getting

succeeding on YouTube and just getting more and more frustrated with the

more and more frustrated with the tooling around it. So, I built a tool to

tooling around it. So, I built a tool to make it slightly easier for me just

make it slightly easier for me just because I needed it. Pick Thing. Pick

because I needed it. Pick Thing. Pick Thing was meant to be a way for me to

Thing was meant to be a way for me to dump a bunch of photos of myself, remove

dump a bunch of photos of myself, remove the backgrounds, and use them for making

the backgrounds, and use them for making thumbnails. We still use it to this day.

thumbnails. We still use it to this day. You just dump a bunch of files, removes

You just dump a bunch of files, removes the backgrounds. But this version of P

the backgrounds. But this version of P thing, the one we're in here is the

thing, the one we're in here is the third full from scratch rewrite on a new

third full from scratch rewrite on a new repo. Because what had happened that was

repo. Because what had happened that was really interesting is while we were

really interesting is while we were building P thing, I once again got super

building P thing, I once again got super frustrated with the S3 setup process. So

frustrated with the S3 setup process. So Mark and I put a bunch of time into

Mark and I put a bunch of time into building in abstraction on top of it.

building in abstraction on top of it. That became upload thing. We never

That became upload thing. We never shipped pick thing because it very

shipped pick thing because it very quickly got folded into upload thing.

quickly got folded into upload thing. The original S3 bucket for the original

The original S3 bucket for the original upload thing release was named P thing

upload thing release was named P thing because through the building of P thing,

because through the building of P thing, the dev tool that was missing became

the dev tool that was missing became very clear to us and we went all in on

very clear to us and we went all in on that and upload thing was the first

that and upload thing was the first thing I built that got real traction in

thing I built that got real traction in the dev world other than the T3 stack

the dev world other than the T3 stack and we were so proud. I still think to

and we were so proud. I still think to this day that upload thing is one of the

this day that upload thing is one of the best examples of modern DX and

best examples of modern DX and integrations in Typescript fullstack

integrations in Typescript fullstack projects. It's significantly better than

projects. It's significantly better than even like Verscell's blob in terms of

even like Verscell's blob in terms of how well it integrates and how well it

how well it integrates and how well it thinks about the parts that come

thinks about the parts that come together. And it was clear that was what

together. And it was clear that was what we would be doing next. I went hard on

we would be doing next. I went hard on upload thing. We got a bunch of content

upload thing. We got a bunch of content made about it. We had other creators

made about it. We had other creators making content about it too. Really

making content about it too. Really proud of what we did with upload thing.

proud of what we did with upload thing. And it made money. Didn't make a lot of

And it made money. Didn't make a lot of money, but it did make money. At this

money, but it did make money. At this point, it was just Mark and me, but

point, it was just Mark and me, but Julius from the community, who you guys

Julius from the community, who you guys probably know from all the crazy stuff

probably know from all the crazy stuff he's done with turbo repo and monor

he's done with turbo repo and monor repos, as well as mostly running create

repos, as well as mostly running create app. He has started contributing to

app. He has started contributing to uploading and getting more and more

uploading and getting more and more useful on the open source side. At which

useful on the open source side. At which point I said, screw it, we can eat some

point I said, screw it, we can eat some of our runway. We need Julius. And I

of our runway. We need Julius. And I started paying him part-time while he

started paying him part-time while he was in college. And now Julius is one of

was in college. And now Julius is one of the most essential parts of everything

the most essential parts of everything that we build. I that kid powers [ __ ]

that we build. I that kid powers [ __ ] everything. I didn't want to hire. I

everything. I didn't want to hire. I desperately didn't want to hire, but the

desperately didn't want to hire, but the fact that we could pay him a bit of

fact that we could pay him a bit of money as he helped us out with stuff was

money as he helped us out with stuff was really cool. We've paid him a lot since.

really cool. We've paid him a lot since. He's doing quite well, and I'm really

He's doing quite well, and I'm really pumped that he has been as helpful

pumped that he has been as helpful throughout as he has. Even then,

throughout as he has. Even then, remember that Ping capped at around 8K

remember that Ping capped at around 8K MR, which is recurring revenue. Upload

MR, which is recurring revenue. Upload Thing peaked at around 5 12K. It had a

Thing peaked at around 5 12K. It had a much better potential longtail, but to

much better potential longtail, but to be frank, a lot of the adoption on

be frank, a lot of the adoption on upload thing was because I was a popular

upload thing was because I was a popular YouTuber. Upload thing is great. I still

YouTuber. Upload thing is great. I still really believe in the product. I think

really believe in the product. I think when you compare upload thing to the

when you compare upload thing to the other options that exist, the gap

other options that exist, the gap between upload thing and everything else

between upload thing and everything else is bigger than the gap between ping and

is bigger than the gap between ping and everything else. And I don't even say

everything else. And I don't even say it's bigger than the gap between T3 chat

it's bigger than the gap between T3 chat and everything else. It felt so much

and everything else. It felt so much better, but the problem just wasn't big

better, but the problem just wasn't big enough or cared about enough for it to

enough or cared about enough for it to make a big splash, but the adoption from

make a big splash, but the adoption from my audience was enough that we felt like

my audience was enough that we felt like it was. And we kept pushing really hard

it was. And we kept pushing really hard to make it work for big businesses, to

to make it work for big businesses, to make it enterprise ready, to make bring

make it enterprise ready, to make bring your own bucket possible. We went ham.

your own bucket possible. We went ham. We rebuilt the entire way the upload

We rebuilt the entire way the upload process worked to make it way faster.

process worked to make it way faster. When I say we, I mostly mean Julius. I

When I say we, I mostly mean Julius. I came up with the idea, but he did all

came up with the idea, but he did all the work. Really proud of what we made

the work. Really proud of what we made with upload thing. It sucks that it

with upload thing. It sucks that it doesn't matter. Eventually, and now

doesn't matter. Eventually, and now we're getting quite a bit later. We got

we're getting quite a bit later. We got to be in like April of 2024 at this

to be in like April of 2024 at this point. I am talking to creators a lot

point. I am talking to creators a lot more often now. I have a ton of thoughts

more often now. I have a ton of thoughts about like the process and the pipeline

about like the process and the pipeline for making good content. And I was

for making good content. And I was spinning up more things, not expecting

spinning up more things, not expecting any of them to make money or even be

any of them to make money or even be useful just cuz I needed them. I mean,

useful just cuz I needed them. I mean, marker thing, which is the tool that we

marker thing, which is the tool that we use for managing all of the things I

use for managing all of the things I talk about on stream and exporting the

talk about on stream and exporting the CSV so I can use it to chop the topics

CSV so I can use it to chop the topics up into the individual things I talked

up into the individual things I talked about. We wouldn't be able to make the

about. We wouldn't be able to make the videos as fast as we do without this.

videos as fast as we do without this. Really proud of what we've built with

Really proud of what we've built with marker thing. Learned a bunch with this

marker thing. Learned a bunch with this one, too. This is my first big server

one, too. This is my first big server component project. Never charged for it

component project. Never charged for it and it's fully open source. We just made

and it's fully open source. We just made it cuz we wanted it. I made P thing for

it cuz we wanted it. I made P thing for real this time because I was talking to

real this time because I was talking to these other creators and I showed them

these other creators and I showed them my workflow. They were jealous of P

my workflow. They were jealous of P thing. So I made it for the third time

thing. So I made it for the third time cuz the first time became upload thing.

cuz the first time became upload thing. The second time was my private tool.

The second time was my private tool. Third time was an actual monetizable

Third time was an actual monetizable thing with Stripe integrated and

thing with Stripe integrated and whatnot. And it did okay. It makes like

whatnot. And it did okay. It makes like 1 to 2K a month still. The margins are

1 to 2K a month still. The margins are good on it too. We were still losing

good on it too. We were still losing money. Ping T3 Tools Inc. was still not

money. Ping T3 Tools Inc. was still not a profitable business, especially

a profitable business, especially considering that we had to pay Julius

considering that we had to pay Julius and Mark real salaries and myself a fake

and Mark real salaries and myself a fake salary, very small one. We were still

salary, very small one. We were still burning money, but I was surprised that

burning money, but I was surprised that Pick Thing did as well as it did. That

Pick Thing did as well as it did. That it was making around 2K a month. That

it was making around 2K a month. That meant we didn't need to go much further

meant we didn't need to go much further to get profitable. If we could do what

to get profitable. If we could do what we did with Pick Thing like four more

we did with Pick Thing like four more times, the additional 8K a month on top

times, the additional 8K a month on top of the like 13 we were making, like it

of the like 13 we were making, like it was like 22 to 25K a month and we'd be

was like 22 to 25K a month and we'd be break even. We would no longer have to

break even. We would no longer have to worry about dying because we ran out of

worry about dying because we ran out of money. So that became my goal. I wanted

money. So that became my goal. I wanted to find other things like Pick Thing

to find other things like Pick Thing that we could build to pat out our

that we could build to pat out our revenue and our long-term like

revenue and our long-term like everything so we didn't have to worry.

everything so we didn't have to worry. We were as YC phrases it default alive.

We were as YC phrases it default alive. if we didn't grow anymore, we wouldn't

if we didn't grow anymore, we wouldn't go out of business. We were still

go out of business. We were still grinding hard on upload thing at the

grinding hard on upload thing at the time, too. And honestly, I think Pick

time, too. And honestly, I think Pick Thing was closer to the end of the year,

Thing was closer to the end of the year, if I recall, and I mostly did it because

if I recall, and I mostly did it because other creators wanted it badly. December

other creators wanted it badly. December 2024 comes

2024 comes at this point. AI is taking over

at this point. AI is taking over everything. Everyone's hyped on it. I'm

everything. Everyone's hyped on it. I'm still not super into it. I'd made the

still not super into it. I'd made the move to Cursor. I was making a lot more

move to Cursor. I was making a lot more money on YouTube. I was doing pretty

money on YouTube. I was doing pretty well, but I still wasn't that into AI.

well, but I still wasn't that into AI. But a big thing happened. Deepseek

But a big thing happened. Deepseek dropped their groundbreaking model V3.

dropped their groundbreaking model V3. Not R1, V3. V3 was super slept on

Not R1, V3. V3 was super slept on because it wasn't the fastest model and

because it wasn't the fastest model and it wasn't the smartest model, but it was

it wasn't the smartest model, but it was a really good compromise between the

a really good compromise between the speed, the performance, and the

speed, the performance, and the intelligence, and most importantly, the

intelligence, and most importantly, the price. Deepseek V3 felt like Claude, but

price. Deepseek V3 felt like Claude, but cost under a 50th or so as much as

cost under a 50th or so as much as Claude did. and it was slightly faster

Claude did. and it was slightly faster than Claude 3.5 was. And I was blown

than Claude 3.5 was. And I was blown away with the model. There was a problem

away with the model. There was a problem though. The website was useless. Not

though. The website was useless. Not like bad the same way Claude and ChatgBT

like bad the same way Claude and ChatgBT were. And I hated those websites.

were. And I hated those websites. Deepseek V3 was hosted on a site they

Deepseek V3 was hosted on a site they clearly had thrown together in a few

clearly had thrown together in a few hours at most. That was just entirely

hours at most. That was just entirely broken and sucked to use. I wanted a

broken and sucked to use. I wanted a better interface to play with Deepseek

better interface to play with Deepseek V3. And I was really frustrated with the

V3. And I was really frustrated with the state of Claude and Chat GPT. I wanted

state of Claude and Chat GPT. I wanted something that was fast and nice and

something that was fast and nice and good to use. And as you can guess, I

good to use. And as you can guess, I started Better Chat. The goal with

started Better Chat. The goal with Better Chat was to make a better chat

Better Chat was to make a better chat app to talk to Deepseek V3. By the time

app to talk to Deepseek V3. By the time I was ready to ship Better Chat, which I

I was ready to ship Better Chat, which I built almost entirely during like the

built almost entirely during like the Christmas holidays where I didn't have

Christmas holidays where I didn't have to do support or calls or any of the

to do support or calls or any of the other things I do, I just from like

other things I do, I just from like December 24th, January 1st went hard on

December 24th, January 1st went hard on it. Then Mark got back from his vacation

it. Then Mark got back from his vacation for the holidays. We went even harder

for the holidays. We went even harder getting payments and everything set up

getting payments and everything set up and then shipped it and it did way

and then shipped it and it did way better than we ever would have expected.

better than we ever would have expected. Right before we shipped, we had a couple

Right before we shipped, we had a couple decisions to make. Chat thing was taken

decisions to make. Chat thing was taken as a name and better chat was a bad

as a name and better chat was a bad name. And while I was hunting, I found

name. And while I was hunting, I found the domain T3 chat available for retail

the domain T3 chat available for retail price. Since we were based on V3 and I

price. Since we were based on V3 and I had mostly retired the T3 branding at

had mostly retired the T3 branding at that point, I thought it'd be funny to

that point, I thought it'd be funny to call it T3. So, we called it T3 chat.

call it T3. So, we called it T3 chat. But another important thing happened,

But another important thing happened, which is that V3 got way slower. As more

which is that V3 got way slower. As more people were catching on about how good

people were catching on about how good V3 was, and we were all using the

V3 was, and we were all using the official APIs, it went from like 180 to

official APIs, it went from like 180 to 200 tokens per second down to like 60 or

200 tokens per second down to like 60 or 70. And I felt awful to use. So I did a

70. And I felt awful to use. So I did a bunch of hunting. I got really deep on

bunch of hunting. I got really deep on things like artificial analysis and

things like artificial analysis and found that 40 Mini on Azure was really

found that 40 Mini on Azure was really fast. And there was one other important

fast. And there was one other important thing about that. Since we had went

thing about that. Since we had went through Y Combinator, we had a bunch of

through Y Combinator, we had a bunch of crazy deals with Google Cloud, with

crazy deals with Google Cloud, with Azure and their cloud, especially for

Azure and their cloud, especially for the AI stuff, even with AWS, even though

the AI stuff, even with AWS, even though we'd burn most of the AWS credit on

we'd burn most of the AWS credit on upload thing. So, part of why I was

upload thing. So, part of why I was building this and releasing it was, oh,

building this and releasing it was, oh, I can convert these credits we have with

I can convert these credits we have with Azure and OpenAI and Enthropic and

Azure and OpenAI and Enthropic and Google into real dollars. And I had this

Google into real dollars. And I had this giant pile of Azure money to burn. So, I

giant pile of Azure money to burn. So, I said, screw it. We'll use Azure with 40

said, screw it. We'll use Azure with 40 Mini as our default model, so it feels

Mini as our default model, so it feels really fast and nice to use. Earn some

really fast and nice to use. Earn some credits, charge eight bucks a month for

credits, charge eight bucks a month for unlimited. The $8 price point was Mark's

unlimited. The $8 price point was Mark's idea. It's one of the best ideas. I

idea. It's one of the best ideas. I really think helped a ton with our

really think helped a ton with our success. And we dropped this on a whim.

success. And we dropped this on a whim. Julius was fully maintaining upload

Julius was fully maintaining upload things, so we could go nuts on this

things, so we could go nuts on this ourselves. And it went way bigger, way

ourselves. And it went way bigger, way faster than we ever would have expected.

faster than we ever would have expected. I am still not in a position to share

I am still not in a position to share revenue numbers because I'm holding

revenue numbers because I'm holding those close because they help with

those close because they help with investor negotiations and whatnot. But

investor negotiations and whatnot. But what I will say is that it took about

what I will say is that it took about two weeks for us to be making more money

two weeks for us to be making more money from T3 Chat than we were from the rest

from T3 Chat than we were from the rest of the business combined. And very soon

of the business combined. And very soon after we were profitable. I never

after we were profitable. I never thought one product would get us there.

thought one product would get us there. I thought better chat would be one of

I thought better chat would be one of those five products that made 2 to 4K a

those five products that made 2 to 4K a month that would come back together and

month that would come back together and in the end make us enough money to

in the end make us enough money to survive. I did not think T3 Chat would

survive. I did not think T3 Chat would so quickly trounce everything we had

so quickly trounce everything we had built combined and become one of the

built combined and become one of the fastest growing companies in the valley,

fastest growing companies in the valley, much less one of the like better

much less one of the like better products we'd built. I am amazed that it

products we'd built. I am amazed that it went as far as it did and that when we

went as far as it did and that when we dropped T3 chat in January, it was

dropped T3 chat in January, it was received as well as it was. I've gotten

received as well as it was. I've gotten way deeper into AI stuff. I was actually

way deeper into AI stuff. I was actually quite frustrated cuz there were all

quite frustrated cuz there were all these AI influencers that made it seem

these AI influencers that made it seem really hard and complex. It took me like

really hard and complex. It took me like 2 months to get ahead of them just by

2 months to get ahead of them just by like reading the reports and talking to

like reading the reports and talking to people like Simon and going deep on

people like Simon and going deep on artificial analysis and talking to the

artificial analysis and talking to the people who made the things that I used.

people who made the things that I used. I feel like I got to learn so much so

I feel like I got to learn so much so fast that it frustrated me how much

fast that it frustrated me how much grifting there was in the space. And

grifting there was in the space. And that's how we got where we are today.

that's how we got where we are today. Where I can bring on Yosh as our intern

Where I can bring on Yosh as our intern for the summer. Where I can hire Julius

for the summer. Where I can hire Julius full-time and do my best to get him out

full-time and do my best to get him out here. Where I can fight in all these AI

here. Where I can fight in all these AI spaces and get flamed all the time for

spaces and get flamed all the time for thinking AI is good but also not great.

thinking AI is good but also not great. It's why I can egg on companies like

It's why I can egg on companies like Anthropic and Open AI because their UI

Anthropic and Open AI because their UI suck. It's why a lot of you are here

suck. It's why a lot of you are here today. I don't know where I'm going with

today. I don't know where I'm going with this one at this point. So, I guess the

this one at this point. So, I guess the lesson to learn is that at no point did

lesson to learn is that at no point did I correctly predict what was next for

I correctly predict what was next for us. I could only ever see what wasn't.

us. I could only ever see what wasn't. So, I tried my best to upon

So, I tried my best to upon acknowledging that a path was the wrong

acknowledging that a path was the wrong one, like acknowledging that staying at

one, like acknowledging that staying at Twitch was the wrong path and quitting,

Twitch was the wrong path and quitting, like acknowledging that this startup was

like acknowledging that this startup was terrible for me and I could do it

terrible for me and I could do it myself, like talking to these other

myself, like talking to these other companies and accepting that I could do

companies and accepting that I could do a startup when I didn't think I could.

a startup when I didn't think I could. Like, when I thought YC was the wrong

Like, when I thought YC was the wrong choice and the people most important to

choice and the people most important to me told me I was stupid.

me told me I was stupid. All of these things were me not

All of these things were me not predicting the future correctly. It was

predicting the future correctly. It was much more me accepting that my current

much more me accepting that my current path was wrong and I needed to try

path was wrong and I needed to try something else. I never thought any of

something else. I never thought any of the big things would be the big ones. I

the big things would be the big ones. I never thought any of the failed things

never thought any of the failed things would be the failed ones. But by going

would be the failed ones. But by going at it for eight years and from YC

at it for eight years and from YC starting to today, learning to deeply

starting to today, learning to deeply care about how the things I make are

care about how the things I make are used and perceived and engaged with,

used and perceived and engaged with, caring too much about the users of the

caring too much about the users of the thing and what experiences they are

thing and what experiences they are having. Just making [ __ ] that's good

having. Just making [ __ ] that's good that I am proud of. That was the key

that I am proud of. That was the key that got me where I am. And I don't know

that got me where I am. And I don't know if you can really do that at a

if you can really do that at a traditional role at a company. And I'm

traditional role at a company. And I'm not trying to convince you to go start a

not trying to convince you to go start a company. I'm not trying to convince you

company. I'm not trying to convince you that you need to predict the future and

that you need to predict the future and make a bunch of crazy bets. I'm telling

make a bunch of crazy bets. I'm telling you that this was not the path I ever

you that this was not the path I ever thought would make sense for me. But by

thought would make sense for me. But by cutting off all the other paths, I

cutting off all the other paths, I somehow ended up here. And as crazy as

somehow ended up here. And as crazy as it was, and as many mistakes as I made

it was, and as many mistakes as I made and as much pain as I felt and as much

and as much pain as I felt and as much [ __ ] as I've taken, which has been a

[ __ ] as I've taken, which has been a lot, I I hope this video doesn't get

lot, I I hope this video doesn't get flamed with shitty comments with me

flamed with shitty comments with me being this transparent, but it probably

being this transparent, but it probably will be. I know how this goes. I never

will be. I know how this goes. I never would have thought that 2020 Theo, super

would have thought that 2020 Theo, super focused on the enterprise grind,

focused on the enterprise grind, climbing the ladder, making the best

climbing the ladder, making the best product for their users, would end up

product for their users, would end up the YC bro AI dude building a product

the YC bro AI dude building a product that's a chat app, competing with

that's a chat app, competing with billion-dollar companies that it is

billion-dollar companies that it is today. But you don't get here by

today. But you don't get here by correctly predicting things. You get

correctly predicting things. You get here by making a bunch of stupid

here by making a bunch of stupid mistakes. And the biggest unlock for me

mistakes. And the biggest unlock for me was probably

was probably right here.

right here. when I no longer had a salary, when I no

when I no longer had a salary, when I no longer had a boss, most importantly, I

longer had a boss, most importantly, I no longer had a rival and I hadn't

no longer had a rival and I hadn't realized how much that was distracting

realized how much that was distracting me and getting in my way. If I knew what

me and getting in my way. If I knew what we needed to build for our users, and

we needed to build for our users, and some manager or PM or exec or reorg plan

some manager or PM or exec or reorg plan was getting in the way, I would fixate

was getting in the way, I would fixate on them and make them the enemy. I just

on them and make them the enemy. I just couldn't stop thinking about how much

couldn't stop thinking about how much better things would be if they were out

better things would be if they were out of my way. And this line here is the

of my way. And this line here is the point where that stopped because now the

point where that stopped because now the only thing in my way is me. And that was

only thing in my way is me. And that was the unlock that changed my life. Getting

the unlock that changed my life. Getting to a point where I had nobody to blame

to a point where I had nobody to blame or be angry at or fixate on as the

or be angry at or fixate on as the reason things can't be better. In

reason things can't be better. In accepting I am the one who blocks

accepting I am the one who blocks things. I am the reason things aren't as

things. I am the reason things aren't as good as they could be. I'm the one who

good as they could be. I'm the one who makes the wrong decisions and has to

makes the wrong decisions and has to suffer the consequences. And also, sadly

suffer the consequences. And also, sadly enough, my team had to suffer as well.

enough, my team had to suffer as well. The worst thing I've ever had to do. But

The worst thing I've ever had to do. But those were my mistakes. And I had to get

those were my mistakes. And I had to get over my desire to blame those in my way.

over my desire to blame those in my way. And the only way I could do that is by

And the only way I could do that is by getting them out of my way. That was the

getting them out of my way. That was the moment where everything changed. And in

moment where everything changed. And in retrospect, I think success was

retrospect, I think success was inevitable at that point as long as I

inevitable at that point as long as I didn't burn out or grow to hate myself.

didn't burn out or grow to hate myself. I can't believe where I'm at now, where

I can't believe where I'm at now, where I went from the startup hater that was

I went from the startup hater that was grinding enterprise to investing in like

grinding enterprise to investing in like a fourth of the YC batch and grinding

a fourth of the YC batch and grinding hard on a startup with two and a half,

hard on a startup with two and a half, three people now and Yash with a fourth.

three people now and Yash with a fourth. The fact that my team is smaller than it

The fact that my team is smaller than it was when I left YC. It's smaller than I

was when I left YC. It's smaller than I was when I left the team I formed at

was when I left the team I formed at Twitch by an order of magnitude. and

Twitch by an order of magnitude. and we're building bigger, better, more

we're building bigger, better, more successful things that have more users

successful things that have more users that care more. None of these paths

that care more. None of these paths would have made any sense to me if you

would have made any sense to me if you laid them out to me when I started. I

laid them out to me when I started. I would have laughed at you if you told me

would have laughed at you if you told me how I would get where I am today. But

how I would get where I am today. But I'm proud as [ __ ] of it. And I'm proud

I'm proud as [ __ ] of it. And I'm proud as [ __ ] of all the people who rode with

as [ __ ] of all the people who rode with me here, who let me do this, who trusted

me here, who let me do this, who trusted me to figure out how to make it work. My

me to figure out how to make it work. My friends who invested on a thing that

friends who invested on a thing that made no sense at all and waited for

made no sense at all and waited for years for it to make any sense at all.

years for it to make any sense at all. for the community that's had my back,

for the community that's had my back, for each and every one of you guys

for each and every one of you guys hanging out in chat and watching on

hanging out in chat and watching on Twitch, and also watching this video

Twitch, and also watching this video when it's done. To my team, both for

when it's done. To my team, both for building T3 tools and for keeping this

building T3 tools and for keeping this channel afloat. I didn't go into this

channel afloat. I didn't go into this planning to have three jobs. Certainly

planning to have three jobs. Certainly not planning to be famous in our space.

not planning to be famous in our space. I just rode the [ __ ] wave. And here

I just rode the [ __ ] wave. And here the [ __ ] I am now. So, take this as you

the [ __ ] I am now. So, take this as you will. I got nothing else. I hope this

will. I got nothing else. I hope this helps somebody. At the very least, it

helps somebody. At the very least, it helps me get this all off my chest. It's

helps me get this all off my chest. It's nice to have a reference to point at

nice to have a reference to point at when people wonder how the hell I got

when people wonder how the hell I got here. And hopefully people will stop

here. And hopefully people will stop assuming that I made a startup to cash

assuming that I made a startup to cash out on my YouTube now because that is

out on my YouTube now because that is not how any of this [ __ ] went. That's

not how any of this [ __ ] went. That's all I got. Let me know what you guys

all I got. Let me know what you guys think. Until next time.

Click on any text or timestamp to jump to that moment in the video

Most transcripts ready in under 5 seconds

One-Click Copy125+ LanguagesSearch ContentJump to Timestamps

Paste YouTube URL

Enter any YouTube video link to get the full transcript

Most transcripts ready in under 5 seconds

Get Our Chrome Extension

Get transcripts instantly without leaving YouTube. Install our Chrome extension for one-click access to any video's transcript directly on the watch page.

Add to Chrome — Free

Works with YouTube, Coursera, Udemy and more educational platforms

Get Instant Transcripts: Just Edit the Domain in Your Address Bar!

YouTube

←

→

↻

https://www.youtube.com/watch?v=UF8uR6Z6KLc

YoutubeToText

←

→

↻

https://youtubetotext.net/watch?v=UF8uR6Z6KLc

YouTube TranscriptPreparing your results…

YouTube Transcript:The painful truth about startups (my story)