0:04

Welcome everyone. I'm

0:06

Jared and you are listening to

0:08

the Change Log where each and

0:10

every week we sit down with

0:12

the hackers, the leaders and the

0:14

innovators of the software world to

0:16

pick their brains, to learn from

0:18

their failures, to be inspired by

0:20

their accomplishments and to have a

0:23

lot of fun along the way.

0:25

Kendall Miller is a bubbly extrovert

0:27

who sticks his fingers in a

0:29

lot of pies. He advises tech

0:31

companies like Fusion Off, positions tech

0:33

products like Cebo and Tenser Lake,

0:35

organizes tech networks like CTO Lunches,

0:37

and even sells whiskey and gin

0:39

to tech people like us via

0:41

his Friday Deployment Spirits brand. Kendall

0:43

has learned a lot since he

0:45

first entered the industry and he's

0:47

eager to share what he knows

0:49

and who he knows with the

0:51

world. But first, a quick mention

0:53

of our partners at fly .io, the

0:56

public cloud built for developers who

0:58

ship. If you ship code on

1:00

the public internet, then you owe

1:02

it to yourself to check out

1:04

fly .io. Okay, Kendall Miller on

1:06

the changelog. Let's do it. Well

1:14

friends, it's all about faster builds,

1:16

teams with faster builds, ship

1:18

faster, and win over

1:20

the competition. It's just science. And

1:22

I'm here with Kyle Galbraith, co -founder

1:24

and CEO of Depot. Okay,

1:26

so Kyle, based on the premise

1:28

that most teams want faster builds, that's

1:31

probably a truth. If they're using

1:33

CI provider for their stock configuration or

1:35

get -of -actions, are they wrong?

1:37

Are they not getting the fastest builds possible? I

1:39

would take it a step further and say if

1:41

you're using any CI provider with just the

1:43

basic things that they give you, which is if

1:45

you think about a CI provider. It

1:48

is, in essence, a

1:50

lowest common denominator generic VM.

1:52

And then you're left to your

1:54

own devices to essentially configure

1:56

that VM and configure your built

1:58

pipeline. Effectively pushing down to you,

2:01

the developer, the responsibility

2:03

of optimizing and making those builds

2:05

fast. Making them fast, making them

2:07

secure, making them cost effective, like

2:09

all pushed down to you. The

2:11

problem with modern day CI providers

2:13

is There's still a set of

2:15

features and a set of capabilities

2:17

that a CI provider could give

2:19

a developer that makes their builds

2:21

more performant out of the box,

2:23

makes their builds more cost effective

2:25

out of the box and more

2:27

secure out of the box. I

2:29

think a lot of folks adopt

2:31

GitHub Actions for its ease of

2:33

implementation and being close to where

2:35

their source code already lives inside

2:37

of GitHub. And they do care

2:40

about build performance, and they do

2:42

put in the work to optimize

2:44

those builds. But fundamentally, CI providers

2:46

today don't prioritize performance. Performance is

2:48

not a top -level entity inside

2:50

of generic CI providers. Yes.

2:52

OK, friends. Save your time.

2:54

Get faster builds with Depot, Docker

2:56

builds, faster get -of -action runners, and

2:58

distributed remote caching for Bazel, Go,

3:00

Gradle, Turbo repo, and more. Depot

3:02

is on a mission to give

3:04

you back your dev time and

3:06

help you get faster build times

3:08

with a one -line code change. Learn

3:10

more at depo .dev. Get started

3:13

with a seven -day free trial. No

3:15

credit card required. Again, depo

3:17

.dev. Today,

3:32

I'm joined by Kendall Miller, a

3:35

bubbly extrovert. So I have expectations,

3:37

Kendall. But extrovert is a service,

3:39

not just bubbly extrovert. Don't act

3:41

like I don't sell it, Jared.

3:43

OK. How do

3:45

you price such services? Well,

3:50

that's complicated because what bubbly

3:52

extrovert means at every company is

3:54

a little different. But it's

3:56

surprisingly valuable in a world of

3:58

technology nerds to have I

4:01

don't want to say an ounce of

4:04

personality. There's a lot of people with

4:06

an ounce of personality to bring a

4:08

truckload of personality is a lot. So

4:10

yes. So I imagine you walk into

4:12

a meeting and everything changes. Like there's

4:14

like before Kendall and after Kendall. Is

4:16

that fair? When people notice it

4:18

is when I take vacation and I

4:20

come back and they're like, oh, thank God,

4:22

because at least somebody's going to talk

4:24

before the meeting starts. Somebody's going

4:26

to, you know, Interrupt

4:28

occasionally and put in

4:31

something bubbly and entertaining. Yeah.

4:34

As a service. As a service to

4:36

others. It's for everybody else. It's

4:38

really, you know, I'm here to be

4:40

a blessing to the world. And

4:42

we're surrounded by software engineers and product

4:44

people and like, let's just be

4:46

honest. I love them. I

4:48

don't want to bash them. My entire world and

4:50

my entire career has been in tech basically. Well,

4:53

most of my career has been in tech. Well,

4:56

you wear many hats. One thing I

4:58

read about you that I'm sure you

5:00

wrote about yourself is that you help

5:03

builders of technical products turn them into

5:05

successful technical businesses. And so I thought

5:07

we'd start there. What's the difference?

5:09

How do you help? Why do we need help,

5:11

et cetera? Well, oh

5:13

man. So I

5:15

work with a lot of early

5:17

stage founders, almost always technical founders. I'm

5:19

not the technical founder. I have

5:21

a technical background, but it's been 20

5:23

years since I've written meaningful code.

5:25

I've never been paid to write code

5:28

and that's probably a positive thing. I

5:31

find recursion much too interesting, which as

5:33

an aside, I do believe that recursion

5:35

is our way out of the coming

5:37

AI apocalypse. We just got to get

5:39

really creative with recursion and they'll just

5:41

never be able to take over. And

5:43

get stuck. Yeah, see, it seems really

5:45

obvious to me. I

5:47

thought we just pulled a plug, but that's

5:50

my move. We just pulled a plug, but yours

5:52

is way more diabolical. The plug is getting

5:54

more and more complicated as we've created these power

5:56

sources that are portable, unfortunately. This

5:58

is true. Yeah. Anyway, so

6:00

I'm often working with a technical

6:02

founder. Technical

6:05

founders have some particular quirks

6:07

in general, and most of them

6:09

seem to be kind of

6:11

common across the industry. And we

6:13

can get into some specifics

6:15

there. But the first and most

6:17

important thing that a technical

6:19

founder does is they almost always

6:22

think they can build their

6:24

way to a successful company. And

6:26

that happens occasionally, but it

6:28

is definitely the exception because the

6:30

best products do not win. And

6:33

I'm gonna on a very large company here,

6:35

but Microsoft exists. And we know

6:38

that the best products do not

6:40

win, right? You win

6:42

because you learn how to sell.

6:44

Does having a good product help?

6:46

Absolutely. But there are terrible

6:48

products out there where the. The

6:50

founder has figured out how to sell

6:52

it and they have wildly successful businesses.

6:55

And there is an incredibly large graveyard

6:57

full of the world's best products that

6:59

nobody ever used or nobody ever sold.

7:01

Even sometimes they, they, they can get

7:03

millions of users and they can't figure

7:05

out how to sell it. So the

7:07

difference between a technical product and a

7:09

technical business, a technical product is a

7:11

hobby until you have people who are

7:14

paying you for it. Right. And so

7:16

you think sales are positioning is. what

7:18

it takes in order to take one

7:20

to the other. Because I think you're

7:22

right, but you're also crushing my soul.

7:24

You know, I just want the best

7:26

thing to win. I mean,

7:28

I could literally one of the companies

7:30

that I'm working with right now, the

7:32

founder built Nomad and Nomad over Kubernetes.

7:34

Every company I've ever spoken to that

7:37

used Nomad and used Kubernetes said Nomad

7:39

is. by far the better product. Like

7:41

night and day, the better product. And

7:43

I used to run a Kubernetes consulting

7:45

shop and helped, I think, the largest

7:47

user of Nomad migrate to Kubernetes. And

7:49

the reason they moved, they said, Nomad's

7:51

better in every way. We hate that

7:53

we have to do this. But when

7:55

we get online and we Google for,

7:57

you know, has anybody ever run into

7:59

XYZ edge case? There's nobody out there

8:01

who's ever seen it before, right? And

8:03

we're the biggest company running it at

8:05

the biggest scale. And we're sick of

8:07

that. We just want to go someplace

8:09

that everybody knows. And it's, it's kind

8:11

of the same reason everybody still uses

8:13

WordPress. It's not that it's the best

8:16

CMS. It's just that it

8:18

has incredible market share. There's, if you

8:20

Google any problem. You're going to find a

8:22

plugin that somebody's already written that already

8:24

solves your problem because it's been around for

8:26

forever and it's been wildly successful. So

8:28

yeah, it's, it's not just positioning though. You

8:30

can have fantastic positioning. It doesn't mean you're talking

8:33

to the right people. It doesn't mean your

8:35

pavement pounding. And a lot of founders are like,

8:37

they'll build something. They put it out there

8:39

and they wait. Oh, no, nobody

8:41

found it. Well, I could,

8:43

um, do some advertising, write a

8:45

few more blog posts, or I bet

8:47

if I built this other feature,

8:49

then they're gonna love it. And you

8:52

know, I mean, it's just so

8:54

much easier to build. And

8:57

I know this, I run a couple

8:59

of companies myself, and I would always

9:01

rather just keep building, because it's more fun

9:03

than go sell. It's hard to sell,

9:05

especially early on, because you're like, hey,

9:07

I know you've never heard of us, and

9:09

we have no credibility, and you don't

9:11

know anyone that... vouch for us and also

9:13

our products really good, we swear, right?

9:17

It's hard. So how do

9:19

you help these technical people? I mean,

9:21

they hire you and you help them, or

9:23

is there general advice that could apply

9:25

to people that necessarily hiring

9:28

your services, that they could walk away with

9:30

and say, okay, here's, I mean, obviously getting

9:32

out there, you said pounding the pavement. I

9:35

don't know what else you said, but. I'm sure

9:37

there are other steps in there that I could

9:39

make down to do. Like, what

9:41

are some actionable steps? I mean,

9:43

so the first thing is you've heard

9:45

it said first time founders build second

9:47

time founders sell. Or, you know, there's

9:49

there's a lot of different variations on

9:51

that, but it's a it's a common

9:53

thing. the

9:56

selling, what makes a salesperson a salesperson

9:58

is not that they've learned this trick.

10:00

Oh, if I say this and this

10:02

way, then they're gonna buy. Like there

10:04

are a few tricks like that. I've

10:06

learned a few that I've been surprised

10:08

actually are very effective, but it's not

10:10

like I'm gonna turn my close rate

10:12

from 5 % to 90 % because I

10:14

learned to be a sleazy dirtbag sales

10:16

guy, right? The... Truth

10:19

of sales is number one,

10:21

it's a numbers game. You knock

10:23

on enough doors with the

10:25

world's worst vacuum. Eventually, you

10:27

knock on a door where somebody's vacuum just

10:29

broke and they have company showing up in a

10:32

half an hour and they are stressed the

10:34

heck out and they don't care how bad your

10:36

vacuum is, they need it right now, right? So

10:38

at the end the day, it is just a

10:41

numbers game. Having a great vacuum that does a

10:43

great demo helps you sell more vacuums, but you

10:45

still gotta knock on a million doors. What

10:47

makes a salesperson a salesperson is they've

10:49

knocked on every door in the neighborhood. They

10:51

have no idea what to do next

10:53

and they wake up and keep going. And

10:56

that's not in the DNA of the

10:58

average technical co -founder or technical founder. And

11:00

so. It's a thing that they have to

11:02

learn the hard way and almost always

11:04

by the hard way without him putting in

11:07

air quotes. What that means is

11:09

they've failed once they built a, they built a

11:11

company and they couldn't get enough people to

11:13

use it or they, they reached a couple million

11:15

and they needed to reach a lot of

11:17

million because they were venture backed. Um, and then

11:19

they realize in the second round that they

11:21

have to go bigger, harder. So what do I

11:23

do? Well, so, so you asked a couple

11:25

of questions there. First of all, general advice, sell,

11:27

sell, sell, sell, sell, sell it before you

11:29

build it. Also. Build something much

11:31

much smaller than you think because

11:34

a lot of companies go build

11:36

a product this big right huge

11:38

And then they find out after

11:40

talking to everybody that the products

11:42

should go this way and 10

11:44

% of it is relevant And

11:46

it's really hard to pivot when

11:48

you've boiled the ocean already right

11:50

if you build a product that

11:52

really should be a feature And

11:54

you could build in a weekend, then

11:56

when you go out and talk to the

11:58

market and find out, oh, actually the interest is

12:00

over in this direction or I need to

12:02

build it that direction. It's not that stressful to

12:05

pivot. It's not that hard. You can listen

12:07

to the advice from customers and actually build a

12:09

thing that they want. My favorite companies are

12:11

the companies that when I talk to technical people,

12:13

they're like, why does that even exist? I

12:15

can build that in a weekend. Sure, you could

12:17

build that in a weekend. You'd build it

12:19

poorly. It would be not feature rich. And

12:22

you wouldn't be able to sell it because you have

12:24

no design sense and you don't know who to talk

12:26

to, right? But if you build

12:28

it in a weekend and then you add on, you

12:30

can add on in the right direction without being stressed

12:32

about it because you already boiled the ocean. So that's

12:34

the first thing. Build a lot smaller than you think.

12:37

Sell it a lot sooner than you think.

12:39

If you can't get somebody to pony

12:41

up money for it really early on, It's

12:43

not gonna be easier when it's bloated

12:45

and big and heavier and you've put 10

12:47

,000 times more of your opinions into it.

12:51

The other thing that I see a lot of

12:53

technical founders do is they don't do the

12:55

math. Like you have to build a business around

12:57

something that actually makes sense, right? One

13:00

technical co -founder I worked with

13:02

a long time ago built

13:04

a... a B2B SaaS product that

13:06

engineering leaders were going to

13:08

buy. And he priced it

13:10

as such that I was talking with him

13:12

saying, well, now hang on a second. You've

13:14

sold to one of the world's biggest companies.

13:17

And it's a $5 a month deal.

13:19

Well, more engineers are going to take

13:21

it up. OK, but how many of

13:23

these do we have to sell before you can

13:26

earn a living? And this is a

13:28

bootstrapped business. It wasn't even venture backed. And

13:30

he's sitting there and doing the math. guess

13:33

I gotta sell about 10 ,000 more, right?

13:35

And so like either you can sell

13:37

it for $500 a month or $1 ,000

13:39

a month or $10 ,000 a month or

13:41

whatever it's gonna take you to actually get

13:43

to the sustainability that you need within

13:45

a timeframe that makes sense or you need

13:47

to not sell it. But selling it

13:49

for $5, patting yourself on the back that

13:51

you got a success story and finding

13:53

out that you could do this for 25

13:55

years and you might be making $2

13:58

,000 a month, that doesn't work. So

14:01

there's a lot of things that people

14:03

miss. But how do

14:05

you pair these two ideas?

14:08

Idea number one is build

14:10

something small and focused and

14:12

arguably, because of those two things,

14:14

not all that valuable. And

14:16

charge $500, $1 ,000. Like

14:18

charge a lot, build

14:21

a little. It takes a

14:23

lot of confidence to do something like that. Like, well,

14:25

I got this little thing that does one thing. It's

14:27

going to cost you a lot. It seems

14:29

like one or the other, but you're

14:31

saying you can accomplish both. So

14:34

first of all, you can get people to buy something that

14:36

you haven't built. And there's

14:38

a difference in the conversation between is this

14:40

interesting to you and would you pay

14:42

for it? There's a lot of people in

14:44

tech who will say, oh, that's super

14:46

interesting. I love it. But that's different from

14:48

I would happily pay for that. So

14:51

there's a little bit of nuance

14:53

there. But how do you jive those

14:55

two things? That's the hard part

14:57

about picking a niche, right? I mean,

14:59

there's times where you need to

15:01

go boil an ocean and build say

15:03

a cloud, right? Before you're going

15:05

to have adopters, but it's not like

15:08

Amazon started with a cloud. It

15:10

started with compute, right? And then found

15:12

out all the other things. And

15:14

so. So when you're gonna go

15:16

compete with the world or build something new and interesting

15:18

in the world, I mean, there's a reason they call it

15:20

an MVP. Viable is not

15:22

the smallest I'm capable of building in

15:24

minimum viable product, right? Viable is the

15:26

smallest thing that I can get somebody

15:29

to buy is really how you should

15:31

think about that and buy at a

15:33

reasonable price. But if somebody says, hey,

15:35

you know, I really need XYZ problem

15:37

solved, you say, well, what's the first

15:39

bit of that problem that I could

15:42

solve for you today? and

15:44

you'd sign on a dotted line because it would be

15:46

valuable enough to move. And can I go build

15:48

that in a weekend? And a weekend

15:50

is a little hand wavy, but you know, maybe it

15:52

takes you a week, but if it takes you nine

15:54

months, the chances are

15:56

more likely than not you

15:58

built the wrong thing. It

16:02

seems to me

16:04

obvious that there

16:06

are unicorns out there, but

16:08

most technical people do better

16:10

if they're paired with somebody. Salesy,

16:14

design -y, extrovert -y, perhaps.

16:16

Someone like yourself. Like,

16:18

perhaps you and I could team up

16:20

and build a good business because I have

16:22

builder chops and you have sales chops

16:24

and we jive, et cetera. That

16:27

combination is really hard to find. It might be

16:29

easier to find a spouse at this point. So,

16:32

I mean, seriously, so many people.

16:34

Is there a tender for co -founders?

16:36

Probably there is. There is, yeah. Is

16:38

there? There's a

16:41

couple of systems out there for, I

16:43

mean, even Y Combinator has like

16:45

a co -founder matching setup, I

16:47

believe. And there's a couple of

16:49

different places that do something similar

16:51

to that. But I

16:53

think the hard part is finding the

16:55

bubbly extrovert who's technical enough to understand

16:58

the product at a deep enough level

17:00

to talk coherently about it. Now, for

17:02

a lot of those technical co -founders,

17:04

they never really understand it at a

17:06

deep and intellectual technical ability. They understand

17:08

it in the same way that you

17:10

learn a foreign language, right? If you're

17:13

around it long enough, you can kind

17:15

of sound like the other person, right? And

17:18

you're just kind of bs -ing,

17:20

but you start to sound

17:22

like a native if you've spoken

17:24

it long enough. It

17:28

is difficult, but what I

17:30

think is interesting is that there

17:32

are VCs that back an

17:34

awful lot of technical founders without

17:36

that business person and just

17:38

assuming that they'll figure it out.

17:40

But I mean, this morning

17:43

I have an old sales leader

17:45

that I worked with, I'm

17:47

trying to be vague here, who

17:49

is working with a technical

17:51

founder who the VCs told him.

17:53

we will back you if

17:55

you hire a sales background co

17:57

-founder because we believe that you

17:59

can make this go on

18:01

your own. So that is a thing

18:03

that does happen. And there's

18:05

currently a lot of business

18:07

people, or as I sometimes call

18:09

them, idea guys who are

18:11

drooling right now over the prospects

18:13

of vibe coding their way

18:16

without a technical person to sales

18:18

success. And I think both

18:20

those routes are probably fraught. So

18:22

how did you get into this

18:24

then? How did you as a non

18:27

-technical person or I don't know what's

18:29

your background? Were you a computer

18:31

science major? Probably not like how'd you

18:33

get in and then how'd you

18:35

eventually learn the language? So,

18:37

well, so it is complicated. So my distant

18:39

background is technical. I started writing code

18:41

in fifth grade and my first programming language

18:43

was AUK, which don't, don't start with

18:45

AUK would be my encouragement. I went from

18:47

AUK to basic. finish with AUK either.

18:49

Don't finish with AUK either. You

18:52

know, and I worked up, but I

18:54

spent the most of my time with C++. I

18:56

entered college as a sophomore almost entirely

18:58

with computer science credits that I had

19:00

taken in college classes and APs in

19:02

high school. And with the assumption that

19:05

I was going to study computer science.

19:07

But I made a decision late in

19:09

high school. I think

19:11

I like people better and computers

19:13

completely consume me. And so I

19:15

will fail at choosing between the

19:17

two. So I do have the

19:19

technical background. I worked for a

19:21

startup in 2004 in Denver that

19:23

built Twitter about five years too

19:25

early. And then

19:27

I left the field completely. I

19:29

got an English literature degree, two masters

19:31

in the history of religion. I

19:33

moved to China and worked for a

19:35

nonprofit for 10 years. You

19:38

know, so wildly different background that happened

19:40

for a while. But when I moved

19:42

back to the States about a little

19:44

over a decade ago, I was the

19:46

first hire at a small startup. Not

19:49

realizing I was the first hire because

19:51

I didn't know to ask, do you

19:53

have money and do you have people

19:55

that work for you? I know now,

19:57

but. Would

20:00

you ask them like, are you hiring? Yes,

20:02

okay. Yeah, he offered me more money

20:04

than anybody else offered and I took the job

20:06

because nobody knew what to do with me. They're looking

20:08

at my background in China going, what the heck

20:10

is this, right? So I did sales for them. We

20:13

succeeded at sales. I hired a

20:15

sales team. I hired a marketing team.

20:17

I took over operations. I actually

20:19

ended up taking over engineering when the

20:22

CTO made an engineer cry and

20:24

then eventually took over the company and

20:26

we sold the company. So it's

20:28

that experience where I did literally everything

20:30

at the company over the course

20:32

of about. eight years that put me

20:34

in a position where I could

20:36

go out. I started meeting

20:38

with other technical founders. I

20:41

ended up on the board of another

20:43

company in Denver that has grown, been

20:45

successful, had a big exit to PE

20:47

about a year and a half ago,

20:50

ended up just advising a number of different

20:52

places. And you know, I say advising, it

20:54

sounds like they're paying me an equity or

20:56

cash money. And most of the time it

20:58

was just me helping somebody out. Hey, I've

21:00

been there before. Let's talk about this. Hey,

21:02

I've got somebody to refer your way. Hey,

21:04

talk to this person, talk to that person. And

21:07

then they end up coming back. So

21:11

is it pure altruism? I mean, it

21:14

makes it sound like it's ever that clear

21:16

of a thought. No, I think it's,

21:18

I think. Networking has network effects,

21:20

right? When I moved back to the

21:22

States from China and I logged into

21:24

Slack at this new company and found

21:26

out that I was the only employee,

21:29

I panicked. And I was

21:31

like, yeah. And so I was like, well,

21:33

I better go network because I'm going to

21:35

be out of a job in no time.

21:37

And so I went out and just networked

21:39

like crazy. Who can I meet? I'm going

21:41

to be genuinely interested in anyone I can

21:43

find. And, you know, finding out that that

21:45

led to hires, you know, the first sales

21:47

hire that I brought in was that, the

21:49

first engineer that I brought in was that.

21:51

That led to all kinds of hires, that

21:53

led to sales, that led to eventual business

21:55

opportunity. I mean, a guy I hired nine

21:57

years ago is a guy that I'm in

21:59

business now with that I have opportunity to

22:01

talk about this. I make a liquor called

22:03

Friday Deployment Spirits. This is our generative A.

22:05

Rye. It's a Rye whiskey.

22:08

We also make a gin called Force Push

22:10

Gin focused on the tech world. Anyways, it's a

22:12

guy that I hired eight years ago because

22:14

I met him at a meetup. I

22:16

needed a person who did exactly what

22:18

he did. Um, we worked together great for

22:20

a little while and then years later

22:22

we ended up starting a business together. Right.

22:24

So networking has network effects. So when

22:26

I'm out meeting people and I'm genuinely interested

22:28

in getting to know them, which I

22:30

am because I find people fascinating. I find,

22:32

Hey, you need a job and I

22:34

know somebody who's hiring for you. And if

22:36

I refer you to that person and

22:38

you know, I don't take, uh, A lot

22:40

of people have asked me because I

22:42

refer a lot of people for jobs. So

22:44

a lot of people have asked me,

22:46

like, why don't you take a recruiter fee?

22:48

I would probably make a lot more

22:50

money if I did. But as soon as

22:52

I monetize the network directly, the network

22:54

loses all of its value. And

22:56

so because I've referred a person over

22:58

here, that director of engineering now owes

23:00

me a favor, right? And so now

23:02

I'm. I take a job at this

23:04

new company, and I think that

23:07

this product might be relevant to that director of

23:09

engineering. And I don't mean owes

23:11

me like, you know, it's not quid

23:13

pro quo. But when I call him and say, hey,

23:15

would you take a look at this? He's going to

23:17

say yes, because I helped him hire the last five

23:19

people he's hired, right? Actually,

23:21

I mean, it's a good friend of

23:23

mine in San Francisco who got fired from

23:26

a very big company. I actually don't

23:28

think he got fired. I think he left.

23:30

I don't know. I don't know what

23:32

happened. Very large company. And, you know, he's

23:34

one of these guys that's too big

23:36

of a big cheese, like VP at very

23:38

large organization. And I

23:40

don't nobody in my network is

23:42

the CEO of Microsoft, right?

23:45

Like, I know a lot of

23:47

people, but I don't know

23:49

those people. So

23:51

I only know one very large cloud companies

23:53

execs. And I wrote them and said, Hey,

23:55

this person came available on the market. He's

23:58

more expensive than almost any network that I

24:00

know, so you're the only place I can

24:03

refer him to. Well, they just hired him.

24:05

They just actually, I found out yesterday. So

24:08

now there's a C level executive at one

24:10

of the largest cloud companies in the world that,

24:12

you know, doesn't, I'm not saying owes me

24:14

one, but we'll definitely pick up the phone when

24:16

I call. Um, and so those

24:18

kinds of things just add up, right?

24:20

And so what ends up happening is

24:22

somebody pings me, has a question about,

24:24

Hey, I'm building XYZ. Here's why it

24:26

works. Here's why it doesn't. Can I

24:29

get some feedback? I look at it

24:31

70 to 80 % of the time.

24:33

I just. all

24:35

over their ideas. Sorry. I'm not supposed to curse

24:37

here because you're going to bleep them out.

24:39

I say bad things all about, you know, I

24:41

shut them down. I tell them it's a

24:43

terrible product. Well, I myself is, I worked for

24:45

a religious organization in China for 10 years,

24:47

Jared. And one of the things I decided when

24:49

I moved back was that I was going

24:51

to curse more and I've been wildly more successful

24:53

than I anticipated. Um, but, um,

24:55

yeah, I mean, I ended up all over

24:57

their ideas and, uh, but about 30 % of

25:00

the time or 20 % of the time

25:02

I'm like, wow, this is actually really interesting.

25:04

And here's three or four companies I think

25:06

might need that right now. And I go

25:08

make those intros. Those turn into

25:10

meaningful conversations for those companies. And that company

25:12

comes back and says, can we hire

25:14

you? You did that really fast and really

25:17

easy. And, um, you know, my

25:19

goal is not to make sales. I don't want

25:21

to be sales guy, uh, but I can go

25:23

in and get those first few product feedback conversations.

25:25

I can help you build a sales team. I

25:27

can help you build a marketing team. I can

25:29

help you understand what you're doing wrong. Um,

25:32

But even then, I mean, honestly, a huge percentage of

25:34

what these guys are doing wrong is like, they

25:36

don't know that you got to show up on time

25:38

to an interview, right? It's the

25:40

little things you got to communicate. Yeah,

25:43

because you've been an engineer in a hole

25:45

where the thing that was valued was that

25:47

you outputted code and you were really good

25:49

at that. And that's different than what

25:51

it takes to run a company and to get

25:53

people raw rod behind you and get people excited

25:55

about the vision for where you're going and give

25:57

people the freedom to go build on their own.

25:59

And I'm. talking for a

26:01

long time here as I tend to do.

26:03

No, that's fair. I'm having lots of thoughts.

26:05

One of them was, have you seen the

26:07

episode of Friends where Joey tells Phoebe that

26:09

she cannot commit a selfless deed. And so

26:11

the entire episode, she goes about trying to

26:14

have a selfless act. Cause I asked her,

26:16

is it pure altruism? Which is a stupid

26:18

question. Cause there's no such thing as pure

26:20

altruism. Cause it's like, no, but you try

26:22

to do a selfless act and then somehow

26:24

it comes back to you anyways. She

26:26

ends up carrying a baby on behalf

26:28

of somebody, but realize she goes away about

26:30

it. Yeah. I mean, she felt good

26:33

at the end the day. It's like, no,

26:35

she was very angry. I love that

26:37

particular. set up. The other

26:39

one, as you say, you don't want to be a

26:41

sales guy, but I'm trying to figure out what

26:43

kind of guy you want to be. Are you, do

26:45

you like zero to one and then move on?

26:47

Are you just trying to help out where you can?

26:49

Like part of it excites you. So

26:51

pseudo sales is always going to be a part

26:53

of my job, right? I'm always going to be

26:55

talking to people. It's always going to lead to

26:57

sales conversations. When there's a lot of

26:59

times I end up at a new company and somebody in

27:02

my network pings me and says, Hey, that's kind of interesting.

27:04

Can I talk to you about it? And it turns into

27:06

a sales conversation, right? I

27:08

don't like being sales guy because then

27:10

I have to hassle my network, you

27:12

know, check in, follow up. And so

27:14

I usually, I want to hire somebody

27:16

else that's going to do those kinds

27:18

of things. So it doesn't have to

27:20

be Kendall hassling somebody. I don't want

27:22

to hassle. I don't mind if a

27:24

sales person who works with me or for me is hassling

27:26

the hell out of one of my friends, that's their

27:28

job. And, but it doesn't, if

27:30

it's affecting my relationship with that person,

27:32

that's a problem, if that makes sense.

27:34

So what do I want to do?

27:36

I mean, That's the messiest part.

27:39

That's the hardest part for me every time

27:41

Jared is because what excites me is just

27:43

solving business problems, right? Like I like getting

27:45

in, finding out that there's things that are

27:47

a mess and just going and helping solve

27:49

it. So I mean, to give

27:51

you an example, like in the

27:53

last two days, I've written job

27:55

descriptions. I've come up

27:57

with prospecting lists. I've interviewed

27:59

people for, you know, marketing

28:02

roles for. uh, operating

28:04

rules for sales roles, um, interviewing

28:07

for a CEO right now. I'm, you

28:09

know, there's, I touch every single part

28:11

of the business cause I'm comfortable in

28:13

almost all of it. I'm not going

28:15

to go review a PR. That's the

28:17

one spot that I'm probably not going

28:19

to be useful. Although at previous company

28:21

that I was running, I would about

28:23

every year go issue a PR just

28:26

to scare all the engineers. Uh, you

28:28

know, there's always something small

28:30

like, um, a typo or something

28:32

in documentation. So

28:35

you have your spirits company, but

28:37

I'm thinking why not apply your talents

28:39

at something that you're building for

28:41

yourself or with somebody versus helping other

28:43

people? Is this one of the

28:45

things you do is building this spirits

28:47

company? So yeah, so there's

28:49

a bunch of things I do. So what I

28:52

get paid for directly, I work with

28:54

a cloud company called Sevo out of the UK.

28:56

I work with an AI company called uh,

28:58

tensor lake out of San Francisco. Um,

29:01

and then I'm on the board of

29:03

a company called fusion off. Um, that's, you

29:05

know, thinking off zero competitor. Um, but

29:07

then I run a couple of things myself.

29:09

So there's the spirits company. I have

29:11

a business partner with, uh, we make a

29:13

gin and a whiskey and we're looking

29:15

at some other interesting things that we may

29:18

do next. Uh, had, had a really

29:20

crazy opportunity come up yesterday that it's too

29:22

early to talk about, but it's exciting.

29:24

I run a network of CTOs globally called

29:26

CTO lunches. Uh, so. We have about

29:28

1600, 1700 CTOs worldwide. We put

29:30

on lunches all over the world.

29:32

So that's nice because that is a

29:35

way that builds the network. Honestly,

29:37

what I like about it is I

29:39

get to fly places on the

29:41

company dime, eat lunch on the company

29:43

dime. And now I'm in the

29:45

liquor business. So liquors, you know, a

29:47

company expense. And, you know, there's

29:49

all these different things, but CTO lunches

29:52

is another business that I own

29:54

and run. I co -own, but.

29:57

I do think that CTO lunches has

29:59

the capacity to get to, if

30:01

everybody fired me tomorrow, I think within

30:03

six months I could make enough money off

30:05

of it to make it my full -time

30:08

gig. But I like touching other things

30:10

enough that I'm not walking away. Yeah. So

30:12

how does that work? Well, come back

30:14

to that. There's one last thing that I

30:16

do also is I do have an

30:18

advisory group called Grow Big Advisors, where I've

30:20

paired together with a handful of friends

30:22

to do this sort of startup advising. They

30:24

have slightly different skill sets than me. sales

30:27

leader, one's an engineering leader. Um,

30:30

one's got the marketing background. So, uh,

30:32

it's different, different things, but in exchange

30:34

for a small bit of equity, we

30:36

provide this kind of just advisory rather

30:38

than hands -on. So that's a little different.

30:41

So those are the things that I'm

30:43

building for me. Um, so

30:45

you have a lot of stuff going on and you're

30:47

not just helping other people with their business. stick my fingers

30:49

in a lot of pies. Yes. Well,

31:02

friends, you know, I love Notion.

31:04

It is the number one tool

31:06

in my tool chest. It is

31:08

a daily driver for me. I'm

31:11

using Notion constantly organizing everything I

31:13

possibly can with Notion. And I

31:15

just love it. Of course, there's

31:17

other specialized tools out there that

31:19

have you jumping from app to

31:21

app to app. But I find

31:23

that Notion lets me do most

31:26

of everything I need to and

31:28

collaboratively with the rest of my

31:30

team and even extended team. And

31:32

now it has AI built right

31:34

in. I can ask Notion AI,

31:37

hey, can you help me build this

31:39

new page for this new purpose? And

31:42

it goes and builds it.

31:44

So I don't have to go

31:46

and build each independent new

31:48

workflow by hand. Notion AI

31:50

is there for me. It

31:52

helps me find things, solve things,

31:55

write things, draft things,

31:57

build things. It's

31:59

amazing. So yes, Notion combines

32:01

your docs, your notes, your projects

32:03

all in one place for

32:05

you and your team to

32:07

collaborate on. And the fully

32:09

integrated Notion AI suite lets you

32:11

work faster, smarter, easier, better,

32:13

all the things. And I

32:16

know because I use it every single

32:18

day. You can try Notion today for free

32:20

when you go to Notion. So

32:41

why the

32:44

spirits pie and

32:46

So we

32:49

were I had just

32:51

come back into town after some time away

32:53

and I was meeting with four old friends

32:55

and we were sitting around at a brewery

32:57

talking about, why can't I buy a

32:59

bottle of whiskey called, I did sock two

33:01

and all I got was this bottle of

33:03

whiskey. The list

33:05

of names was immediately

33:07

very long. Rishi,

33:12

my business partner, and I

33:14

stuck around after the other two guys left and Rishi goes, well,

33:16

Kendall, I think this is... a good idea. And I said, well,

33:18

Rishi, I think this is a good idea. And he goes, well,

33:20

I'm not messing around. I said, well, I'm not messing around. I

33:22

said, well, I kind of think we should do it. I said,

33:24

well, I kind of think we should do it. And the next

33:26

day I was on the phone with a distillery saying, hey, how

33:29

do I buy a whole bunch of liquor if I

33:31

want to do this? And we ended up in a

33:33

partnership and it's, it's worked out. So

33:35

do you think like targeting. So

33:37

I've noticed this as a trend,

33:39

which is why I think it's a

33:41

probably a good business and be

33:43

a relatively easy one to get into

33:45

is, you know, every celebrity has

33:47

their tequila or their. whiskey or their

33:49

gin named after them or whatever

33:51

it is. It's like branded alcohol for

33:53

this celebrity. Aviation jet. I'm

33:56

Ryan Reynolds. Thank you, Jared. I didn't want

33:58

to make the connection, but since you

34:00

did, we can also talk

34:02

about my football team in a little bit.

34:04

But anyway, sorry, keep going. No, so

34:06

I just have noticed that. I'm like, well,

34:08

this must be lucrative and relatively easy

34:10

because you're basically just branding, aren't you? You're

34:12

not, you're not opening a distillery. And

34:15

you just said you went and bought a

34:17

bunch of whiskey, right? Yeah, well, so

34:19

the whiskey and the gin are made for

34:21

us. It's not white labeled something else.

34:23

So a lot of places will go, man,

34:26

I don't know if you want to get me talking

34:28

about liquor, but we can talk about liquor. So

34:30

the vast majority of distilleries are buying

34:32

a base liquor from Indiana, because that's

34:34

where a huge percentage of the distilleries

34:37

are for hashtag reasons. The

34:39

alcohol laws in America will blow your

34:41

mind that are left over from prohibition.

34:43

So anyways, this.

34:45

Our distillery makes it from grain.

34:48

So it's not a base spirit. That's

34:50

just distilled again, nine more times like

34:52

Tito's does. And then Bragg's about it

34:54

nine times distilled. How bad was what

34:56

you started with? You had to distill

34:58

it nine times. And

35:00

it's distilled once because it's

35:02

a fantastic product. OK. Anyways, and

35:05

it's made for us. It's

35:07

a recipe for us. So it's

35:09

different. It's contract distilling rather

35:11

than just white labeling something else.

35:14

But, uh, but yes, that it's, I

35:16

did not open a distillery. I do

35:18

not want to open a distillery. I

35:21

just handle once it's in the bottle.

35:23

Uh, and I, well, I have to

35:25

buy the bottle, but in distributions, a

35:27

gigantic pain in the butt because, uh, because

35:29

of American liquor, because of American

35:31

laws. So, so I'm thinking then like

35:33

your, I mean, of course I'm

35:35

target market, right? Cause I'm a nerd

35:38

and like. generative ary, like

35:40

lands perfectly on me. But

35:42

there's not very many. also really

35:44

good whiskey. Well, that's fine. That's all

35:46

well and good. But the naming

35:48

is what I care about. Like you're

35:50

you're targeting alcohol at technical people,

35:52

basically. Isn't that like a small cross

35:54

section of the world? Are you? Well,

35:57

that's why one bottle is $1400.

36:00

Oh, okay. No, I'm kidding. You just

36:02

sell it to your CTO lunches. That's

36:04

your entire market. The whiskey

36:06

is 125 and the gin is 95.

36:08

So it's not cheap. It is a

36:10

luxury item. It is a very, very

36:12

good item. So

36:15

let me be clear about a couple of things. One, I

36:18

would have totally done this exact same

36:20

thing with a terrible distillery and a

36:22

terrible end product because I think the

36:24

marketing has legs. We have

36:26

a fantastic product that I can't believe how good

36:28

the product is. I actually think the gin

36:30

is the best gin I've ever had, and I

36:32

am a gin snob. And I don't feel

36:34

like I can say that about my own gin,

36:36

but it is so very good. And I

36:38

can say it because I didn't make it. I

36:41

had some input into the recipe, but that's

36:43

all. And the whiskey is

36:45

also very, very good. So it is a

36:47

luxury product, and it costs a lot

36:49

of money. And I'm selling it to a

36:51

group of people who have a lot

36:53

of disposable income, like to make jokes like

36:55

this. You know, my thought is not

36:57

that Jared will become a daily drinker of

36:59

generative AI. My thought is sometime in

37:01

the next year, Jared is going to want

37:03

to give a gift to somebody he

37:05

knows in tech and he's going to ship

37:07

a bottle of generative AI or, you

37:09

know, Um, director of engineering at

37:11

XYZ company is going to buy his team

37:13

all a bottle of whiskey for Christmas.

37:15

And, uh, they're going to buy generative a

37:18

right. That's, that's my goal is that

37:20

it's not, I don't think anybody's going to

37:22

buy a $95 bottle of gin. And,

37:24

uh, even though it's a very pretty bottle

37:26

of gin and it says force push

37:28

right there on the label and it's really,

37:30

it's Friday deployment spirits, Jared. It's funny. Um,

37:32

I still don't think it's going

37:34

to be anybody daily, anybody's daily drinker.

37:37

Gotcha. Well, I just wanted to

37:39

tee up your sales pitch and see how

37:41

you, how you do. How did I do?

37:43

Can you give me a rating? I would

37:45

drink it. I mean, I would, I don't

37:48

know if I would, uh, 1400 bucks is,

37:50

that's a call order. I'm no, I'm not

37:52

a CTO of anything though. It's not 1400

37:54

bucks. I want to say again, 125 bucks

37:56

versus 95 bucks. Oh, I'm sorry. What'd you

37:58

say was 14? No, I started at 1400

38:00

bucks to screw with you so that the

38:02

125 bucks that it actually costs would, uh,

38:04

stress you out less. See, talking about sales

38:07

tactic. Yeah. 100%. So sell me some

38:09

of your other, I

38:11

won't say use car salesman tactics, but just

38:13

what are the winners where you can just say

38:15

something and it'll work? You said you got

38:17

a couple of those. Yeah,

38:19

so a couple of things that I think

38:21

are really interesting. If you're

38:23

selling to technical people, don't

38:25

capitalize their name when you send the email.

38:28

Make some spelling mistakes. Yeah.

38:30

Otherwise it looks like it's a, it's a

38:32

mail merge, right? And if you, if you

38:34

can just do the littlest thing to make

38:36

it look like this was actually written by

38:38

a human. And if you can write it

38:40

by a human, it's even better, but you

38:42

misspell somebody's name. You know how much more

38:44

likely they are to respond to you because

38:46

they're like, Oh, this, this ass hat didn't

38:48

just plug it into a CRM and send

38:50

me in a mail merge, right? They're actually

38:52

reaching out to me. Don't pay attention to

38:54

the details. 100 % and sometimes they

38:57

respond with like a dude, come on, you

38:59

could do better, but they responded. So

39:01

first of all, there's little things like

39:03

that, even if it's just lowercase the

39:05

name, make a spelling mistake

39:08

in your message so it doesn't look like

39:10

it was sent by a machine. Nobody

39:12

wants the machine outreach. That's number one. Number

39:15

two, one of the

39:17

most interesting things is every

39:19

company needs to know, where

39:22

do I start? Founders

39:24

always underpriced their product. I

39:26

do this myself. You

39:28

weren't willing to pay 1 ,400 bucks, but 125

39:30

seemed cheap to you once I got back to

39:32

it. I probably could have sold the bottle for

39:35

200 bucks and Jared might have bought it. I

39:37

don't know how that's, I'm probably hand

39:39

waving. But founders always underpriced their product.

39:41

And one of the most important things

39:44

for... fixing that is finding out what

39:46

are people's budgets, right? So how much

39:48

am I saving you? Those kinds of

39:50

conversations, but they're really hard to have.

39:52

So there's this, uh, uh,

39:54

there's this tactic called the bucketing method. And

39:56

so what I say is Jared, I've got a,

39:58

got a nice bottle of whiskey here and

40:01

I want to put you into, it's a 2025.

40:03

You're going to like the way that it

40:05

drives and feels. good year. Uh, it was a

40:07

good year. And, um, You know, how much

40:09

you got to spend on this, Jared? You got,

40:11

you hoping to spend $5 on a bottle

40:13

of whiskey? You hoping to spend $100 on a

40:15

bottle of whiskey? Or do you have, you

40:17

know, $5 ,000 to spend on a bottle of

40:19

whiskey? And there's something about that last number being

40:21

so big, it's outrageous, that

40:23

almost everyone, almost every time will

40:26

go, well, it's not $5 ,000. I

40:28

was thinking more like 50. Okay,

40:30

great. Now I've got a ballpark

40:32

for what was in your head.

40:34

So this method and you name

40:37

a number, You name a number

40:39

twice as big, 10 times

40:41

as big. That last number has to

40:43

be so outrageous that it makes them

40:45

get mad. Uh, and then they'll almost

40:47

always come back with what was in

40:49

their head. So that's called the bucketing

40:51

method. That's an interesting thing. And then

40:53

there's one last one, which just in

40:55

general in sales, if you can ask

40:57

a question like if this was existed

40:59

and it was perfect for you to

41:01

meet your needs, you know, something like

41:03

that, what would it look like? And

41:06

now. you're gonna sell the

41:08

product to me. And if I can

41:10

sit still and shut up and let you just

41:12

sell to me the less I say in a sales

41:14

call, the better you're gonna feel about what I'm

41:16

selling, even though I haven't told you anything about it.

41:20

How do you mean I'm gonna sell it to you? Cause

41:22

I'm gonna describe what it would be and then you're

41:24

gonna say, that's what it is or what do you

41:26

mean? Yeah, I mean, it depends on what the, you can't,

41:28

you can't say, oh, well, that's exactly what it is,

41:30

but you can, you can go back with, you

41:33

know, how do I say this? Here's

41:37

a concrete example that'll be relevant to everyone

41:39

in your audience, even if they're not founders. Interviewing.

41:41

When you're interviewing, you're trying to

41:44

sell yourself, right? The

41:46

worst possible thing you can do

41:48

is go into an interview and say,

41:50

tell me what you need. And they

41:52

list 50 criteria. And then you try

41:54

to go through those 50 criteria

41:56

and say, here's why I

41:58

meet every single one of those criteria.

42:00

You will fail because you don't.

42:02

You're never exactly what they have in

42:04

mind. Never, ever, ever. Right? So

42:07

if you can say, Hey, in

42:09

your in your perfect world, what are

42:11

the problems that I would solve? Not

42:14

who exactly am I? Right?

42:17

What not? I'm going to show up on day

42:19

one. I'm going to wear this color. I'm going

42:21

to talk to this person and do this thing.

42:23

But what is the problem I'm going to solve?

42:25

Then they start selling to you the problem that

42:27

you're going to solve for them. And at

42:30

the end of the conversation, you say, well, I might

42:32

not be exactly what you have in mind, but here's how

42:34

I'm going to solve your problem anyways. And

42:36

keep it nice and short and sweet. And

42:38

you convinced them that what

42:40

you've just said meets their criteria.

42:43

Now they have just sold you to

42:45

themselves. And that sounds

42:47

super sketchy. It's really not that complicated. There's

42:49

one book on this that I encourage people

42:51

to read. It's called Spin Selling. And

42:53

spin is not like I'm going to spin

42:55

you and it's an acronym for something. But

42:58

the one useful thing they talk about is

43:00

this bit of like get the other person

43:02

to be pitching you instead. Yeah.

43:04

Interesting. It's a small thing

43:06

and it's easy to lose track of.

43:09

But I mean literally interviews is like

43:11

the one place where particularly technical people

43:13

get caught up. Yes,

43:15

I'm exactly that. I've written Python for

43:17

17 years exactly like you asked. And

43:19

I've known Kubernetes since, you know, Wozniak

43:21

was born. Please

43:23

hire me. Well, you

43:25

mentioned Kubernetes. I know you've dabbled

43:27

in the cloud or you're

43:29

kind of hanging out in like

43:31

the infrastructure section of technology. Is

43:34

that really where you besides

43:36

the alcohol section? What are

43:38

the aisles that you hang out in? you in the infra aisle

43:40

in the cloud? Where do you hang? So,

43:44

um, yeah, I'm

43:46

currently working with this company, Sevo, which

43:48

is interesting because Sevo is a cloud

43:50

company. I mean, think an AWS or

43:52

a Google, uh, what their, their Kubernetes

43:54

first, what makes them different is a

43:56

couple of things. One, um, you

43:58

know, it's UK based, which right now happens

44:01

to be really good for business, uh, for hashtag,

44:03

I don't know reasons, but, um, the, uh,

44:05

The thing that's also interesting is they will ship

44:07

you the whole software stack. So it'd be

44:09

like if you used AWS for a few weeks

44:11

and you're like, Hey guys, this is great. Can

44:14

we have it? Right. And then AWS just shipped

44:16

you AWS. So Civa will do that. You can

44:18

buy the software stack. They'll also

44:20

ship you a hardware appliance where you can throw

44:22

it in a rack. So what's interesting to

44:24

me about that is the 10 years that I

44:26

was in China, I basically missed the data

44:28

center. world. Because in 2004,

44:30

at that startup, I mentioned that built Twitter a

44:32

little too early. We had some server rooms

44:34

in the back that we would shut down a

44:37

machine at a time and install the newest

44:39

version and then shut down the next one and

44:41

install the newer version. Or

44:43

we put up a website that said down

44:45

for scheduled maintenance. You remember those? And

44:48

then when I came back,

44:50

it was cloud and blue -green

44:52

deployments. I

44:54

missed the whole data center world. So it's

44:56

SIVO is kind of interesting to me because

44:58

it's my first time touching anything data center

45:00

related. And, you know, not that the customers

45:03

necessarily are all doing that. Most of them

45:05

are probably just using our cloud, but I

45:07

find the data center bit really interesting, especially

45:09

as people are starting to flee the cloud.

45:11

But so that's that's one bit. It does

45:13

tend to be infrastructure. About a

45:15

year a half ago, I almost two years

45:17

ago, I went and raised money to start

45:19

an AI startup. So I did plan to

45:21

do an AI startup. My

45:23

co -founder after the VC said, we're ready to write

45:26

the check. And my co -founder called me and said, so

45:28

I think my wife's going to divorce me if I

45:30

do this. So we shut the whole thing down. Really?

45:33

Still good friends. But that

45:35

was in infrastructure because I

45:38

do think there's some interesting

45:40

infrastructure problems around AI in

45:42

particular that are still unsolved. The

45:45

reason infrastructure is interesting

45:47

is because a single

45:49

SaaS product with a

45:52

single SAS vertical, you

45:54

either hit the market perfectly or you don't.

45:57

With anything infrastructure, you get to ride the waves

45:59

because what you're selling is the pickaxes, right? And

46:02

it's not even pickaxes, it's blue jeans.

46:04

So it doesn't really matter if it's

46:06

gold rush or we're cutting down a

46:08

bunch of trees and clearing the forest,

46:10

everybody needs blue jeans. And so that's

46:12

the fun thing about infrastructure is you're

46:14

solving the problems that everybody needs to

46:16

go solve their problem. It's an easier

46:19

bet. So I find

46:21

that more interesting. And there

46:23

really aren't workarounds. I mean, you're going to need

46:25

it or you're not going to need anything.

46:27

Like infrastructure is going to be there, whether it's

46:29

rented, bought, cloud, on -prem, like

46:31

it has to be there in some form. And

46:33

so like you said, it is kind of

46:35

more like the blue jeans where there's not going

46:37

to be really a, there might be a

46:39

down market, but never a no market. Well,

46:42

and there's, okay. So then the thing

46:44

that's relevant here that I think is

46:46

interesting is part of what was

46:48

fun about the Kubernetes boom 10 years

46:50

ago when I was first involved in it

46:52

is it's like there's this new kind

46:54

of workload a container. And it's not really

46:56

that different, right? But turns out the

46:58

way that we orchestrated container is slightly different.

47:01

And the tools that ended up winning

47:03

in that orchestration were slightly different. And there

47:05

was a whole new paradigm in declarative,

47:07

you know, infrastructure rather than having to orchestrate

47:09

everything like we did in the days

47:11

of puppet or Ansible, right? It was kind

47:13

of exciting because it was a new

47:15

kind of workload. And the all the paradigms

47:17

we're going to change. And KubeCon in

47:20

like 2015, 2016, 2017, whatever I,

47:22

2015 might have been too early. Maybe I

47:24

started in 2016, but early on, it was a

47:26

whole bunch of nerds sitting around in a

47:28

room feeling like we're changing the internet and this

47:30

is fun. Right. Um, and

47:33

I feel like that's the only place

47:35

that that's also happening is in AI and.

47:37

AI is, you know, inference servers, the

47:39

way that we train them, the way that

47:41

we run them, the way that we

47:43

audit them, the way security looks in AI.

47:45

Like it's a new kind of workload

47:47

that requires a different kind of interaction. It's

47:49

not that different. It's really, you

47:52

know, it's just, it's just a workload at

47:54

the end of the day and you box

47:56

it up and some kind of wrapper at

47:58

the end of the day. But it's, there's

48:00

interesting problems because the things that go in

48:02

it don't come out in the same way.

48:04

And the things that come out. don't

48:06

come from the same places. And

48:08

so auditing that, life -cycling that, et cetera,

48:10

it's new, it's interesting. And the people

48:12

who are working hard to solve those problems,

48:14

I think, are having a lot of

48:17

fun. It's definitely

48:19

different enough that I

48:21

think there's ML Ops community.

48:24

And there's these like... which is basically was

48:26

DevOps or whatever you want to call it

48:28

prior, like whatever operations look like. Well, which

48:30

was platform engineering prior, which was, you know,

48:32

the Linux sys admin before, which was titles

48:34

in the cloud. You know, if you go

48:37

back far enough, it was a lot of

48:39

guys writing a lot of arc, just saying.

48:41

Oh, good point. Yeah. Anyways, keep going.

48:43

And they still might be the ones, you

48:45

know, the old graybeards still using their arc where

48:47

it makes sense. Well, I

48:49

was just going to say there's.

48:51

There are differences. Can you mention

48:53

security? Can you enumerate perhaps what

48:55

makes it different? Why would

48:58

there be a subculture that comes out

49:00

of this? It's just compute and data,

49:02

right? Like what makes it different enough?

49:05

Yeah, it should just be compute

49:07

and data. Okay,

49:10

so a concrete example, a friend of

49:12

mine, just this morning in the CTO

49:14

lunches community, a friend of mine works

49:16

at a very large, very regulated industry

49:19

and he's a VP at

49:21

this big company and he's arguing

49:23

that their company can't use

49:25

any AI products and it's super

49:27

frustrating. Just today,

49:30

Shopify tweeted something like, it

49:33

should be a reflex to use

49:35

AI. And that is the absolute standard

49:37

for all of our engineers at this point. And

49:39

so this friend who's at VP is like, well,

49:41

how do I get to that when I'm in a

49:43

company where I can't use any of it? And

49:46

so there are some solutions that have come

49:48

out that address it from his perspective. And

49:50

I'm going to start there, and I'm going

49:52

to come back to the infrastructure bit. But

49:54

from his perspective, the things are, who's

49:57

putting what where? right? And

49:59

when you have a company like his

50:01

that's big and super well regulated and

50:03

everything's blocked, you know what people do?

50:05

They pull out their phone and they

50:07

type something into chat. So now they're

50:09

putting data that they shouldn't be putting

50:12

into the big public models that are

50:14

probably, they've got the free version. So

50:16

the big public models are training off.

50:18

So that's the worst possible outcome, right?

50:21

So There's a company, an old co -worker

50:23

of mine is one of the early engineers

50:25

at called SurePath AI. I don't know

50:27

if you've heard of this, but SurePath AI

50:29

sits in between almost like a proxy

50:32

for a big company like this so that

50:34

every single thing that goes to say

50:36

ChatGPT, they can stop and say, oh, that

50:38

has PII in it. And they redact

50:40

it so ChatGPT doesn't even get it. So

50:42

when you say, hey, ChatGPT, who is

50:44

Jared from the change log? It's going to.

50:46

say, hey, you said who is named

50:49

from the changelog. Did you mean to give

50:51

me a name? It seems more like

50:53

a placeholder, right? Like that's going to be

50:55

the actual response from ChatGPT because they're

50:57

going to intercept that. Yeah. And

50:59

so there's those kinds of cases.

51:02

Now there gets to be more

51:04

interesting edge cases there that like, well,

51:06

I want HR to be able to use

51:08

AI and train on enough of our data

51:10

internally that they can ask the AI like,

51:13

Give me a list of all the salespeople

51:15

and it can do it because that's going

51:17

to increase their productivity. But I

51:19

don't need finance to have access to that.

51:21

And I definitely don't want HR to have

51:23

access to what finance has access to, right?

51:25

So there ends up being fine grained security

51:27

there, even if you're self hosting those models.

51:29

So that's one thing. The

51:31

next thing is when you're talking about

51:33

even just self hosting a model without

51:35

a product like SurePath, and this is

51:37

not an endorsement for SurePath. The guy

51:39

that works there is a good friend

51:42

of mine, I think very highly of

51:44

him, and I hope they're making a kick

51:46

-ass product. I've never used it, so

51:48

I don't know, and I can't vouch for

51:50

it. But the folks that are building

51:52

something internal have to look at it,

51:54

and how do you audit what the

51:56

model's been trained on? How do

51:58

you audit what the model's learned? How do you

52:00

audit who's had access to that model? How

52:02

do you audit if a malicious actor got in

52:04

and started asking questions to that model? That's

52:07

not that different from just

52:09

looking at audit logs of a

52:11

traditional server of any kind,

52:13

right? But there ends up being

52:15

some unique challenges there. And

52:17

especially when you lifecycle those servers.

52:21

those inference servers and you have a new

52:23

model on it or you put in some

52:25

new training data. Who had access to the

52:27

new training data? When did that training data

52:29

go in? Like, now what did they have

52:31

access to? The problems actually get pretty interesting. And

52:34

the rabbit hole goes deeper and deeper

52:36

and deeper and deeper and deeper. I'd

52:39

say it seems like they snowball and

52:41

kind of fractal off even to now you're

52:43

multiplying problems against each other. Well, and when

52:45

you're using AI to detect what you shouldn't

52:47

be doing or you should, you know, and

52:49

that's the other thing that AI is really

52:52

good at is AI can watch you type

52:54

a name in and go, oh, that's a

52:56

name. Right. Oh,

52:58

you shouldn't, you shouldn't be saying that. So where does

53:00

that model live? Who's hosting that

53:02

one? Who's auditing that one? Yeah.

53:04

I mean, it's. This

53:07

is a whole new world and the

53:09

people who are missing out on the AI

53:11

gold rush because there's a lot of

53:13

us are our Luddites over in the corners

53:15

going this this is one step farther

53:17

than I want to go open stack was

53:19

enough I mean we tempted fate when

53:21

we You know trained the computers and I

53:23

need a command prompt and awk. That's

53:25

all I need See you get me Jared

53:27

now I need to be better at

53:30

reg X if awk is gonna be all

53:32

I'm gonna do because there's a Anyways,

53:35

so say we all. So

53:38

say we all. But this is where

53:40

all the fun is and there is

53:42

a whole bunch of things changing. It's

53:44

also the only companies with money right

53:46

now. Right. Yeah,

53:48

I was expecting you would say that eventually is like, well,

53:50

you know, this is where the money is. So

53:52

why not be involved in the exciting thing

53:54

where the money is? Just kind of makes

53:56

sense. Well, Sevo has a

53:58

AI offering that is actually pretty

54:00

compelling. It's a private AI offering so

54:02

that you can be comfortable with

54:05

your data. And then this

54:07

other company in San Francisco that I'm

54:09

working with, Tensor Lake, that does understanding

54:11

of unstructured documents. It's actually really fascinating

54:13

for me. And I'm learning a lot.

54:15

I'm excited to be in the AI

54:17

world because that is where things are

54:19

moving. And I'm surprised how much I'm

54:21

learning about it. I

54:24

end up having conversations with chat gpt

54:26

every day, you know yesterday I was asking

54:28

a million questions about mcp servers and

54:30

like or mcp It's protocol or whatever, you

54:32

know model context protocol. Yeah, thank you

54:35

Like how does this work? Why does it

54:37

work? What are the limitations? How does

54:39

it interact with a browser? How does it

54:41

interact with AWS? Is it gonna write

54:43

this kind of code for me or do

54:45

this kind of interaction for me and

54:47

and it's You know just things the rabbit

54:49

hole always goes deeper it does It

54:51

does. And I was just going to ask

54:53

how much you're using these tools in

54:56

your own work because you have so many

54:58

disparate tasks, it sounds like, lined up

55:00

for any given day that do you tell

55:02

the AI, please spell Jared's name wrong

55:04

for me? As I send this email out,

55:06

or how do you, how

55:08

much are you using these tools

55:10

in your work? What's interesting is, so

55:14

for internal things, I

55:16

actually feel like ChatGPT has

55:18

just recently, and I mostly use

55:20

ChatGPT. I dabbled with Claude

55:22

and a couple of the other

55:24

things. You know, at

55:26

Sevo, I use Sevo's models, which

55:29

are llamas that

55:31

we're hosting. But

55:33

ChatGPT just in the last

55:35

few months feels like it's crossed

55:37

over from, hey,

55:39

I need an outline on XYZ

55:41

problem. What should I think

55:44

about that I haven't maybe thought about? And it would

55:46

give me, you know, one or two decent ideas

55:48

and the rest of it would kind of be fluff.

55:50

And now it's so good at it that I'm

55:52

regularly thinking I'd be stupid not to use this, right?

55:54

And so a lot of times I'll type up

55:56

a few things. Hey, here's the 10 things I'm thinking

55:58

about as it relates to this problem. What am

56:00

I missing? And it will structure a

56:03

few of the things that I put together

56:05

into the same category and then give me

56:07

new categories. I have it set up to.

56:09

be extremely obscenely brief, which is

56:12

funny because I'm being obscenely

56:14

verbose on this podcast. But

56:16

in life, I value brevity a lot.

56:19

And I have it set up to always

56:21

prompt me with, here's what you didn't ask

56:23

that you probably should have asked. And

56:25

so at the end of every

56:28

single response, it says, you should

56:30

ask this. And It

56:32

surprises me how often it catches things that

56:34

I didn't think. So to give you a

56:36

concrete example, and the short answer is I'm

56:38

using it all the time. It's

56:40

just completely gone through the roof. But

56:42

just in the last month, I kept trying

56:44

to use it and I couldn't find

56:46

times and spaces where it made sense. it's

56:50

become a second nature for me.

56:52

But I had done a write

56:54

-up on a bunch of things

56:56

that needed to happen in this

56:58

one organization, at this one company. And

57:01

I fed it in and said, you know,

57:03

make this better. And it said, you know,

57:05

at the bottom, there was this one section

57:07

where I said, here's some of the problems

57:09

that I'm seeing. And ChatGPD said, If

57:11

you wanna sound like a senior executive,

57:14

don't express the problems. Talk about how you're

57:16

working to address the problem. Don't just

57:18

bring them up. And I was like, ah,

57:20

this is like embarrassing. Like I should,

57:22

should 100 % know this? Should 100 % be

57:24

interacting this way? And it slapped me

57:26

on my hand before I went and took

57:28

that to the person I, you know,

57:30

was turning it into. And that was, that

57:32

was for me, I think the like

57:34

eye opening, like I'm stupid every time I

57:36

don't do this. because

57:38

it's so It's like an outbound filter, more than

57:40

anything else. It's just like, before

57:42

outbound. It comes up with, yeah, I mean, it

57:44

comes up with categories I don't think about. It

57:47

comes up with, you know, it gives me feedback

57:49

on my writing that I didn't think about. It

57:51

tells me to be more brief because I'm overly

57:53

verbose in my writing. And

57:55

I mean, even today I asked it to put

57:57

together a couple lists for me. Hey, I need

57:59

a list of this kind of company doing this

58:01

kind of thing. And I want 50 of them.

58:04

It takes 20 seconds. It's wild. It

58:07

is. It really

58:09

is. Now, does it name your liquors for

58:11

you or you do those yourself? So

58:14

thus far, we have a very

58:16

long spreadsheet of potential names. I

58:18

will say the generative A rise

58:20

amusing because in, I popped into

58:22

the, so the CTO lunches, we

58:24

have the free group that's 1600

58:26

CTOs that all meet for lunches

58:28

all over the world, right? And

58:32

then we have a paid group that's a small community

58:34

that hangs out at a slack. And

58:36

right before we launched the whiskey, I

58:39

went to the group and said, hey,

58:41

I have a couple of ideas for the

58:43

whiskey name. And here's what I think

58:45

it's going to be. And one of the

58:47

people in CTO Lunches said, well, what

58:49

about Generative A Rye? And that ended up

58:51

being what we use. So it was

58:54

not. He's asked for

58:56

a commission. I think I gave him a bottle. I

58:59

promised him a bottle. So if I haven't given

59:01

him a bottle, you know who you are, Topher. Give

59:03

him that $1 ,400, that special

59:05

edition. That special edition. Hey,

59:07

I hand wrote the number on all

59:09

of them. I'm holding bottle 22 right

59:11

here. That's cool. That's cool. Gosh,

59:14

you're really selling them. all small batch. I'm

59:16

wanting one of those. Dang it. Now

59:18

you now you price anchor me. I'm thinking 125,

59:20

you know, affordable. Feels like no big deal.

59:22

Yeah. Knock on the door. Would

59:24

you like to buy a $50 ,000 vacuum?

59:26

No. Well, great, because I've got a

59:28

hundred dollar vacuum. So

59:31

can you I know you haven't.

59:33

These are probably a private spreadsheet, but

59:35

can you humor us? Can you

59:37

can you workshop a few options out

59:39

there? Maybe Oh my gosh, let

59:41

me let me see if I can

59:43

I think that I might not

59:46

be willing to share Okay, take a

59:48

moment and review Maybe give me

59:50

your bottom half, you know, don't give

59:52

me the real good ones. Give

59:54

me what you even spit out Pull

59:56

it up real quick. Let's see

59:58

That's what I found specifically with creative

1:00:00

work is that chat GBT is

1:00:02

so mid on Everything

1:00:04

funny or ironic or tongue -in -cheek like

1:00:06

it's like no dude like the

1:00:08

the cheese ball factor is at 10

1:00:10

and They can't quite do anything

1:00:13

funny, but you know boring prose is

1:00:15

fine Okay, so I open it

1:00:17

up. There is a number in here

1:00:19

that I'm happy to share So

1:00:23

we thought about in

1:00:25

gin earing for a gin.

1:00:28

Again, I missed it. I missed it

1:00:30

entirely. In gin earing, but the

1:00:32

gin is all capitalized. Gotcha. Gotcha. One

1:00:35

thing that I definitely hope that we do

1:00:37

is I want to have a

1:00:39

vodka that we sell and it's two

1:00:41

different labels. It's the exact same fucking vodka

1:00:43

and you can only buy them in

1:00:46

a box set and one is called tabs

1:00:48

and one is called spaces because it's

1:00:50

vodka and they're both empty and I just,

1:00:52

I love this idea. I like that

1:00:54

one. I think we will get there eventually.

1:00:57

We have emergency response

1:00:59

or incident response

1:01:02

emergency kit. Okay.

1:01:04

As a name, we

1:01:06

have I did sock two and all I

1:01:08

got was the stupid bottle of whiskey. There's

1:01:11

there's a one of my favorites is

1:01:13

actually I work in computers because I think

1:01:15

for the people who aren't in tech

1:01:17

but want to buy a gift for their

1:01:19

person who does work in tech, I

1:01:21

mean, most of them don't know why force

1:01:23

push is funny. And they're like, that's

1:01:25

almost too anything. Yeah, maybe it's

1:01:27

inappropriate. They don't know what to buy.

1:01:29

So the I work in computers as the

1:01:31

gift to give to the tech person.

1:01:33

I think it's really funny. I

1:01:36

do like that one. Yeah.

1:01:38

So we've got, we've got a

1:01:40

lot more. Um, yeah,

1:01:42

I'm saving the good ones. Yeah, I would

1:01:44

too. You know, why waste them on a

1:01:46

silly podcast? I mean, it started with Friday,

1:01:48

a Friday deployment was one of the names

1:01:50

of the liquors. And we later decided that

1:01:52

was going to be the name of the

1:01:54

company, um, Friday deployment spirits. Cause you want

1:01:57

to get in that Friday deployment spirit, either

1:01:59

you don't want to deploy on Friday or

1:02:01

you have to deploy on Friday and you're

1:02:03

going to need to drink. Uh, so.

1:02:05

I will be some gin. That's

1:02:08

exactly right. What else do you want

1:02:10

to talk about, Jared? Anything else that

1:02:12

I haven't covered? I just wonder how

1:02:14

you like gin, personally. I think gin's

1:02:16

disgusting and it tastes like a pine

1:02:18

tree. Oh, man, that's

1:02:20

exactly right. That's what you're missing. You like to

1:02:22

drink pine trees. Well, you are from Colorado, so

1:02:24

a lot of pine trees are there. There you

1:02:26

go. Well, when I was a kid, my dad

1:02:28

would drink gin occasionally, and I

1:02:31

would smell it, and I was like, that smells

1:02:33

disgusting. He'd be like, have a sip, and

1:02:35

I'd take a little sip, and I'd be like,

1:02:37

that's disgusting. That's disgusting. And I remember growing

1:02:39

up and being like, is that the one that

1:02:41

tastes like pine needles? Ew, no. Exactly. And

1:02:43

then I became an adult, and I had some

1:02:45

pine needles, and I was like, this is

1:02:47

amazing. still

1:02:50

the same taste. For what

1:02:52

it's worth, Jared, Arjen

1:02:54

has some juniper in it, and you can

1:02:57

tell, but it also has a lot more going

1:02:59

on. So if you're not a big juniper

1:03:01

guy and you want to try a G &T

1:03:03

that you might be really excited about, but

1:03:06

I like it. If you send me a bottle, I'll give it a

1:03:08

try. If I send you a bottle,

1:03:10

will you do an entire hour and a

1:03:12

half episode where you just sit there and

1:03:14

drink it and then give a review of

1:03:16

it that gets increasingly unhinged as you get

1:03:18

farther and farther through the bottle? Because

1:03:20

then I will send you two bottles. I can

1:03:22

get you halfway there. Well, I don't like to

1:03:24

taste the gin. So I can't promise I'm going

1:03:26

to drink it for 90 minutes, you know. But

1:03:28

I will try it on the show. And

1:03:31

I'll talk with and maybe send one to

1:03:33

Adam. We could have a drink and we could

1:03:35

maybe do a review. Probably not 90 minutes, but.

1:03:37

Uh, yeah, I could do something with it. I

1:03:40

believe you can stretch it to 45 minutes. I

1:03:42

will 100 % said you a bottle of liquor

1:03:44

if you do a review on the changelog. All

1:03:46

right. Thus closes my

1:03:48

sales pitch. So,

1:03:51

so wait, we came into this hoping to have a

1:03:53

bubbly conversation. You got a little bit of that and

1:03:55

you're going to end with a bottle that you want.

1:03:57

You want the whiskey to, I mean, let's go. I

1:03:59

actually do like whiskey, whereas Jen, I don't like. So,

1:04:01

you know, do you like, do you like rye whiskey?

1:04:04

Um, probably I'm not, I'm not

1:04:07

a connoisseur. But what's the difference?

1:04:09

What's Rye specific? Rye

1:04:11

is more peppery. It's

1:04:13

a kind of different. I mean,

1:04:15

I guarantee you've had Rye whiskey

1:04:17

if you're a whiskey drinker and

1:04:19

you might just not. Unless you

1:04:21

try a Rye next to a

1:04:24

wheat, you might not notice

1:04:26

the big differences. Fair. Yeah, I

1:04:28

don't have an advanced palate. I do

1:04:30

know that I prefer bourbon. And that's I

1:04:32

don't have a refined palate either and

1:04:34

I make liquor for a living. So it's

1:04:36

okay. Sorry. What were you saying? No,

1:04:38

I was just going to say I generally

1:04:40

like bourbon, which I think has more

1:04:42

of a sweetie, you know, more sweet than

1:04:44

peppery, but yeah, yeah, it's the corn

1:04:46

gets the corn gets sweet, but also a

1:04:48

little bitter. A wheat one is like

1:04:50

really smooth and sweet too. But

1:04:52

a wheat makes bourbon feel.

1:04:56

Downright offensive if you drink wheat if you drink

1:04:58

wheat whiskey for a week and you tried

1:05:00

bourbon again You'd be like whoa because it's got

1:05:02

so much stronger of a kick. Gotcha. So

1:05:05

it's way smoother. Yeah, and a rise

1:05:07

gonna be the middle I mean a riot depends

1:05:09

on what the other grains are at the the

1:05:11

the rye tends to be more like a I

1:05:13

say pepper in part because it's a it's a

1:05:15

little bit of a spice that's sort of added

1:05:17

on to whatever the rather flavors are that are

1:05:19

going on so You can have a very sweet

1:05:21

one with a rye cake. You can have a

1:05:23

bourbon with a rye cake. Well, that's I'm I'm

1:05:26

excited. I'm excited for these. I'm excited. You're

1:05:28

excited. I wish Adam was here because he's actually

1:05:30

more of the whiskey guy than I am.

1:05:32

He's got all the all the stuff he just

1:05:34

said. He would have been saying the same

1:05:37

thing. Just make him feel guilty. It's true. Guilt

1:05:39

covers a multitude of wrongs. That's right. All

1:05:41

right. Well, Kendall, awesome meeting you. Great speaking with

1:05:43

you. Anything. Thanks for having me. Anything you were

1:05:45

wishing, I'd ask you. I never did. And you're

1:05:47

like, gosh, this guy's a dope. I've been waiting

1:05:50

the whole time. I mean, I

1:05:52

think the obvious question that you didn't

1:05:54

ask is, how can a man be

1:05:56

so friendly and look so good and

1:05:58

be so humble? And I

1:06:00

just don't have an answer to that. Some

1:06:03

of the world's mysteries will just have

1:06:05

to be left as such. Thanks

1:06:08

for having me. This was fun. Yep,

1:06:11

thank you. Kendall

1:06:15

is a man of his

1:06:17

word, and my complimentary bottles

1:06:19

of generative ary and force

1:06:21

push gin are in route

1:06:23

as we speak. I

1:06:26

am also a man of my word,

1:06:28

so even though I don't like gin,

1:06:30

we'll give these spirits a taste test

1:06:32

with real -time reactions on an upcoming episode

1:06:34

of Change Dog and Friends. Who knows?

1:06:37

Maybe my old man taste buds like

1:06:39

the taste of pine trees, and

1:06:41

I just don't know it yet. Okay,

1:06:43

let's thank our sponsors one more

1:06:45

time. Thank you to fly .io, to

1:06:47

depo .dev, and to Notion, Notion .com slash

1:06:49

changelog. And thank you to Breakmaster

1:06:51

Cylinder for producing so many custom beats

1:06:53

for us that we've just released

1:06:55

our fourth album on streaming services all

1:06:57

around the world. Search for changelog

1:06:59

beats and give the brand new after

1:07:01

-party album a listen. Your flow state

1:07:04

will thank you. That's all for

1:07:06

now, but we'll talk to you again

1:07:08

on Friday. That show is a

1:07:10

little bit bonkers. Matt Ryder joins

1:07:12

us and he's taking us to Matt

1:07:14

World, which I hear is pretty cool, but

1:07:16

maybe not as cool as Jared World.

1:07:18

You be the judge on Friday. Bye

1:07:20

y 'all. You