0:04

Welcome to Fall Through, your source

0:07

for deep conversations about technology, software,

0:09

and computing. I'm your host, Chris

0:11

Brando, also known as scriptible, and

0:13

on today's episode, we are talking

0:15

with the one and only gopher

0:17

man himself, Bill Kennedy. We talk

0:19

about courses, teaching, what even is

0:21

a senior and staff engineer, and

0:23

so much more. This episode has

0:25

an hour of bonus content. So

0:27

if you want to hear more

0:29

in-depth conversations with Bill, head on

0:31

over to fall through dot f

0:33

n slash subscribe to sign up today.

0:35

And as always, if you're not already, please

0:37

give us a follow on social media. And

0:40

if you want to see our lovely faces

0:42

instead of just hearing our lovely voices, head

0:44

on over to YouTube and smash a subscribe

0:46

button and ring that notification bell so you

0:48

can find out exactly when we drop all

0:51

of our video episodes. And with that, let's

0:53

getting into the the show. Welcome

1:11

back to Fall Through. In this

1:13

episode, it's going to be myself,

1:15

Matthew Snaabria and Chris Brando. How

1:17

are you doing here, Chris? I'm

1:19

doing great. It's a really sunny

1:21

day out. So it's nice. I

1:23

like the sun. It was raining

1:25

yesterday. So yeah, it is nice.

1:27

And it reminded me that I

1:29

have to go tend to my

1:31

grass. It needs some raking and

1:33

loving for the spring season. And

1:35

we're joined with a very special

1:38

guest. Probably. No, I'm, you know,

1:40

I want some of that rain, Chris. We're

1:42

on a severe drought in Miami, and

1:44

everything around me is dying, and I

1:46

have to be out there with the hose

1:48

every other day to keep my

1:51

investment in flowers and plants and

1:53

trees. So I'm jealous that you said

1:55

it was raining. Yeah. I mean, it's nice

1:57

what it rains like for a little

1:59

bit. then we have my sunny day,

2:01

so I don't mind that. Certainly I don't

2:04

want another drought. You're entering the rainy

2:06

season in Florida though, right? Dude, I

2:08

think we've had less than a half an inch

2:10

of rain since October. Oh no. Yeah, that's not

2:12

good. I know, usually in the summer, like... No,

2:14

yeah, I know, but if it doesn't start

2:17

raining soon, we're gonna have a lot of restrictions.

2:19

Yeah, that's bad. When I was living down there

2:21

every day in the summer, like two o' in

2:23

the summer, like two o' o' o' o' o'

2:25

o' o' o' o' o' o'clock in the afternoon.

2:28

nice downpour. You'd get that like like clockwork.

2:30

And it's good. Goes away if your

2:32

plants are watered and you move on.

2:34

Yeah, it's just true. It's funny. I,

2:37

my wife obviously knows who you are,

2:39

right? Like I've, you know, I think

2:41

you probably met her, I don't even

2:44

remember, but she knows who you are

2:46

because of the hat, right? So she

2:48

says, oh, that's Mr. Gopher man. That's

2:51

how she refers to. She calls you

2:53

Mr. Mr. Mr. Gopher. Gopherman. stickers that

2:55

are modeled after you or the gophers

2:58

with the little hat so for anyone

3:00

that's listening we are interviewing the the

3:02

mister gopher man himself but find me

3:05

out of conference you can get some

3:07

of those stickers yeah yeah no in

3:09

this episode we're gonna chat all things

3:11

you know go art and labs training

3:13

education hiring senior engineers we're gonna chat

3:16

about a bunch of this stuff in

3:18

this episode so Why don't you start

3:20

us off, Bill, with a brief introduction

3:22

of who you are, what you're what

3:24

you've been up to? Name's Bill Kennedy.

3:27

I have a small company in Miami

3:29

named Arden Labs, where we primarily do

3:31

consulting, staff, all training, all that good

3:34

stuff. Check out the website. You can

3:36

also check out Arden Labs.a. I've been

3:38

spending the last six months building AI

3:41

demos. There's really some cool fun stuff

3:43

out there as well. We're going to

3:45

client projects. clients. I try to choose

3:48

one project every year and right now

3:50

I'm doing work with prediction guard, which

3:52

is Daniel Whiteknacks Company, and they're basically

3:55

a AI platform that gives you that

3:57

security and all that stuff you need

3:59

to actually run. AI apps that aren't

4:01

being consumed by open AI and

4:03

Google and God knows what they do

4:05

with your data, right? All that good

4:07

stuff. Yeah, I guess still going to

4:10

conferences, doing workshops, did a

4:12

meet up yesterday and yeah, you

4:14

know, try to keep my favorite thing

4:16

now for the last three months

4:18

is ultimate or ultimate software design

4:21

live. So every twice a week,

4:23

I'm online for two hours and

4:25

we're hanging. building out code with one

4:27

of my junior engineers and I was working with

4:29

him a lot. I'm trying to get him up

4:32

to speed and I started realizing all the conversations

4:34

that I was having with him should be recorded

4:36

somewhere. So then I said why don't we just

4:38

start building something online that everybody can see and

4:41

I can have these conversations not just for you

4:43

but everyone else. That's been a lot of fun.

4:45

Nice. I love how you like you just hit

4:47

all the topics right off the gate, AI hiring,

4:50

education, all need-ups. We're gonna get to all of

4:52

that stuff too, which is great, which is great,

4:54

which is great, which is great. Well, we'll get

4:56

there slowly. I remember, I think the

4:58

first time I met you was when

5:00

I think I took your ultimate go

5:03

course way back in Florida. I went

5:05

to the physical location and did that.

5:07

And it was life changing. It was

5:09

really, really fun for me. And I

5:11

remember taking the course when I was

5:13

newer to go and I didn't understand

5:15

everything that you were talking about at

5:17

the time because I was still new

5:19

to it. It really did change my perspective.

5:21

Is Ultimate Go still like your primary

5:23

thing? Or I guess you said you

5:26

switched to like, you know, now you

5:28

have the live stream. Is that kind

5:30

of the bread and butter now for you?

5:32

So what I did, you know, I taught

5:34

the original class for a good five years

5:36

straight, and I was about to be

5:39

put into an insane asylum if I

5:41

had to say those words one more

5:43

time. Like at some point, I don't

5:45

know how musicians sing the same song

5:47

every night for... Three like I did it

5:49

for three years. So imagine me having to

5:51

sing the same song every night for like

5:53

I couldn't do it Yeah, so I had

5:55

to come up with a new class which was

5:57

the ended up being the ultimate service class

6:00

But everybody thinks they know

6:02

how to build a service. Nobody wanted

6:04

to take that class. So eventually

6:06

I retrofitted the class around some

6:08

Kubernetes and software design, which is

6:10

really what I was teaching. And

6:12

so that class gained a lot of

6:15

popularity, and I love that class

6:17

because it's a more practical class.

6:19

If you remember my first class, it's

6:21

great, right? If performance matters, these

6:23

things, and this is how goes

6:25

giving you performance. And we don't

6:27

go into some. deep level semantics.

6:30

But what I always hated about

6:32

that class was that at the

6:34

end of it, you weren't necessarily

6:36

any more prepared to go off

6:38

and start building product. From the

6:40

ultimate go class, you mean? Yeah,

6:42

right? Like you had an amazing

6:44

wealth of knowledge now, but you

6:46

didn't really now, it wasn't set

6:48

to start building something. So the

6:51

ultimate software design class is the

6:53

complete opposite. At the end of that,

6:55

you're ready to start building, and you

6:57

have a bunch of code. You've just

6:59

added to your toolbox to do that.

7:01

So what I've been doing is sneaking

7:03

in the important parts of Ultimate Go

7:05

into Ultimate Software Design. So you

7:07

essentially get both classes sort of mixed

7:10

in one, and then you're able to,

7:12

it's a much more practical sort of

7:14

class, so you can start building things

7:17

sort of out of out of the

7:19

gate. And that's kind of what

7:21

the Ultimate Software Design Live is.

7:23

We started building a, I had a

7:25

list of projects, like you should

7:27

build, if you're learning a new

7:29

programming language, build this list of projects,

7:32

right? Like the first one was a

7:34

crud. Well, that's the ultimate software

7:36

design, right? We built Crud there. But

7:38

the next one was a chat system.

7:40

So I said, Kevin, we're gonna build

7:43

a chat system, right? But we're gonna

7:45

do things in a way that

7:47

anybody that's watching should be

7:49

able to just clone and run, and

7:51

they don't need anything. Actually, it

7:53

is pretty cool. We've built,

7:55

I mean, it is decentralized.

7:57

I'm going to cheat a little bit.

8:00

I used gnats for the basically the

8:02

back plane fabric. So we built chat

8:04

server connects to gnats. So chat servers

8:06

can talk to each other. And then

8:09

we built a client and go that

8:11

connects to your little chat server. And

8:13

we have a pretty distributed system at

8:15

this point, right? Because anybody can start

8:18

a gnats for gnats system and gnats

8:20

can be sort of plugged together. And

8:22

yeah. And we just did. We just

8:24

did what I'm called decentralized identification this

8:27

week, basically using the same Web 3

8:29

technology. Ethereum does with ECDSA keys, so

8:31

you get your own ID. And you

8:33

sign the data, and then the system

8:36

knows that this was signed by you

8:38

because you're the only one that could

8:40

sign it. And now nobody can pretend

8:42

they're you anymore. So we keep adding

8:45

this sort of stuff. So now the

8:47

next thing is going to be some

8:49

levels of encryption. And eventually. When we're

8:51

completely done, we'll have an AI. You

8:54

can turn your chat app into an

8:56

AI spot, and hopefully nobody even knows,

8:58

either talk, or they're talking to you

9:01

or the AI at that point, right?

9:03

That's awesome. Yeah. So we've got a

9:05

few more months of work. Nice. I

9:07

caught a few of the early streams,

9:10

and this was before I had to

9:12

go, my little conference tour that I

9:14

did, like last week and the week

9:16

and the week before. But I caught

9:19

a few of the streams stream, really

9:21

nice. It's really nice. to this as

9:23

well or? There's no, I didn't, you

9:25

don't need Kubernetes for this. So you

9:28

literally could go to the repo which

9:30

is Arden Live/USDL, Ultimate Software Design Live.

9:32

You can bring down that repo. You

9:34

literally could start chatting with somebody today

9:37

right now with that. Nice. If you

9:39

know their their ID, so you know,

9:41

it's just like anything. I need to

9:43

know your phone number to send you

9:46

a message. Here's my. right? My 20-byte

9:48

ID based on a private key that

9:50

we just gave you and you could

9:52

have a communication right now today talking

9:55

so much. See,

10:01

there is a group of people

10:03

on the internet whose daily

10:05

life is all about negativity and

10:07

I don't understand it. Like,

10:10

you took 10 minutes out of

10:12

your life that you're not going to

10:14

get back to say this to me. Like,

10:16

even the book reviews, I go on

10:18

Amazon sometimes and I read some of

10:20

the book reviews and I laugh. If

10:22

they're really bad, I'll publish them on

10:25

Twitter because it's just so. And the

10:27

only thought in my head is that's

10:29

fair if you feel that way, but

10:31

you could have done a

10:33

thousand other things in that

10:35

10 minutes than than what you

10:37

did. And I and that's just not my

10:39

person. Like, I don't understand it. Yeah,

10:41

it's it's tough, right? You're a very well

10:43

-known person in the go community. I'm sure

10:45

you've had your fair share of people

10:48

saying things to you on the internet. And

10:50

I often find that this this industry,

10:52

the tech industry specifically has a very weird

10:55

relationship with feedback. It's it's like

10:57

this industry is afraid to to give

10:59

feedback to others. And dare I

11:01

say, criticize others, right? There's like a

11:03

there's a weird paradigm here that

11:05

if I criticize your work bill, right?

11:07

And I say, hey, like, what

11:09

do you why you return that error

11:11

there? What's going on? Or why are you

11:13

panicking there? And some people take that as

11:16

a personal attack, whereas some other people say,

11:18

oh, yeah, Matt, you that. Why am I

11:20

panicking there? And they have they take it

11:22

as a learning opportunity and they can have

11:24

the discussion back and forth. And I often

11:26

find with a lot of these people that

11:28

publish the bad book reviews or some weird

11:30

slack comment or some, you know, post on

11:32

social media. When you ask them the direct

11:34

question back, like, hey, what? What have

11:37

I done to you? Or like, what's wrong? They

11:40

fold and they don't answer. And it makes

11:42

me think that maybe you didn't have anything

11:44

there to begin with. You just wanted to

11:46

complain for the sake of complaining. It's

11:49

interesting. Very interesting behavior out

11:51

there. When I started blogging

11:53

back in 2013, it

11:56

was horrific. One, I didn't know nearly

11:58

what I knew now. and

12:00

I was publishing stuff and it was

12:02

not all of it was as

12:04

accurate as I wish it had

12:06

been in the beginning because I

12:08

was learning to in some cases

12:11

I was throwing stuff out there

12:13

just so somebody would correct it

12:15

at some point right because especially

12:17

back then I couldn't get the

12:19

go team to talk to me right

12:21

and there was a high-end click of

12:24

people back then that it was a

12:26

hard sort of group to to crack.

12:28

And so if I asked for help,

12:30

I never got anything. But if I

12:32

put a bad thought or idea or

12:34

anything out there, it would get corrected,

12:37

right? Yeah. But at the same time,

12:39

not only is it getting corrected, but

12:41

you're getting beat up because I'm

12:43

supposed to be even early on,

12:45

I'm training people. So imagine this

12:47

elite group of engineers who have

12:50

more experience than I do, went

12:52

to better schools than I do.

12:54

Know this stuff already. I'm just

12:56

learning it. and they're beating me

12:59

up left. It was hard. I

13:01

wouldn't even, at a conference, I

13:03

would almost stay away from everybody

13:05

because it was a horrible sort

13:07

of, so my skin has gotten

13:09

really thick over the years because

13:11

of that. But the other thing is,

13:14

why don't you just quit and give

13:16

up? Because at the time I had

13:18

five kids, it wasn't an option to

13:20

not take the beating. Right? I could

13:22

have easily just not taken the beating

13:25

and not grown my company the way

13:27

I did, but it wasn't an option.

13:29

Every time I wanted to quit, I

13:31

would just go, you can't quit. It

13:33

doesn't matter, just take the beating.

13:35

And then eventually I got to

13:37

a point where the material got

13:39

really sound because I worked at

13:42

getting there. And then the complete

13:44

opposite sort of happened. Then I

13:46

was kind of ignored. It was

13:48

like no feedback at all from

13:50

anybody anymore. And that was kind of

13:52

like a weird point in time too,

13:55

right? That you sort of get this

13:57

transition from getting beat up severely to

13:59

such... Generally, either people thought, well, it's

14:01

Kennedy, so it must be right, which

14:04

wasn't necessarily. Or the people that were

14:06

beating me up were just finally tired

14:08

of beating me up. I don't know,

14:10

dude. Yeah, yeah. Everybody found somebody knew

14:12

to beat up. Maybe, yeah. But I

14:15

don't leave comments on the blog. I

14:17

don't, I don't do that. I have

14:19

comments on the videos, but again, they're,

14:21

they're all super, you know. Dude, I'll

14:23

be doing a live stream stream and

14:26

somebody will ask a question. I'm like,

14:28

dude. Read the room, buddy. Like, read

14:30

the room. Yeah. Like, and I'm in

14:32

the middle of coding and I see

14:34

it. And my brain's just like, really,

14:37

this is what we're doing. Like, you

14:39

came, again, you spent five minutes your

14:41

life on that message today on this,

14:43

on this stream. On the Lord's Day,

14:45

you did this. Yeah. I don't know.

14:48

I don't know. Look, I always say

14:50

there's gonna be a moment where I'm

14:52

laying on my deathbed, hopefully, and I'm

14:54

thinking about all the time I wasted,

14:56

right? And there's plenty of it. But

14:59

I won't be any of that stuff.

15:01

I won't be leaving a bad review

15:03

because I had nothing else to do

15:05

for that. Like, you know, I don't

15:07

do. But it comes with the territory.

15:10

If you're gonna put yourself out there.

15:12

This is true. If you're going to

15:14

put yourself out there, if you're going

15:16

to make content, if you're going to

15:18

do things, you do open yourself up

15:21

to criticism and good, bad, everything in

15:23

between, and it's true. I just wish

15:25

the industry itself embraced it a little

15:27

more and discussed it rather than like

15:29

either throwing the ban hammer around real

15:32

quick or saying, well, actually I'm leaving,

15:34

or I just wish people discussed it

15:36

more rather than just making a blanket

15:38

statement and walking out, you know. The

15:40

more you say, the more you speak,

15:43

you're going to make a mistake. When

15:45

I would do a 20-hour training, there

15:47

is no way on God's Green Earth.

15:49

I wasn't going to say two things

15:51

that week that I wish I hadn't

15:54

come out of my mouth. Right. The

15:56

probability of it happening is just high.

15:58

Right? You just hope to... survive it,

16:00

what it does. Right. People

16:08

have changed, like, I can't say in a

16:11

class anymore, we're going to take a

16:13

walk because people get offended, like,

16:16

there's a bunch of jokes. There are

16:18

a bunch of jokes and stories I

16:20

used to tell in training that over

16:22

the years I had to stop, somebody

16:25

would get offended in the class and then

16:27

I'd hear it. And I got to the

16:29

point where I almost had to just be

16:31

completely and I have jokes that

16:33

I know land and suddenly over

16:35

time they stopped landing that it

16:38

was wild to me too, right? And

16:40

so that's the other problem when you're

16:42

speaking as much as, let's say, I

16:45

do and you're trying to tell a

16:47

joke and it doesn't land that day

16:49

because there's somebody in that classroom that took

16:51

a fence to it for whatever reason. It

16:53

doesn't even matter why, but they did. And

16:55

now you're right. Now you got

16:58

that going on too, which is funny

17:00

because these same people will go and turn on

17:02

like kitchen nightmares with Gordon Ramsay and love the

17:04

fact that he takes the chefs into the back

17:06

to like yell at them or scold them. And

17:08

it's like you love it in that context. But

17:10

then when you're the one when you're the chef

17:12

getting told like, hey, I'm the come outside. I

17:14

want to talk to you for a second. Now

17:17

you're offended. Like what? How

17:19

does that make sense? It's very weird to

17:21

me. And I don't I don't I don't

17:23

get it either. It's I don't know. It's

17:26

interesting. There's a part of it too, where

17:28

I think it's those same people that I

17:30

mean, you know, Matt, you and I have

17:32

talked about this. That'll like go out there

17:34

and champion. Like when we talked about not

17:36

on the podcast, we talked about like the

17:38

whole thing a few years ago where everybody

17:40

had this crusade against certain words that we

17:42

were using like Blacklist and Master Slave and

17:44

all of this and a giant crusade about

17:47

it. And I remember so many conversations with

17:49

people and they're like, well, we have to

17:51

do this because these words, they're offensive. They're

17:53

harmful to people like you. And I'm like,

17:55

I've not talked to a single Black person

17:57

that has had a problem with any of

17:59

these. But I will tell you, here's

18:01

the long list of things that have

18:03

been problems. Can we work on these?

18:05

And people come back and say, and then

18:08

they want to change it to me

18:10

and it's so much easier. How dare

18:12

I do it? And I'm like, okay, well,

18:14

there's this balance that we got a, I

18:16

think we have, it's off balance at the

18:18

moment, right? Like I think we're, we're being

18:20

a little too, we're aiming for the easy

18:22

stuff to try and. fix, but when the

18:24

hard stuff comes, we don't want to do

18:26

it. And I think part of it too

18:28

is like, we have ascribed to this idea

18:31

that everybody should be comfortable all of the

18:33

time. And I feel like that plays into

18:35

this a little bit, where it's like, there's

18:37

a lot of places and a lot of

18:39

things where you're just not going to be

18:41

comfortable. And we have to accept that. Like

18:43

learning is an uncomfortable thing. That's the point

18:45

of it. You're moving yourself out of your

18:47

comfort zone to absorb some new knowledge. We

18:49

wound up in this really weird spot

18:51

where, like, okay, well, yes, you're in

18:53

this space to be uncomfortable, but you

18:55

can only be uncomfortable in certain ways.

18:57

And if someone makes you uncomfortable in

18:59

a different way, then we have to

19:01

remedy that, like, immediately, or that, like,

19:03

we have to find a way to

19:05

make it so that you get back

19:07

to feeling comfortable again. And it does

19:09

feel like it's gone, like, in a

19:11

very excessive direction at this point. And

19:14

I know we've spent a number of

19:16

episodes. communities in general, but I think

19:18

the go community too when it comes

19:20

to a lot of the stuff. And

19:22

I think we've seen over time kind

19:24

of what happens when when you overemphasize

19:26

uncomfortable too much, you wind up with

19:28

communities kind of being a

19:30

bit bleh, where you can't have conversations

19:33

or you can't focus on the right

19:35

parts of a conversation because somebody has

19:37

been very like, oh, like you said

19:39

that thing and I'm offended by the

19:41

thing and now we've got to talk

19:44

about that. So here's my two cents.

19:46

I'm a kid from the 70s, okay?

19:48

So, you know, and things, some

19:50

language had to change.

19:53

I'm not saying it didn't.

19:55

But for me, if I say

19:57

something that affends you.

20:00

And it's not readily in my

20:02

brain that I've done that, because

20:04

it's possible, because the way I

20:06

grew up and where I grew

20:08

up and all that stuff. If

20:10

you tell me straight up, Bill,

20:12

that upset me, I would say

20:14

thank you, I will never say

20:16

it again, because I didn't know.

20:18

I will never say it again,

20:20

because I didn't know. There has

20:22

to be a teaching moment. What

20:24

drives me crazy is when I

20:26

say, oh my God, I didn't

20:28

know that, I'm sorry, and that's

20:30

not taken at face value. It's

20:32

like, you should have known. Well,

20:34

slow down, slow down, right? Like,

20:36

you might think I should have

20:38

known, but I didn't, I'm telling

20:40

you, I'm apologizing, right, and move

20:42

on. And do, when I started

20:44

Gopridge, okay, I had to start

20:46

working with a lot of women.

20:48

And I had three women that

20:50

taught me so much in the

20:52

first year. All the mistakes I

20:54

was making with just seeing a

20:56

woman in a classroom that wasn't

20:58

smiling and ask her why she

21:00

wasn't smiling. And I would ask

21:02

one of the women I was

21:04

working with. You know, I said

21:06

this today and things really went

21:08

backwards and they'd be like, no,

21:10

you didn't. And I'm like, yeah,

21:12

no, Bill, you can't do that.

21:14

I had a year of Bill,

21:16

you can't do that. great enough

21:18

to sort of deal with that

21:20

and teach me and after a

21:22

year I stopped making all these

21:24

mistakes. I didn't know, right? And

21:26

so I just wish that I'm

21:28

okay with you saying what I

21:30

just said offended you. I'm okay

21:32

with that. I'm okay with you

21:35

saying what I just said offended

21:37

you. I'm okay with that. I

21:39

don't have to agree. So if

21:41

you said you offended me and

21:43

I said it again, like, come

21:45

on. Right, but I just want,

21:47

I just want the chance to

21:49

have the opportunity to learn and

21:51

let's say be forgiven at that

21:53

point. Agreed. I think you hit

21:55

it on the head when you

21:57

said earlier that there's just some

21:59

people out. that thrive in that

22:01

negativity, they want that negativity.

22:03

And there's a difference between

22:05

saying something in passing one time

22:07

and then apologizing for it if

22:09

you didn't know, versus consistently having

22:12

that behavior. Like you said, that

22:14

does make you think soul, right?

22:16

If you have that behavior consistently.

22:18

And it's just, it's interesting that

22:20

this permeates in the learning space,

22:22

because like you said, Chris, when you're

22:24

in the learning space, you have to

22:26

there's a there is a sense of

22:28

discomfort there right you you're not in

22:30

your comfort zone the whole purpose of

22:32

you in there to learn is to

22:34

be uncomfortable to stretch your mind to

22:36

observe absorb that new knowledge as you

22:38

said and I I can readily think

22:40

of times where I was either disciplined

22:43

by my parents or in a learning

22:45

situation in karate where I was learning

22:47

new like moves or whatever or where

22:49

I was like getting a pull request

22:51

reviewed and I was uncomfortable

22:53

And I was like, wow, I'm stretching

22:55

out here. When I walked away from

22:57

that, I grew, right? In all of those aspects,

22:59

I grew, I learned the thing that

23:02

I was being disciplined on, I learned,

23:04

you know, how to better do this

23:06

in Karate. My code got better because

23:08

the reviewer actually, like, gave it to

23:11

me rather than coddled me. And I

23:13

walked away from those, all of those

23:15

things, stronger and better. And I

23:17

think that concept has fell off

23:19

a little bit. We've... over index

23:21

on the coddling aspect. And

23:24

like you said to yourself, Bill, you

23:26

stop making some of those jokes,

23:28

you stop bringing some of this

23:30

stuff to your course, to your

23:32

classroom. And we can argue whether

23:35

that's for good or for bad,

23:37

but in a part of that, you left

23:39

part of yourself on the

23:41

sideline while you're teaching your

23:43

class. And there has to be some

23:46

like. Negative thing that comes out of

23:48

that, right? Because you're not bringing your full genuine

23:50

self to the class anymore, because you have to

23:52

tiptoe a little bit. And does that affect the

23:54

learning? Maybe, right? Yes, no, I don't know. Depends

23:56

in the audience, I guess. Depends in the content.

23:59

It's just, it's sad. talking from a point

24:01

where you appreciated the little jokes and

24:03

yeah we're gonna take a walk and

24:05

you knew it was a joke where

24:07

somebody else I guess had some experience

24:09

where maybe that happened to them and

24:11

now there's trauma around it right right

24:13

and again I don't want to bring

24:15

up trauma so if you tell me

24:17

that's upsetting you I won't say it

24:19

again when I'm in my third class

24:21

and I'm starting to see the same

24:23

pattern like mhm-hmm okay I guess

24:25

for whatever's happened in society for the

24:27

last end number of months, this is

24:29

no longer anything I can say, right?

24:32

Like, I have to gauge it too.

24:34

I mean, I'll stop getting work if

24:36

everybody's like, you can't bring Kennedy in

24:38

a classroom. You're friends, everybody. You know,

24:40

there's, but you know, I tell my

24:43

kids this, I didn't make this up,

24:45

but you learn more from your mistakes

24:47

than your successes, right? And it's easy

24:49

to beat yourself up every time you

24:52

make a mistake. I make 100 mistakes

24:54

a day, I make one really bad

24:56

mistake a week. That's what it is.

24:58

I lie in bed every night thinking

25:01

about, all the mistakes you made today,

25:03

right? We're going to do better tomorrow.

25:05

And the next time I'm like, well I

25:07

didn't do better today. But you're learning

25:10

more from the mistakes than you

25:12

are the successes. You need successes

25:14

to keep your positivity go. Can't

25:16

be like, right? And I try

25:18

to tell the kids that. My daughter just

25:20

made a mistake last week, it costs

25:22

you're like five grand. And I said

25:24

to her, this isn't, and she's 29,

25:26

I said, this isn't gonna be the

25:29

most expensive mistake you ever make in

25:31

your life. I hope it is, but

25:33

I promise you it's not going to

25:35

be. So now you've learned, now we

25:37

move on, right? And it's hard, man.

25:39

It's hard not to beat yourself up,

25:42

but I had seen an interview with

25:44

an astronaut. I can't tell you which

25:46

one was. And they said when they're

25:48

up in space and they make a

25:50

mistake. They're told they have

25:53

30 seconds to beat themselves up

25:55

and then they have to move on because

25:57

they don't have time not to up in

25:59

space That's it. So you get

26:01

30 seconds, everybody looks at you,

26:03

start your 30 seconds, beat

26:06

yourself up, and then you move

26:08

on. And I loved that

26:10

because I don't, I need more

26:12

than 30 seconds still. But,

26:14

you know, I'm, I probably, depending

26:16

on the mistake, have got

26:18

it under an hour at least

26:20

where I can finally just let

26:22

it go. Right. I think that's

26:24

a skill that you need to develop. You

26:26

need to recognize you're beating yourself up. You need

26:28

to recognize there's nothing more you can do about

26:31

that. Forgive yourself and move on. The faster you

26:33

can do that, the better your life's going to

26:35

be. Yeah.

26:37

I think there's, at the end the day, there's

26:39

a lot of, it comes down to having a

26:41

lot of nuance when you approach all of this.

26:44

It's like there's no, like, right or wrong way to

26:46

handle a particular situation. You've just got to handle

26:48

it the best that you can and learn from

26:50

that. And if you see that happen again in

26:52

the future, adjust accordingly. Kind of like what you

26:54

just said, Bill. Where it's like, yeah, if you

26:56

keep seeing people that are feeling a little bit

26:58

uncomfortable in the class because you told that joke,

27:00

like you got to weigh it. Like, is this

27:02

joke integral to the course? Or can I just

27:04

throw it out and not have any problem? And

27:06

if you throw it out have any problem, well,

27:08

that's an easy decision. Like, okay, well then, that's

27:10

not. Sometimes I need to tell the jokes to

27:12

wake everybody up. I don't even know if they're

27:14

paying attention. So sometimes you might say something just

27:16

to see who's, whoa, like who's,

27:18

who's still involved after four

27:20

hours, right? Again, nothing horrible, but

27:22

enough to just go off

27:25

the beating path a little bit,

27:27

you know, I don't

27:29

know. It can help with attention too. But

27:31

yes, I mean, I don't know. I've been

27:33

navigating it for almost 10 years now. I've

27:37

seen shifts of people in the

27:39

classroom. And I don't want to say

27:41

it's generational because I don't think

27:43

it's generational. But at the same time,

27:45

I ain't getting any younger, right?

27:47

I got, I'm 55 and I've got

27:49

young 20 year olds in the

27:51

classroom. Some level, you got to stay

27:53

hip too. Can't be the old

27:55

man in there either, right? Yeah. So,

27:57

so that's the other thing too,

27:59

right? I'm all over the planet. You have

28:01

to know where you are too sometimes because

28:04

things land. I had to learn the hard

28:06

way. I had to learn the hard way

28:08

on some of that too, right? Right. Yeah,

28:10

things always land correctly, depending on your audience

28:12

and culture. You're going to make, at the

28:15

end of the day, the point is, if

28:17

you're out there publishing, speaking, doing something,

28:19

the more you do, the probability of

28:21

making mistake is high, own it, apologize,

28:23

move on, don't do it again. Right. I

28:25

want to change topic slightly to the

28:27

overall like learning landscape, right? We talked

28:30

about, you know, the feedback and the

28:32

criticism stuff, but I want to talk

28:34

about the overall programming learning landscape. What

28:36

have you, what are you seeing in

28:38

the industry, right? What, what programming skills

28:40

are in demand? What have, what skills

28:43

have fallen off in recent years? Anything

28:45

about your curriculum that you've kind of

28:47

pivoted and changed on? I want to

28:49

talk about that stuff. I'm not the...

28:51

If you want a highly customized sort

28:53

of training, I'm not your person because

28:55

I know what you need. I know

28:57

this class is what you need. And

28:59

every time I got to get into

29:02

a meeting with a company saying they

29:04

want this, this, and that. I'm saying

29:06

you don't really want that, you need

29:08

this. And then I do the class.

29:10

They're like, Bill, this is exactly what

29:12

we needed. I know what you guys

29:14

need. I'm dealing with a day in

29:17

and a day out. What I'm

29:19

seeing a little bit more of

29:21

recently, five plus years old that

29:23

have been architected with a level

29:25

of complexity that nobody wishes was

29:27

there, but it happened. And as

29:29

people leave these companies, as these

29:31

engineers leave and they got to

29:34

hire new engineers, what they're looking

29:36

for are resumes that show enough

29:38

well-rounded experience where they feel that

29:40

they could bring this person in

29:42

and they can handle the complexity

29:45

without having to bother them too

29:47

much. Like learn the complexity of

29:49

the system. They're not even

29:51

testing for programming skills anymore

29:54

so much as, okay, they've

29:56

worked with enough different databases

29:58

and enough messaging systems. at least

30:00

it says on the resume, and enough of this,

30:02

and enough of that, I feel comfortable that I

30:04

can bring them into this chaos, and they can

30:06

be productive within a certain amount of time. That's

30:08

sort of what I'm seeing. Now, let me

30:10

tell you what I'm seeing that's making me

30:13

cry, because I'm doing a lot of

30:15

interviews for the last month, okay? A lot

30:17

of interviews with people that have anywhere from

30:19

three to 15 years of experience of experience

30:22

on the resume. And the more experience

30:24

you have on the resume, the resume, the

30:26

skater I get-der I get-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a I have found

30:28

a few number of people

30:30

who can actually take direction

30:32

and write an algorithm at

30:34

this point. I'm not going to tell

30:36

you what the algorithm is because

30:39

then everybody's going to

30:41

study it and then they're going

30:43

to come to Arden. But things

30:46

like iterations and math, people

30:48

have a big problem with

30:50

doing this. And I think

30:52

what's happening is everybody's at

30:54

work copying and pasting code.

30:57

writing web APIs, finding code,

30:59

copying it, hey it works,

31:01

moving on, and people have

31:03

gotten very very good at

31:06

finding some code that works,

31:08

patching it somewhere and getting

31:10

things, but not really understanding

31:12

what the code is doing, how that

31:14

code, why the code even works,

31:17

and it's scaring the bottom of

31:19

me. It's scaring that. When you

31:21

say, when you say iteration and

31:23

math, do you mean iteration as

31:25

in like iterators or do you

31:28

just mean the concept of small

31:30

changes over time? Like here's construct

31:32

a slice with something in it.

31:34

It could be an unlimited number

31:36

of these things in it. Now

31:39

iterate over that and do something

31:41

every nth item, right? Gotcha. So

31:43

you do mean the iteration aspect

31:46

of iterate over this? Do something

31:48

and every time, every four or

31:50

five iterations, do this other thing.

31:53

They can't, nobody can code

31:55

it. Nobody can code it. Nobody

31:57

can code it. Nobody can code

31:59

it. I had a guy that couldn't

32:01

code this little algorithm, but I kept

32:04

looking at his resume. I helped him

32:06

through it. So after about 10 minutes

32:08

with helping him, he got it working.

32:10

And I don't need you to be

32:12

able to just do it completely. I

32:14

just want to see how you interact

32:16

with me and how you get through

32:19

it. So he got through it. With

32:21

his resume, my brain said he should

32:23

have... done better, but I gave them

32:25

the next challenge that I have, which

32:27

is a concurrency problem. I give them

32:29

a function that runs serially, right? Again,

32:31

another sort of iteration that does things,

32:34

and I say, I want you to

32:36

throw a go routine at every one

32:38

of these items in the collection. Don't

32:40

do it one at a time on

32:42

one go routine. Like a fan-out pattern,

32:44

right? Code that for me. And this

32:47

guy coded that easier than he did.

32:49

the other algorithm that I asked him

32:51

to code. And then my light bulbs

32:53

really went on, right? Because he had

32:55

some experience already with some concurrency patterns.

32:57

And again, I had to help him

32:59

a little bit. I didn't care about

33:02

that, right? But he did a 10

33:04

times better job with what I would

33:06

have considered more complex than the simple

33:08

little algorithm. And then the light bulb

33:10

went on. I think people have almost

33:12

forgotten how to program in a traditional

33:15

sense, at least the way I was

33:17

taught. And now it's again all patterns,

33:19

and I let people keep the AI

33:21

on. I go, if this is how

33:23

you work every day, then interview that

33:25

way. Right? I mean, I don't really

33:27

care. Sometimes the AI helps you. Sometimes

33:30

it doesn't. I've seen the AI helping

33:32

people and they're ignoring it. And I'm

33:34

like, the AI just gave you the

33:36

answer, man, and you just like completely

33:38

threw that away. Right? It's scary, man.

33:40

It's scary. And I find the kids,

33:43

I'm gonna say kids, the people that

33:45

are out of school a year or

33:47

two. Do better on my interviews than

33:49

the ones that have more experience. That's

33:51

mind blowing to me. A lot of

33:53

the recent college graduates often over index

33:55

on algorithm study. So it would make

33:58

sense that they would like probably perform

34:00

better in that aspect versus people that

34:02

have been in the industry. And I

34:04

know there's probably listeners listening to this

34:06

right now that are like, well, the

34:08

algorithm stuff, leacote, sucks, this and that.

34:11

And yeah, you're right. Like that's not

34:13

a great way to interview. But also

34:15

when you have to work through some

34:17

problems, you do need to have like

34:19

algorithmic thinking. You need to you need

34:21

to describe what it is you're going

34:23

to do with this code in plain

34:26

language in an algorithm before you actually

34:28

just go hack at the code. And

34:30

then a lot of people's responses to

34:32

that are like, oh, well, I prototype

34:34

better with my my programming language. It's

34:36

like, but then how are you going

34:39

to tell the next person? You know,

34:41

you're going to tell them to read

34:43

your, your, code that nobody can read

34:45

because you're too clever. If you can't

34:47

explain to me the algorithm in plain

34:49

language, how are you going to fix

34:51

it or troubleshoot it when things go

34:54

wrong? That scares me too. It reminds

34:56

me of one of the things that

34:58

Leslie Lamport has said a lot, where

35:00

it's like, he draws this distinction between

35:02

programming and coding. And he's like, people

35:04

today. tend to be able to code.

35:07

Like you take the language, you can

35:09

produce some software with it, but that's

35:11

different from programming, which is this broader

35:13

thing of like actually designing software, designing

35:15

algorithms, designing all the pieces that come

35:17

together, and then coding is a thing

35:19

you do after you figured out all

35:22

of that stuff. And I think a

35:24

lot of people push back against some

35:26

of his ideas like, you know, using

35:28

TLA Plus or. just writing things down

35:30

because that's more programmer and less coding

35:32

and people want to be in the

35:35

coding they want to be in their

35:37

programming language because that's where they feel

35:39

the most comfortable that's where they feel

35:41

like they can be the most productive

35:43

but it does lead to people not

35:45

being able to conceptualize what it is

35:47

that they're kind of building in the

35:50

macro sense which I also think is

35:52

is troubling for businesses right because businesses

35:54

for the most part don't care about

35:56

the code at the end of the

35:58

day. They care if you've programmed it.

36:00

They care if the problem they have,

36:03

you've created a solution for, which you

36:05

don't have to do by writing some

36:07

code. You have to do that by

36:09

sitting down and figuring out what are

36:11

the parameters of this, how do I

36:13

build a solution? And then you can

36:15

go, and as people say, maybe not

36:18

reinvent the wheel, and go and find

36:20

a solution and say, hey, the solution

36:22

fits my problem. I can just use

36:24

this or I can use half of

36:26

this and half of this other thing

36:28

and bring those two things together. But

36:31

I think people want to jump into

36:33

the code a little too quickly. I

36:35

think we're starting to see the repercussions

36:37

of that and the repercussions of people

36:39

just not. wanting to go lower level

36:41

on things and not understand how things

36:43

work. Like a lot of people don't

36:46

understand how garbage collectors work. They don't

36:48

understand how their memory model works. So

36:50

like my program is slow and it's

36:52

like your program's slow and they're like

36:54

it's the garbage collector. It's because I'm

36:56

writing a garbage language and it's like

36:59

no your program is slow because you

37:01

have a garbage problem because you... allocated

37:03

a bunch of memory, then dropped it

37:05

on the floor, and of course it's

37:07

going to be slow, no matter the

37:09

memory model that you're using. But that's

37:11

not something that people are studying as

37:14

much and learning as much, because they

37:16

don't think they need to, because, oh,

37:18

the runtime and the garbage factor will

37:20

do all of this for me. So

37:22

here's something interesting today's show. We started

37:24

using a 20-byte array, which is a

37:27

common address in Ethereum for your ID.

37:29

a 20 byte array fixed no matter

37:31

what, right? And you can produce a

37:33

string from that, which ends up being

37:35

about 40 characters long, okay? So somebody

37:37

was like, Bill, why are you using

37:39

the common address? Why don't you just

37:42

use a string? Won't it be smaller

37:44

and more efficient? And immediately my brain

37:46

went, well, we're dealing with a 20

37:48

byte array. That 20 byte array doesn't

37:50

need anything on the heap. It's 20

37:52

bytes, right. We put that in a

37:54

struck. There's never anything going to be

37:57

on the heap unless the struct itself

37:59

ends up on the heap. But if

38:01

I start working with a string, the

38:03

string value will stay in the struct,

38:05

but the backing array of these bites

38:07

have to be on the heap. So

38:10

when you start thinking about all that

38:12

stuff, right, switching from a string to

38:14

a 25, I mean a 20 bite

38:16

erase, starts, I didn't even think about

38:18

it at the time, right? Plus you

38:20

get the type safetyness of a type

38:22

that isn't string, right? And you start

38:25

sharing all this stuff online and I'm

38:27

starting to realize again, just like you

38:29

said, right? Nobody really understands the, readability

38:31

is about understanding the cost of that

38:33

code, right? And engineering is having an

38:35

understanding of the cost of your decisions.

38:38

But I also always teach that there's

38:40

two modes you're in. You're in programming

38:42

mode and engineering mode. And programming mode

38:44

is find me the 20 lines of

38:46

code for the happy path that makes

38:48

this work. And then engineering is. How

38:50

do I put this in the product

38:53

in a way that's fault tolerant and

38:55

I deal with all the things that

38:57

are about to happen to me, right?

38:59

And to be able to maintain mental

39:01

models of the code base. So when

39:03

things start breaking, I know where to

39:06

go to sort of fix it. These

39:08

are the two things. And I think

39:10

college teaches you how to do some

39:12

programming, which is great, which is why

39:14

I have Kevin on the show, because

39:16

he's taught how to program. He's never

39:18

been taught how to engineer. There's no

39:21

time for it in university in university.

39:23

Right. And that's a little bit more

39:25

also kind of what we see. I

39:27

want to ask a question on the

39:29

learning landscape real quick. You mentioned earlier

39:31

that you see companies coming to you

39:34

with legacy Go code bases, right? Five,

39:36

maybe, five years old, maybe. What really

39:38

makes some legacy? Is it just the

39:40

age? Is it the version of Go?

39:42

Like what, what is actually the finding

39:44

legacy here for you? Legacy means that

39:46

you have a code base where the

39:49

breadth of knowledge about that code base,

39:51

the mental models of this code base

39:53

have almost gone completely lost. Where you've

39:55

had also so much turnover that there

39:57

wasn't even somebody there to teach the

39:59

mental models of the code base. has

40:02

intimate knowledge over code based on ever

40:04

mental model ever. What does that mean?

40:06

That means if they see something in

40:08

the logs, they know where to go.

40:10

You always know where that code is

40:12

at some level. You know how the

40:14

code works, you know where it is.

40:17

And it's not line by line, but

40:19

it's enough of a mental model to

40:21

know where to start. See, the moment

40:23

that's lost, now you're in trouble. And

40:25

God forbid, you don't have enough tests.

40:27

Because at that point, you don't know

40:30

what you're breaking breaking. Right? And I

40:32

don't use tests as a criteria for

40:34

legacy. But the older a code base

40:36

gets and the newer the people are

40:38

on it and the lack of mental

40:40

models, now you get into a very

40:42

scary situation about how do you update

40:45

it? I walked into a client that

40:47

had a code base. It worked. It

40:49

was a mess. They hired me to

40:51

build a version two that we could

40:53

build mental models around, right? I got

40:55

hired to do that. And we did

40:58

it, but the problem was they had

41:00

no tests. The first month was, and

41:02

luckily this was just like a web,

41:04

it was an API, a web API.

41:06

For the first month, all we did

41:08

was write tests on all the end

41:10

points to make sure that we understood

41:13

everything going in. and what we expected

41:15

going out, including all the error messages.

41:17

Once I had the tests, we could

41:19

start rewriting the API. This wasn't simple

41:21

APIs either. There were complex data structures

41:23

going in and out. But now we've

41:26

got a new version of the service

41:28

where there's at least two of us

41:30

that have a full mental model of

41:32

the code base, how to add stuff,

41:34

how to debug stuff, how to do

41:36

that. If I leave and the other

41:38

guy leaves and somebody new has to

41:41

come in, it doesn't matter that I

41:43

feel like this has been architected really

41:45

well. They have no mental model of

41:47

the code base. They're going to start

41:49

hacking on it. They're going to put

41:51

code in the wrong place, even though

41:54

it's going to work, because every time

41:56

I have another person start doing stuff,

41:58

my code reviews are usually, you can't

42:00

put that there. looked good and I

42:02

know it's working but you don't want

42:04

to put that there we got to

42:06

put it over here right and so

42:09

now that becomes a real mess because

42:11

now you're back to the point where

42:13

you can't make a change at all

42:15

because right the whole thing just now

42:17

become and it's not even because the

42:19

initial architecture was bad it's because the

42:22

next person coming in didn't know how

42:24

to maintain that architecture nor adapt as

42:26

things adapted. So that to me is

42:28

sort of legacy. So if you hire

42:30

me right now, and I don't know

42:32

your code base, it's legacy to me,

42:34

until I gain a mental model of

42:37

it, and I feel very comfortable that

42:39

I can run tests, and I can

42:41

make changes, and I know if it

42:43

broke stuff. Hmm. I feel like that's

42:45

one of the best definitions of legacy

42:47

that I've come across. So I feel

42:50

like it is, it puts it in

42:52

the context of not... the code as

42:54

like an objective thing of legacy or

42:56

not, but the subjective view of who

42:58

has to work with it that determines

43:00

whether it's legacy, which I think puts

43:02

more of the onus on us to

43:05

make sure that we do the knowledge

43:07

transfer or even better, actually archiving and

43:09

putting that knowledge into a place where

43:11

people can access it later. Because right,

43:13

if you wrote that code base, but

43:15

you also wrote, you know, we call

43:17

it documentation, but if you also write

43:20

down all of the information about it,

43:22

then somebody who needs to approach that

43:24

can go first, study what you've written

43:26

about it and like, oh, these things

43:28

go here and this is how we

43:30

architected it. This is how it did

43:33

everything. And now they have that mental

43:35

model. So now something that might have

43:37

been legacy before is no longer legacy

43:39

because now someone understands it so they

43:41

can work on it. Right. So it

43:43

also takes it away from a time-based

43:45

thing. Because people tend to think about

43:48

legacy in terms of time, right? Oh,

43:50

it's like, and you kind of. flip

43:52

that definition almost and you say it's

43:54

not legacy because it's old, it's legacy

43:56

because I'm new to the code base,

43:58

right, or I'm not, I, and not

44:01

just new, I'm knowledgeable to the code

44:03

base. I like that definition of legacy

44:05

as well, and that's a good point.

44:07

So then it begs the next question

44:09

of like, what about it is, like

44:11

what are they calling you for then,

44:13

right? Like what do you find yourself

44:16

fixing in these code bases then, normally?

44:18

Most of the time, if I'm in

44:20

those situations, it's really about rewriting it

44:22

to a new architecture. that I know,

44:24

that people that Arden know, and making

44:26

sure that every service that we're now

44:29

building in that company uses the same

44:31

sort of architecture and project structure, same

44:33

design philosophies and guidelines. That way, somebody

44:35

working on this service could jump easily

44:37

on this service, even if they don't

44:39

know the full understanding of what it

44:41

does, they'll know where everything is. Because

44:44

we build the same. the package is

44:46

the same way and we organize things

44:48

the same way. So I'm doing that.

44:50

Now sometimes I get hired to just

44:52

do code reviews and I always ask

44:54

this question first and it always pauses

44:57

everybody. What is your objective of me

44:59

coming in and doing a code review?

45:01

Like what's your objective? Is your objective

45:03

that I'm just going to make you

45:05

feel good that I think you're doing

45:07

things right? Is your objective that you're

45:09

going to feel like at the end

45:12

of this because I'm telling you that

45:14

I don't like what you're doing? start

45:16

rewriting things like what is your objective

45:18

and the funny thing is most people

45:20

don't know what their objective is in

45:22

in say one of these external code

45:25

reviews that's going on and I have

45:27

a client now that they'll reach out

45:29

to me for two hours and I

45:31

end up helping them redesign a package

45:33

that they've been sort of designing right

45:35

right where they're maybe not working with

45:37

concurrency as well as they could or

45:40

and so a lot of times I'd

45:42

like to do it. you know, packaging

45:44

is a brilliant thing in Go, right?

45:46

It's our unit of code. And so

45:48

a lot of times if you're going

45:50

to start trying to clean up a

45:53

code base, there's two things you're going

45:55

to need to clean up. One is

45:57

your project structure and how packages are

45:59

laid out, right? And I teach that

46:01

stuff with layering and domains and things

46:03

like that. in different ways you can

46:05

be successful with architecture and domains and

46:08

layering. These things have to exist. I

46:10

call them firewalls to be honest with

46:12

you, right? Yeah, I remember you talk

46:14

about that. So packaging is a way

46:16

to create firewalls between the most unit,

46:18

right? Discreet parts of your program. You

46:21

want to define layers that define firewalls

46:23

within the scope of data flowing up

46:25

and down. And then you want to

46:27

define domains, which we use microservices a

46:29

lot for. Right? You think about when

46:31

we went microservice crazy. Really what you

46:33

were doing is defining domains of functionality

46:36

at a process level. I'm trying to

46:38

teach domains in proc. How can you

46:40

create those vertical firewalls within the scope

46:42

of your project so you don't need

46:44

the microservice to sort of do the

46:46

domain? And I don't care what your

46:49

architecture is. It's at least from a

46:51

go perspective. You need these three things.

46:53

Good package design for that level of

46:55

firewall, layering. I do application layers, business

46:57

layers, whatever. I don't care what they

46:59

are for you. And then the domains,

47:01

right? These sort of vertical, sort of,

47:04

and if you think in terms of

47:06

firewall, then, right, and if you could

47:08

portray that in your project, where somebody

47:10

can look at it and say, where

47:12

are the layers? Okay, these are the

47:14

domains. If somebody within five minutes can

47:17

have a light bulb. that turns on

47:19

on a project and see the layers

47:21

and see the domains, you're already 50%

47:23

there in terms of... where they should

47:25

go. You can define what the roles

47:27

and responsibilities are for a layer. The

47:29

domain is always a domain of functionality,

47:32

whatever that is, right? And then good

47:34

package design is critical. That's where you're

47:36

going to make all your mistakes in

47:38

the first couple of years, right? We

47:40

talk about all that. And so if

47:42

I come into a project and I

47:45

look at it and I look at

47:47

how much code is there, if there

47:49

isn't a large amount of code, let's

47:51

say 5,000 lines or less. start from

47:53

scratch. Let's move those packages over one

47:55

at a time, put them in the

47:57

new place, and let's... fix those package

48:00

APIs, and we have time to do

48:02

that. We can project of that size.

48:04

We should be able to move over

48:06

in a month or less, depending

48:08

on the tests that we need. If

48:10

it's much bigger, then what we're gonna

48:12

do is you have two choices. You

48:14

start with all the existing packages you

48:16

have, but you come up with the

48:18

new layers. So you come up with

48:20

the new project structure and take all

48:22

the existing packages you have and try

48:24

to organize them in the right place

48:26

to the extent you can, right? because

48:28

that might be really hard to break

48:30

out. But you could always move a

48:32

package somewhere, right? We all agree. A

48:35

package that's here could be moved here.

48:37

We just got to change some imports.

48:39

And so if you get it. If

48:41

you could reorganize the packages you have,

48:43

you start to also understand, you know,

48:45

are these packages really the best ones

48:47

we need it? Are they doing too

48:49

much? And then you can start reworking

48:51

a package at a time and at

48:53

least get to better package design and

48:55

layering, you know, structural layering design. After

48:57

that, you could choose whether or not

48:59

you want to, you know, break it into

49:01

domains. You find that to be a

49:04

hard sell though? Like, a lot of

49:06

people would consider it wasted, right. I

49:08

know, I mean, in your case, you're

49:10

hired to come in, right? I get

49:12

that. They're hiring me because they already

49:14

feel it. Right. In the case of

49:16

maybe the listener or people listening to

49:18

this or even ourselves, when you're trying

49:20

to propose some of this design from

49:22

the inside, do you people come to

49:25

you and say like that's a hard

49:27

sell? Like I'm creating pull requests that

49:29

need three reviews that are just renames

49:31

and moves and people are mad at

49:33

me? Like, do you find that sort

49:35

of pushback? to agree that this has

49:37

to be done. And it's either one

49:40

of two reasons, right? It's the velocity

49:42

of the team is now completely lost.

49:44

They're just not moving fast enough.

49:46

And they're all complaining about the

49:48

same things, right? We don't understand

49:50

the code. We don't put things

49:52

with breaking stuff. And you've got

49:54

to solve that problem. So Bill,

49:56

I can help you with that, right? We

49:58

can figure that out. Sometimes what happens,

50:00

and this is happening more and more,

50:03

some company that wrote these Go services

50:05

just got sold to Company A, they

50:07

bought it, right? And now some brand

50:10

new team of Java developers who want

50:12

to learn Go, and are really strong

50:14

developers, by the way, want to learn

50:17

Go, are told you now have to

50:19

maintain this. Right. So now you're coming

50:21

in. to help them start to understand

50:24

the code, the packages, maybe the decisions

50:26

that they, the old product made, they're

50:28

responsible for maintaining it. And I've never

50:31

met a group of non-go developers in

50:33

that situation where they came from Java,

50:35

C-sharp, that couldn't come in and do

50:38

that work. Like, they don't need syntax

50:40

training, the syntax is there, right? It's

50:42

all more about, but I don't understand

50:45

packaging, yeah, well, it took me five

50:47

years to understand it too, so. Stop

50:49

beating yourself up, right? And then taking

50:52

opportunities. Like on that particular say client,

50:54

all we're doing at that point first

50:56

is trying to get a mental model

50:59

of where everything is and fixing one

51:01

package at a time. After that, then

51:03

you can start talking about restructuring the

51:05

project from some better mental models. But

51:08

at that point, they might feel comfortable

51:10

enough. But they're like, you know what?

51:12

We have the mental models. This isn't

51:15

that bad. The fact that we've just

51:17

cleaned up the packaging and start up

51:19

and config and that's enough. I want

51:22

to move a little bit away from

51:24

the general landscape of learning and package

51:26

stuff. I want to talk more about

51:29

your journey at Art and Labs. I

51:31

want to talk about specifically, you have

51:33

ultimate go, ultimate service, ultimate software design.

51:36

But I saw that you also have

51:38

more, right? You have docory of Kubernetes,

51:40

and more recently now you have rust

51:43

courses. And I found it interesting because

51:45

you are the gopher man, right? Mr.

51:47

Gopher man, as my wife says. And

51:50

Arden Labs has pivoted, I don't want

51:52

to say pivot, maybe pivot is the

51:54

wrong word, maybe expanded, you've expanded to

51:57

rust. Can you talk to us about

51:59

that decision? Dude,

52:01

if you're gonna stay in business for

52:03

a long time, you can't put all

52:05

your eggs in one basket. You have

52:07

to diversify, right? So to diversify from

52:09

an Arden Labs perspective means you can't

52:11

just be about go. Every year I

52:13

would wake up and say, this is

52:15

gonna be the last year for go.

52:17

This is gonna be the last year

52:19

for go. Nobody else's. At some point

52:21

it was for Ruby. Do we all

52:23

agree? Early

52:25

on in Ruby, there were Arden Lab

52:27

companies. There were companies like Arden

52:30

Labs going into corporate environments, training people

52:32

in Ruby, doing Ruby consulting, right?

52:34

And after maybe 20, 25 years, that

52:36

all ended because nobody needed it.

52:38

And every year I'm always like, you

52:40

know what, it's gonna be the

52:43

last year for go because no companies

52:45

are gonna need training. And corporate

52:47

training's come down a lot. Don't get

52:49

me wrong, right? It has. From

52:51

pre -COVID, I was living on an

52:53

airplane. COVID changed a lot of stuff.

52:56

So who knows if they would

52:58

have died a little earlier. But

53:00

corporate training is, right? That demand is

53:02

way down. So how do we

53:04

pivot? We pivot by video, right? COVID

53:06

made us pivot on video. So

53:08

now we have more companies that are

53:11

buying video training than wanting to

53:13

put me in a classroom. Again, nobody's

53:15

going back to work just yet

53:17

too. So, right? So it's, if you're

53:19

running your own company, and I

53:22

say this is the job of the

53:24

CTO, but it's the job of

53:26

the CEO and the CTO from a

53:28

tax perspective, to be constantly looking

53:30

out and saying, where are things changing?

53:32

Where do we have to be

53:35

in six months in order to stay

53:37

where we're at today? Because the

53:39

world is changing all the time. So

53:41

two years ago, I started to

53:43

see that Rust was gonna be here

53:45

to stay. It's not being used

53:48

the way it is now, but I

53:50

saw very clearly the writing on

53:52

the wall that a lot of really

53:54

strong go devs were looking at

53:56

it and playing with it. And a

53:58

few important projects started coming. up where it

54:01

was all going to be rust. So I told

54:03

Miguel at the time, I said, we

54:05

need to find somebody in rust. Somebody's

54:07

written some books. It's just a marathon,

54:09

not a sprint. And so we've

54:12

got to do this for the next

54:14

two years to get to a point

54:16

where if rust does hit, I think

54:18

it will, we're right there in the

54:20

middle of it. And sure enough, I

54:22

mean, Microsoft just announced what? No more

54:24

C code in their operating system. It's

54:26

all rust, right. And so. That was

54:28

several years ago, where we brought

54:30

in Herbert and work with Herbert

54:32

on the training. And there's still

54:35

more go activity at Arden

54:37

than rust activity. But we're

54:39

seeing it now. We're starting

54:41

to get some rust consulting.

54:43

We're probably doing more rust

54:45

corporate training than go corporate

54:48

training. And that makes me really

54:50

happy, Matthew, because at some point...

54:52

this company needs to be able

54:54

to run at a high level

54:56

without me being on the road.

54:58

Right. And it's a major reason

55:00

why I chose not to learn

55:03

rust at any level. Because one,

55:05

Herbert's great at it. He's

55:07

doing great. He's a good trainer.

55:09

And it gives me the opportunity

55:11

at some point when there's more

55:13

rust going on than go to

55:16

be able to, you know. After 10

55:18

12 years relax a little bit

55:20

and go okay. I know we're

55:22

we're there. So Rust was a

55:24

big thing that I saw coming

55:26

and again if you're trying now

55:28

if you try now as a

55:30

consulting company to get into rust

55:32

You're two years late because you

55:34

need a couple years to build

55:36

a reputation and anything. We're at

55:38

all the rest conferences now, right?

55:40

We've got the books where we're

55:42

like but that's two years of

55:44

work. So the next thing I'm

55:46

looking at and I haven't

55:48

pulled the trigger on it,

55:50

is Zig, right? I have

55:53

to ask the question, amazing.

55:55

If you'd asked me a

55:57

year ago, I would have

55:59

said. start learning Zig. If you're

56:01

asking me now, I think Russ

56:03

did a really good job over

56:05

the last year of kind of

56:08

putting Zig in the in the

56:10

background and I think what's also

56:12

happened is thanks to generics and

56:14

some other things the Go team

56:16

has done. More and more enterprises

56:18

are using Go. So a year

56:20

ago I thought Zig was going

56:22

to be the next thing and

56:24

Rust and Zig we're going to...

56:26

I don't think so anymore. It

56:29

doesn't mean I'm not keeping my

56:31

eye on it, because the moment

56:33

I think that, here's the thing,

56:35

every 20 years a new language

56:37

comes out, every 20 years, right?

56:39

C-sharp came out in 2000, then

56:41

what, 2000, say, 2010, it's kind

56:43

of go, but it doesn't really

56:45

get mainstream until maybe what, 2020?

56:47

You see what I'm saying? And

56:50

so, so if you look at

56:52

that, in 2030. Let's say in

56:54

five years from now, I have

56:56

to imagine that there's going to

56:58

be another language that's starting to

57:00

pop up and be the next

57:02

thing. Now, maybe it's still rust,

57:04

because rust is, I think this

57:06

is the year for rust. I

57:08

think with the Microsoft announcement, and

57:11

Linux using it now in the

57:13

OS and these types of things.

57:15

I think this could be the

57:17

year for now. I also think

57:19

that people that went to Russ

57:21

to build web services and stuff

57:23

have come back to go. Okay.

57:25

Yeah. I've seen some of that.

57:27

But I think Russ is now

57:29

is Russ that next language. I

57:32

don't know. I think there's another

57:34

one on the horizon. The only

57:36

one I've seen is ZIG, but

57:38

it hasn't gained that traction it.

57:40

This is what you're looking for.

57:42

Okay, the moment, look at jobs,

57:44

look at jobs, right? There was

57:46

a company or two last year

57:48

looking for ZIG developers. That's where

57:50

my brain, whoa, whoa, whoa, whoa,

57:53

whoa, right? I haven't. I've seen

57:55

it again. And so look to

57:57

see if companies are looking to

57:59

hire anybody in that tech at

58:01

all, even if it's somewhat experimental.

58:03

That will start to give you

58:05

an idea of at least developers

58:07

in house starting to play with

58:09

stuff. Right. Are the people working

58:11

with you for the rust stuff?

58:14

Are they asking the same questions

58:16

about, or are they approaching you

58:18

for the same reasons that they

58:20

were for go? I have some

58:22

legacy code or I have some

58:24

design patterns that I'm not really

58:26

fond of, or is it just

58:28

I want to learn rust? Like

58:30

what's the kind of paradigm there?

58:32

The paid work, which is why

58:35

we've shifted some of our workshops

58:37

now, a lot of the paid

58:39

work is I want to convert

58:41

this C code base to rust.

58:43

That's what we're seeing right now.

58:45

Okay. Okay. And so if you

58:47

look at the workshops we're giving

58:49

like we gave this year already

58:51

at Rust Nation It's essentially workshops

58:53

about how to port C to

58:56

rust at this point And we

58:58

have experience doing that I Herbert's

59:00

got huge experience doing that right?

59:02

So that's sort of what we're

59:04

seeing I expect to see more

59:06

of that as the industry who's

59:08

writing C code wants the benefits

59:10

of what Russ gives you. We

59:12

know what that is, right? The

59:14

type safety and all the other

59:17

stuff that you're not necessarily getting

59:19

from C. So I expect to

59:21

see a lot of that this

59:23

year. We'll see more of that

59:25

next year. And then as these

59:27

code bases are ported, now you're

59:29

going to do the next level,

59:31

which, okay, let's clean this up.

59:33

We'll need more practical classes around

59:35

that. And then we'll see where

59:38

it goes. Here's

59:43

the problem. Here's the problem. Here's

59:46

the problem. They are a very

59:48

small team in Google. Tiniest team,

59:51

right? Who has to constantly justify

59:53

their existence related to cloud revenue.

59:55

Which is very hard when you're

59:58

working on a programming. And so

1:00:00

I have appreciated maybe more than

1:00:02

most their core focus over the

1:00:05

last several years on the enterprise.

1:00:07

My business, right, my ability

1:00:09

to write go code isn't necessarily

1:00:12

going to come from manufacturing

1:00:14

or wasm in the browser. It's

1:00:16

going to come from more and

1:00:19

more companies moving off of whatever

1:00:21

programming languages there are today. Maybe

1:00:23

the two major ones C sharp

1:00:25

in Java and sort of switching

1:00:27

to go, right? And so for

1:00:29

me, having that have been their

1:00:32

focus for the last several years

1:00:34

and being very successful in it.

1:00:36

If you have a small team

1:00:38

of people, what would you rather

1:00:40

than working on? Right? If you're

1:00:42

me and you want to write go

1:00:44

code and you want to make money

1:00:46

on go code, then I want more

1:00:49

people using it than not, right?

1:00:51

So nobody realizes how small

1:00:53

the go team is. Nobody realizes

1:00:55

how like... Small budget it is

1:00:57

in compares to all Google and

1:00:59

nobody knows the pressure that every

1:01:01

developer on the go team is

1:01:03

under in terms of proving their

1:01:05

existence at the end of the

1:01:07

year Which has to be measured

1:01:09

around cloud? It is really hard.

1:01:11

I've tried to do things on

1:01:13

Gobridge to help certain members that

1:01:16

we're helping with education and things

1:01:18

like that when I could but

1:01:20

it's it's really even like Robert

1:01:22

Grissomer and Ian who work on

1:01:24

that lower level language stuff. Sometimes

1:01:26

when I see a proposal come

1:01:28

out like this new one on

1:01:31

air handling, which I'm not a

1:01:33

fan of by any stretch of

1:01:35

the imagination, but I look at

1:01:37

it now as Robert has to

1:01:40

prove that he's doing something here

1:01:42

that's valuable that's going

1:01:44

to help move go deeper into

1:01:46

the enterprise. And so I could

1:01:48

sit here and get on that

1:01:51

issue and fight to the nail

1:01:53

against it. But if I do that,

1:01:55

I'm kind of, what are they cutting

1:01:57

my nose to spite my face? Because

1:01:59

in. reality, if they can come

1:02:01

out with a syntax that's going

1:02:03

to make a whole, another million

1:02:05

developers decide that I don't have

1:02:07

to write if there are not

1:02:09

equals no, which I love writing

1:02:11

and Copilot does it for me

1:02:13

anyway now. But if you're telling

1:02:15

me that if he figures that

1:02:18

out and another million developers in

1:02:20

another thousand companies are going to

1:02:22

switch to go, I'd be an

1:02:24

idiot. to stop it from happening

1:02:26

whether I like it or not.

1:02:28

There. That's actually a good point.

1:02:30

Yeah. I do think I wish

1:02:32

that we could leverage the community

1:02:34

more in these things too. I

1:02:36

think they do to some degree,

1:02:38

but I think that's why, you

1:02:40

know, I like that Tiny Go

1:02:42

is doing the Wasm stuff and

1:02:44

that's a bit separate since it

1:02:46

is the type of thing where

1:02:48

it's like, yeah, it'd be kind

1:02:50

of hard for the Go team

1:02:52

to incorporate that, especially all the

1:02:54

microcontroller stuff that people work so

1:02:56

hard on. going to be a

1:02:58

moneymaker I think for Google in

1:03:00

the long run unless there's a

1:03:02

big a big pivot in what

1:03:04

they care about. But I also

1:03:07

think that it's like maybe the

1:03:09

community needs to push on Google

1:03:11

a bit to be like you

1:03:13

know it's care about this more

1:03:15

yeah worth it to care more

1:03:17

about these future things because that's

1:03:19

also what I push it's not

1:03:21

Google it's every company that's using

1:03:23

this open source project in production

1:03:25

and not paying for it's not

1:03:27

Google. If you're telling me. that

1:03:29

company ABC is using this and

1:03:31

they're making money, then they should

1:03:33

be giving Tiny Go money to

1:03:35

survive for another year. Right. And

1:03:37

or, which is fine enough for

1:03:39

me too, take an employee and

1:03:41

dedicate their existence to the project.

1:03:43

To me, both works. Right. What's

1:03:45

brilliant about AK is that. I

1:03:47

can give them $25,000 and they

1:03:49

can live on that comfortably with

1:03:51

the way they've sort of set

1:03:53

up their lifestyle. They are an

1:03:55

amazing human being that just cares

1:03:58

about this tech and doesn't need

1:04:00

to be driving a Mercedes. And

1:04:02

so 25,000 a year, which is

1:04:04

mind-blowing to me, is they're happy.

1:04:06

Let me just put it that

1:04:08

way, right? I'm sure if we

1:04:10

could get them 50, it'd be

1:04:12

happier, but they're happy at 25.

1:04:14

And so as long as Go

1:04:16

Bridge can continue to fund Tiny

1:04:18

Go and... Google continued to fund

1:04:20

Go Bridge, right, which is the

1:04:22

fastest way to do that because

1:04:24

I don't need Google lawyers and

1:04:26

all the other stuff, right? I'm

1:04:28

good, but company ABC could be

1:04:30

giving them 10 grand and or

1:04:32

an engineer at 100,000 a year.

1:04:34

You see what I'm saying? Yeah,

1:04:36

and this is the same problem

1:04:38

with open source everywhere. Right, and

1:04:40

we can have a whole entire

1:04:42

show. I mean, Felipe, Felipe, has

1:04:44

tried to solve this problem and

1:04:46

he's done some amazing thing. Get

1:04:49

Felipe on the show, have him

1:04:51

talk about what he's done over

1:04:53

the last two years, because he's

1:04:55

one of the few non-googlers right

1:04:57

now, working on the cryptography packages.

1:04:59

He probably has more commits than

1:05:01

anyone on that project right now.

1:05:03

He doesn't work for Google. But

1:05:05

that's not because Google's paying them.

1:05:07

That's because he was able to

1:05:09

able to able to. And it

1:05:11

took him a while. He was

1:05:13

able to find, I don't even

1:05:15

know what it is, 8, 10

1:05:17

companies to write a check. And

1:05:19

if you go to his page,

1:05:21

you'll see who those companies are.

1:05:23

I wish I could, I always

1:05:25

like to shout them out right

1:05:27

now. But get him on the

1:05:29

show. Let him talk about what

1:05:31

he's done, because he's trying to

1:05:33

solve this problem, not just for

1:05:35

himself, but for others. And it's

1:05:38

a big problem. But if you

1:05:40

could sleep at night, knowing. that

1:05:42

you're using something for free that

1:05:44

you're now making money on that's

1:05:46

on you do that's that's your

1:05:48

conscience I have no problem with

1:05:50

hey it's free let's try it

1:05:52

out and if we can get

1:05:54

it to work then we'll pay

1:05:56

for it sure but pay for

1:05:58

it yeah pretty the money back

1:06:00

I think is important us, but

1:06:02

yeah, I think maybe we do

1:06:04

put too much on Google for

1:06:06

being the creators of the language

1:06:08

where it's like it can be,

1:06:10

it is a bigger and broader

1:06:12

thing. I mean, it's been for

1:06:14

years now, it's been more than

1:06:16

50% or around 50% of the

1:06:18

commits come from non- Google people

1:06:20

that go into the language and

1:06:22

the tool chain and the compiler. So

1:06:25

there's no reason why we can't be

1:06:27

like, hey, When you get, when you

1:06:29

have that, you can also work, have

1:06:31

them work on things that will make

1:06:33

you more. It's like, it could actually

1:06:35

be like a net zero for you

1:06:37

if you're people working on a language,

1:06:39

work on something that makes it so

1:06:41

your developers get more efficient, and then

1:06:43

you've basically paid for, you know, your

1:06:45

cost investment and how the community has

1:06:47

something that's better. So I think there's

1:06:49

some good arguments here. It's the same

1:06:51

thing with Caddy. Matt Holt, you know.

1:06:53

Caddy's being used more and more and

1:06:56

I don't think everybody thinks about engine X

1:06:58

right But there are a ton of companies

1:07:00

in Caddy because it's just easier and simpler

1:07:02

and it just works And they're not paying

1:07:05

paying Matt either right Matt right Matt? Matt

1:07:07

right Matt gets by every year and he's

1:07:09

and he loves this thing and that's what

1:07:11

he wants to work on and so He's

1:07:14

like okay in that sense right and at

1:07:16

Arden we did everything we could to help

1:07:18

Matt be able to sustain that to the

1:07:21

extent that we could right now there are

1:07:23

other companies that are helping them but

1:07:25

catty to me is no different than

1:07:27

tiny go right you're using it using

1:07:29

a production and you're not paying for

1:07:31

it like shame on you yeah that

1:07:33

brings a really good point I I was

1:07:35

at cockroach lives before this and I was

1:07:37

a hash corp before that and they used

1:07:39

to be sponsors of like go-for-con or whatever

1:07:42

they used to send money to these things

1:07:44

and now I don't see them on the

1:07:46

sponsor of this for Go for Con. I'm

1:07:48

not, I have no idea if they're, if

1:07:50

they're sending money to like the Go project

1:07:52

or to Tiny Go or to any other

1:07:54

things. But you are right, like these companies

1:07:56

make revenue off of this program

1:07:58

language and what? you just sponsored it

1:08:01

for a year and then you that's

1:08:03

it you left it high and dry

1:08:05

what's what's going on you know I'd

1:08:07

like to see I would like to

1:08:09

see those companies come back and give

1:08:11

money to the go community again to

1:08:13

some degree those companies also faced the

1:08:15

same problem though where like both of

1:08:17

those went through license changes because they're

1:08:19

like hey y'all are using this thing

1:08:21

for free and not giving money back

1:08:23

to us so I feel like it's

1:08:25

like everybody we all need to start

1:08:27

giving back to the things that were

1:08:29

basing the existence of companies off of,

1:08:31

right? There's so much, like everybody loves

1:08:33

to talk about open source, but then

1:08:35

every time you start talking about sustainability,

1:08:37

everybody's like, gee whiz, I don't know

1:08:39

how we're gonna do this. And it's

1:08:41

like, you know, lunchtime, I should get

1:08:43

out of here. It doesn't even have

1:08:45

to be a ton of money, right?

1:08:47

If a large number of people start

1:08:49

giving back money. then you can also

1:08:51

solve this problem. It's not have to

1:08:53

be everybody giving a whole developer. It

1:08:55

could be, you know, you know, $100,000

1:08:57

if a thousand, if a, well, yeah,

1:08:59

a thousand companies give a thousand dollars,

1:09:01

there you go. There's a hundred, I've

1:09:03

been giving Dominic $250 a month, for

1:09:05

God knows how many years for static

1:09:07

check. And at some point, GitHub was

1:09:09

matching it. Right. I've been doing that

1:09:11

for years, because I used that tool.

1:09:13

And I want him to continue to

1:09:15

continue to work on it. Right. What

1:09:17

kills me the most is developers cheating

1:09:19

other developers out of like, I could

1:09:21

see some, some, some higher up who's

1:09:23

not a developer, who's a bean counter

1:09:25

going, why am I going to pay

1:09:27

for that? I'm getting it for free.

1:09:29

But developers, cheating out, you want to

1:09:31

get paid for your work, don't you?

1:09:33

But they don't deserve to get paid?

1:09:35

Like, how do you sleep at night?

1:09:37

So like, you know, I've been funding

1:09:39

static check for, I don't know how

1:09:41

long it is. It's $250 a month,

1:09:44

dude. It's not like breaking my bank,

1:09:46

but that $250 goes a long way

1:09:48

to allow him to work on that.

1:09:50

And obviously, I imagine, he's got others,

1:09:52

but it's... just like you said Chris,

1:09:54

you got enough people giving you 200

1:09:56

bucks a month, you have what you

1:09:58

need to survive and everybody wins. I

1:10:00

think there might be a point to

1:10:02

it as well of like the. I

1:10:04

mean, I'm not going to say open

1:10:06

source should just go away, but I

1:10:08

think, you know, everybody wants to be

1:10:10

like, oh, these companies changing their licenses

1:10:12

so that they're not, you know, quote,

1:10:14

quote, quote, open source as it's defined

1:10:16

by the OSF, is like a really

1:10:18

bad thing. But I also think there's

1:10:20

an element of this that's like, maybe

1:10:22

we need to evolve how we view

1:10:24

all of this, maybe we to add

1:10:26

more nuanced and that needs to come

1:10:28

out in our in our licenseses to

1:10:30

come out in our licenses of like,

1:10:32

like, like, on the show who's a

1:10:34

program return lawyer, we talked a bit

1:10:36

about this idea of like, you know,

1:10:38

Open source licenses have always been kind

1:10:40

of a mess. These new licenses are

1:10:42

also kind of messy But we're like

1:10:44

working our way through them and we

1:10:46

kind of need to have like a

1:10:48

standardized model to like get it in

1:10:50

like for example The llama model from

1:10:52

or the llama, you know weights from

1:10:54

Facebook They say you can use this

1:10:56

for free up until you have 700

1:10:58

million users. Can't use this if you

1:11:00

have less than that great. And I

1:11:02

feel like if we had more space

1:11:04

for people to be like Well, how

1:11:06

successful are you actually being with this

1:11:08

thing? Or like, you know, maybe if

1:11:10

you are someone that just has a

1:11:12

little service, I'm fine for using for

1:11:14

free, but if you're bringing in like

1:11:16

a million plus dollars a year, then

1:11:18

no, you need to, you need to

1:11:20

give back of it. This is clearly

1:11:22

something that you're depending on. And while

1:11:24

the old model of like, okay, we'll

1:11:26

sell you a support contract or something

1:11:28

like that doesn't work for the smaller

1:11:30

developers. So I think we do need

1:11:32

to. be more inclusive of these other

1:11:34

licenses that do really have like a,

1:11:36

okay, no, it's not just a, you

1:11:38

get bad vibes from the community, but

1:11:40

like, no, your legal department's now gonna

1:11:42

have to get involved if you use

1:11:45

this wrong and someone finds out. And

1:11:47

I think that might start pushing people

1:11:49

and companies to start actually contributing back

1:11:51

a bit more because the legal department

1:11:53

is gonna tell you, like, oh, there's

1:11:55

a possibility of a lawsuit, how much

1:11:57

does this thing cost? it's going to

1:11:59

be $1,000 a month. No, go pay

1:12:01

for that thing. Like, I don't, we're not going to deal with

1:12:03

this because it's going to, it's going to cost us that if we

1:12:05

just get the lawsuit. Even if we can win it, we're going to

1:12:07

have to pay so much on legal fees that it's not going to

1:12:09

be worth it. That's not what's going to happen, Chris. It's going to

1:12:11

wait until we get a letter. Just keep using it until we get

1:12:13

a letter. We'll ask for forgiveness and we'll ask for forgiveness

1:12:15

and we'll ask for forgiveness and we'll write a check and we'll write

1:12:17

a check and we'll write a check and we'll write a check and

1:12:20

we'll write a check and we'll write a check and we'll write

1:12:22

a check. It could also happen that way.

1:12:24

Depends on the company. Depends on the people

1:12:26

in the company. But either way, at least

1:12:28

you have some recourse. You could send a

1:12:30

letter. And then it's like, OK, well, fine.

1:12:33

Here you go. We'll production. We give you

1:12:35

money. But I think it's like a thing

1:12:37

to shift the ethos. Because right now, the

1:12:39

ethos, I think, truly is. Oh, this is

1:12:42

free. I can just consume this. And there's

1:12:44

not even a. I don't even have to

1:12:46

think about giving this person any money. Right.

1:12:48

I think that's where so many developers, so

1:12:51

many companies go in of like, I just

1:12:53

got this thing off the street. I don't

1:12:55

have to think about giving them. But as

1:12:57

soon as it's like a, have to start thinking

1:12:59

about it. Doesn't matter. How do

1:13:01

you even know I'm using it? Right. Yeah,

1:13:03

they came on. At home at home. Like,

1:13:06

this is so complicated. But you should have

1:13:08

this conversation with Felico. We'll

1:13:10

talk so. I think

1:13:12

it's in the works.

1:13:14

I think our excellent

1:13:16

producer, Angelica, is working

1:13:18

on getting Filippo on

1:13:21

the show. And Ron, actually,

1:13:23

too. So I think both

1:13:25

of those covered. Clearly,

1:13:28

there has to be

1:13:30

a baton passing, which

1:13:32

we spoke about with

1:13:34

Kelsey High Tower previously.

1:13:36

So. One of the things that come to

1:13:38

mind when you're talking about you know handing

1:13:40

off to next generation or that baton passing

1:13:42

is How the heck do you get someone

1:13:45

up leveled enough to where they are a

1:13:47

senior engineer and ready to take the baton?

1:13:49

Because right now the industry has this problem

1:13:51

where? Companies are hiring people for the senior

1:13:53

level only and I put that in air

1:13:55

quotes because I don't even know what senior

1:13:58

even means anymore and we find the entry

1:14:00

level people coming to the market stuck,

1:14:02

you know, it's stuck in the entry

1:14:04

level and unable to achieve that senior

1:14:07

status. So I kind of want to

1:14:09

hear your thoughts on this matter. An

1:14:11

impossible situation. I have clients that would

1:14:14

love to hire some junior people, but

1:14:16

like I said, they feel like they've

1:14:18

got too much complexity in the environment

1:14:20

and they don't have enough time to

1:14:23

walk. this person. It's not that they

1:14:25

couldn't technically do it. I'm finding people

1:14:27

with less experience that I think are

1:14:30

better software developers than the ones with

1:14:32

experience, but they're feeling it's too much

1:14:34

risk because I'm going to put them

1:14:37

in a complex environment and these people

1:14:39

have never seen this level of complexity

1:14:41

before in architecture, design, and systems. And

1:14:43

I just don't have the time. So

1:14:46

it's almost like I don't care if

1:14:48

they know how to code and go

1:14:50

as much as Have they seen these

1:14:53

types of systems before and these types

1:14:55

of integrations? So I can plop them

1:14:57

in a chair and they already have

1:14:59

a mental model of some of the

1:15:02

chaos and they'll learn the programming language

1:15:04

faster than I can teach somebody else

1:15:06

the architectures and designs. And this comes

1:15:09

back to the idea that you've got

1:15:11

code bases that are five plus years

1:15:13

old now where people are moving on.

1:15:16

Nobody stays anywhere for more than two

1:15:18

years anymore anyway. We can have a

1:15:20

whole conversation about that. I used to

1:15:22

be anti that. I'm not anymore. It's

1:15:25

the only way you're going to get

1:15:27

a raise. You got a bounce. If

1:15:29

you, even if you're there three months,

1:15:32

if somebody's going to throw another $25,000

1:15:34

at you, you almost have a responsibility

1:15:36

if you have a family to take

1:15:38

that one. No, for sure. Yeah. As

1:15:41

long as the work is, it's work

1:15:43

that you're interested in and stuff, right?

1:15:45

And so when I'm doing an interview.

1:15:48

You bounced, you bounced, you bounced, you

1:15:50

bounced. And they say, more money, more

1:15:52

money, more money. Well, OK, that's fair.

1:15:55

I don't know how to solve that

1:15:57

problem, Matthew, because there aren't a. a

1:15:59

lot of projects that are starting with

1:16:01

new development where you can sort of

1:16:04

take on the risk and with the

1:16:06

job market so flooded with these engineers.

1:16:08

You said that companies are looking for

1:16:11

the kind of the experience engineers that

1:16:13

have seen it before, not necessarily like

1:16:15

can code in the language or whatever,

1:16:17

but have seen the systems, the architecture.

1:16:20

Do you define that as well as

1:16:22

being a senior? Is that what you

1:16:24

look for when you look for a

1:16:27

senior? I

1:16:29

hate these words. Senior, I hate them

1:16:31

too. I'm not gonna like you. I

1:16:33

really do hate them. You know how

1:16:35

many resumes I've seen when the person

1:16:37

said senior and I would barely even,

1:16:40

I wouldn't use the word at all.

1:16:42

Like I don't care that you have

1:16:44

15 years of experience, right? A senior

1:16:46

person to me is somebody that within

1:16:48

a month from my perspective, okay? If

1:16:51

I'm hiring for my own project, this

1:16:53

is somebody that's going to do two

1:16:55

things with me. One, I know within

1:16:57

a month they're going to need very

1:16:59

little of my time because they will

1:17:02

have understood the mental models that I

1:17:04

was teaching and the coding standards. And

1:17:06

that they are going to debate the

1:17:08

coding standards and debate the architecture and

1:17:10

come to me with these sort of

1:17:12

thoughts and ideas. I love debating software

1:17:15

design. I love debating these sorts of

1:17:17

things. If you're not capable of engaging

1:17:19

in these conversations when something comes up,

1:17:21

you're not senior to me. You'd be

1:17:23

a mid-level and maybe every once in

1:17:26

a while you ask the interesting question,

1:17:28

but I would expect over six months

1:17:30

that you're fully participating in debate as

1:17:32

much as, right? And debate is hard

1:17:34

for people because when you're in that

1:17:36

kind of- We just spent 30 minutes

1:17:39

in the beginning of the episode talking

1:17:41

about that concept. Debate requires strong personalities

1:17:43

with convictions that are willing to shut

1:17:45

up at times and listen. and reevaluate

1:17:47

what's going on. But if you're a

1:17:50

timid person, you don't feel like you

1:17:52

have enough, then you're not engaging in

1:17:54

that. And I've met guys that I've

1:17:56

been on my team where after they

1:17:58

tell me something and I'd be like,

1:18:00

why didn't you say that during the

1:18:03

meeting? That was interesting. That would have

1:18:05

added value, but they were afraid, right?

1:18:07

Because you got Alpha Bill and Alpha

1:18:09

Powell, growing at it like this, because...

1:18:11

And at the end of the day,

1:18:14

what we're really trying to do is

1:18:16

figure out what's the best solution. And

1:18:18

all my teams know this. If you

1:18:20

ask me to do something that I

1:18:22

can't teach or put back into the

1:18:25

service project, we can't do it here.

1:18:27

We're not going to do it. That's

1:18:29

the measuring stick. If I don't feel

1:18:31

comfortable teaching this, how the heck can

1:18:33

we do it at all? We can't,

1:18:35

right? That's the measuring stick. There have

1:18:38

been times where we've debated something. not

1:18:40

full two or three days, but an

1:18:42

hour here or there. And finally, I

1:18:44

have to make a decision, like, no,

1:18:46

this is what we're going to do.

1:18:49

But to me, that's the senior person.

1:18:51

They can come in, they understand the

1:18:53

architecture design, they understand the business problem,

1:18:55

they understand the problem that's in front

1:18:57

of us, and they can think out

1:18:59

solutions that maintain the architecture, maintain the

1:19:02

coding standards, maintain these things, and be

1:19:04

in debate listening. but also somewhat forcefully

1:19:06

sharing ideas because their opinions and they

1:19:08

have to be strong at the same

1:19:10

time you have to know how to

1:19:13

back off. There aren't a lot of

1:19:15

people that can do that. That to

1:19:17

me is the senior person. And I

1:19:19

have people around me, a guy named

1:19:21

Pavel, Florin. I love these people. I

1:19:23

love when there's a debate going on.

1:19:26

I think I spoke to Pavel before.

1:19:28

Yeah. Like sometimes I'm like, Powell, I'm

1:19:30

too tired right now, dude. I can't

1:19:32

do this. He's like, oh, we got

1:19:34

to talk about it. I do, let's

1:19:37

talk about tomorrow, man. I'm tired, like,

1:19:39

you know. But if you can't do

1:19:41

that, you're not senior. It's not your

1:19:43

coding skill. It's not anything else other

1:19:45

than mental models. You understand the arc.

1:19:48

texture, you see it, you understand the

1:19:50

problem, you understand why we have the

1:19:52

problem, and you've got some thoughts about

1:19:54

solutions, but we're willing to trade those

1:19:56

ideas back and forth. Like, I don't

1:19:58

know what the industry means by a

1:20:01

staff engineer, to be honest with you.

1:20:03

Like, I don't understand the industry term,

1:20:05

because I'm not in that section of

1:20:07

the industry, but I imagine a staff

1:20:09

engineer is similar to what I'm sort

1:20:12

of describing, right? They can, they understand

1:20:14

architecture and design. They know that if

1:20:16

you touch this, that is going to

1:20:18

break over there. And so we really

1:20:20

can't do that unless, right? Yeah. They

1:20:22

see the holistic view of the systems.

1:20:25

Yeah. I've seen like Staff Plus as

1:20:27

the people that can not only just

1:20:29

like know the multiple systems either from

1:20:31

being there a long time, but can

1:20:33

also just kind of enter a space.

1:20:36

and figure out those dynamics really quickly,

1:20:38

and then be able to be able

1:20:40

to like, okay, yeah, like I understand

1:20:42

now if you touch this, this is

1:20:44

things going to blow up, but be

1:20:46

able to do that in like a

1:20:49

matter of weeks, whereas I don't think

1:20:51

you can really expect a senior engineer

1:20:53

to be able to do that. Maybe

1:20:55

for like a single project, but if

1:20:57

we're talking about like a whole, yeah,

1:21:00

for a whole system, I think that's

1:21:02

more of like the staff, senior staff

1:21:04

principle. That's when you expect those people

1:21:06

to be able to be able to

1:21:08

be able to be able to be

1:21:11

able to be able to be able

1:21:13

to be able to be able to

1:21:15

walk in and do that sort of

1:21:17

stuff. But that also requires like a

1:21:19

completely different mindset of how you like

1:21:21

look at and evaluate a system and

1:21:24

then at that point really is not

1:21:26

about code at all It's about like

1:21:28

what is this thing doing and where

1:21:30

are all the parts of it and

1:21:32

I that scares me Chris because How

1:21:35

if you're not engineering from the from

1:21:37

the beginning thinking about these things then

1:21:39

you're not getting there ever right like

1:21:41

yeah I tell Kevin all the time

1:21:43

I'm playing chess all day long on

1:21:45

this code base all day long I'm

1:21:48

trying to think what's going to break,

1:21:50

what's going to break, what's... You're worried

1:21:52

about this. I'm already four moves ahead,

1:21:54

right? All day long. But that comes

1:21:56

with time and practice and putting bad

1:21:59

things in production. So I

1:22:01

don't know how you get to that level

1:22:03

you're talking about, either at a project level

1:22:05

or above, until you've walked that engineering journey

1:22:07

and made all those mistakes to know that's

1:22:09

going to be a problem. Like I don't

1:22:12

know how you can not have been a

1:22:14

developer and get there. I'm not saying that

1:22:16

maybe you haven't coded for 10 years and

1:22:18

you don't know go and that's fine. Right.

1:22:20

I think that's interesting right because I when

1:22:22

I look at my my own career and

1:22:25

I've said this in multiple different ways but

1:22:27

it's like I don't there wasn't a point

1:22:29

where I was like a junior mostly because

1:22:31

of circumstance like every job that I had

1:22:33

probably every job I've ever had in the

1:22:35

in the software community I've been the most

1:22:38

knowledgeable on whatever technology we're using or one

1:22:40

of the most knowledgeable. Like, even for my

1:22:42

very first job when I was a droopal

1:22:44

developer, I walked into a space and it

1:22:46

was like, I was there to fix the

1:22:48

problems because nobody else understood the technology we

1:22:51

were using. So there wasn't anybody I could

1:22:53

be like, help me along. This is when

1:22:55

you got to sit down, you got to

1:22:57

understand all of this. And like, once again,

1:22:59

I didn't enter this from someone that, from

1:23:01

being someone with a CS degree or anything

1:23:04

like this degree or anything like that. Try

1:23:06

to figure out, like, how do you replicate

1:23:08

that with other people? And I'm not really

1:23:10

sure. I think part of it comes from

1:23:12

the fact I have a writing degree. and

1:23:14

when you are a writer, when you are

1:23:17

a storyteller, you have to, you can't just

1:23:19

sit down and write, you have to take

1:23:21

the broader picture and understand where you're going,

1:23:23

what your story universe is, all of that,

1:23:25

and then you can get into everything. And

1:23:27

I think that applies to this as well.

1:23:30

Like if I go out and get another

1:23:32

job, I'm gonna sit down and start looking

1:23:34

at all of the stuff. and then I

1:23:36

can go down and look at that one

1:23:38

code base. And I think that's not the

1:23:40

approach that we give people when they enter

1:23:43

the industry. I think I had a very

1:23:45

unique experience, but I think what most people

1:23:47

get is the. Well, we'll give you this

1:23:49

one section of this one could basically begin

1:23:51

with and then we'll slowly level you up

1:23:53

to the point where now you have a

1:23:56

whole project or now you have multiple projects.

1:23:58

But I think by the time that someone

1:24:00

gets to that point where you're giving them

1:24:02

a whole project, it's been so many years

1:24:04

that they're just used to that other way

1:24:06

of working of that. Here's the box. Like

1:24:09

I've seen this in some engineers I've worked

1:24:11

with where it's like, well, this is my

1:24:13

box. I don't, I don't really know what's

1:24:15

going on any of what's going on any

1:24:17

of what's going on any other surfaces as

1:24:19

well in the, oh, I don't want to

1:24:22

have to talk to that other team about

1:24:24

their stuff to integrate. Can we just do

1:24:26

this in our box instead? So I think

1:24:28

like the way that we've traditionally trained software

1:24:30

engineers has hindered our ability to really get

1:24:32

the number of senior and staffless engineers that

1:24:35

we need because we tell them this is

1:24:37

your box, please stay in your box. And

1:24:39

sometimes the politics of companies also force you

1:24:41

to kind of stay in your box. And

1:24:43

that's just, I don't think that that that's

1:24:46

a path toward being able to being able

1:24:48

to being able to being able to learn

1:24:50

to think from the completely other direction at

1:24:52

the completely other direction at the other direction

1:24:54

at the end of the end of the

1:24:56

end of the end of the end of

1:24:59

the end of the end of the end

1:25:01

of the end of the day. So I

1:25:03

have the luxury with Kevin where I don't

1:25:05

code anything without him. I pair on everything.

1:25:07

Sometimes I drive because I need to move

1:25:09

a little faster that day. Other times I

1:25:12

have him drive. He drives me crazy because

1:25:14

he uses the mouse too much. I'm trying

1:25:16

to get him to learn the keyboard and

1:25:18

there are days where I have patience for

1:25:20

it and there are days that I don't

1:25:22

and that's on me to have more patience.

1:25:25

Okay. For me, the only way I have

1:25:27

found to be able to get Kevin up

1:25:29

to speed at the level I want to

1:25:31

get him, at some point, I want Kevin

1:25:33

to be able to work at a client,

1:25:35

right? I feel comfortable doing that. Is to

1:25:38

pair with him on just about everything that

1:25:40

I'm doing. So I can dump my brain

1:25:42

on him. This is what I'm thinking about.

1:25:44

This is what I'm concerned about, right? I

1:25:46

don't know what the answer is here because

1:25:48

I don't know enough yet, but... I don't

1:25:51

know how else you get that without some

1:25:53

level of pairing. Now I'm saying you should

1:25:55

be pairing. Several years ago, Volkswagen had a

1:25:57

team in Berlin. That team did only pair

1:25:59

programming. So every two, and

1:26:02

it was different developers,

1:26:04

but you weren't allowed to sit

1:26:06

down and code unless you were

1:26:08

pairing with somebody else. Now,

1:26:11

it was an interesting idea.

1:26:13

I think they went too

1:26:15

extreme because nobody was allowed

1:26:17

to touch code at all

1:26:19

without it. Now, what I

1:26:21

would have liked to see is

1:26:23

a mix of, I want a number

1:26:25

of hours a week where you're pairing.

1:26:28

and then where you're sort of working

1:26:30

solo and then come back to do

1:26:32

the review as a pair instead of

1:26:34

that dang PR thing. I hate those

1:26:36

freaking PRs. I can't read them. That

1:26:38

to me is the final sort of

1:26:41

step there, right? Even with anybody I'm

1:26:43

managing and even on client work, it's,

1:26:45

let's sit down and look at the

1:26:47

changes together, not at the PR level.

1:26:49

And here's another thing I do, which

1:26:52

is crazy, with my seniors. I tell

1:26:54

them, if you see a piece of

1:26:56

code. that we're working on and it's

1:26:58

not right. I don't want to comment,

1:27:00

fix it. I don't want to

1:27:02

comment, fix it. What are we doing?

1:27:05

I don't own that code anymore

1:27:07

than you own that code. I

1:27:09

have time for that. Fix it.

1:27:11

Now for the other developers who

1:27:13

haven't earned that yet, then I tell

1:27:15

them we're going to sit together and

1:27:18

look at it. If you see something

1:27:20

you don't like, change it, but that

1:27:22

has to stay. in the PR, that

1:27:24

can't go to Maine without me,

1:27:27

right? On the other developers, that

1:27:29

stuff can go to Maine all

1:27:31

day. Like, let's not slow it

1:27:34

out. You, your stuff can't

1:27:36

go into Maine. Fix it,

1:27:38

and then we'll review it

1:27:40

before anything gets. I don't

1:27:42

know how any other way to maintain

1:27:45

velocity and then get

1:27:47

up speed without some

1:27:49

level of pairing. Yeah, I think,

1:27:51

I think. some degree of it. It's not

1:27:53

pairing in the same way, but I've been a big

1:27:56

proponent. I've been talking so much about this that some

1:27:58

people are like, shut up, but I'm like... You

1:28:00

know, the whole idea of writing more stuff

1:28:02

down and archiving things and really being able

1:28:04

to be like, here's a manual, and not

1:28:06

just like throwing a manual, someone. I think

1:28:08

that is completely useless as an idea. But

1:28:11

give people the, spend the time teaching people,

1:28:13

here's how to, if you have a question

1:28:15

about something, if you're trying to learn something,

1:28:17

if you want to learn a piece of

1:28:19

the code base. This is where you go

1:28:21

in this manual. Like I think of how

1:28:23

it works for like I have seen myself,

1:28:26

the Chicago Manuel style, or any style manual,

1:28:28

and these things are like, you know, Chicago

1:28:30

manual style is like a thousand pages long,

1:28:32

but any publishing house says they're manual of

1:28:34

style. And if you want to write a

1:28:36

piece of writing for them, you got to

1:28:38

write it in that way, but they're also

1:28:40

structured so that if you just know how

1:28:43

manuals of style work, you can sit down

1:28:45

and be like, oh. I need to know

1:28:47

how to do, like, cite this in one

1:28:49

way or how to do this quoting or

1:28:51

this thing. You know where to look in

1:28:53

the manual to find that information so you

1:28:55

can do it in the right way. And

1:28:58

I feel like we as an industry are

1:29:00

missing all of that stuff. And I feel

1:29:02

like when we try to make it, it

1:29:04

usually comes in the form of like, here's

1:29:06

a file you can put so your editor

1:29:08

will force you into doing these things. And

1:29:10

I feel like your approach where you're like,

1:29:12

okay, I want to sit down and pair

1:29:15

with someone, I want to connect on a

1:29:17

human level and do this knowledge transfer, like

1:29:19

being able to do that, but also being

1:29:21

on the other end to scale it up

1:29:23

with. Now I've sat with you a little

1:29:25

bit, maybe for a few weeks or a

1:29:27

month, now you have the knowledge you need

1:29:30

to be able to be able to be

1:29:32

able to be able to be able to

1:29:34

be able to be able to be able

1:29:36

to be able to be able to go.

1:29:38

use that manual to learn more, to do

1:29:40

more, to do things in the way that

1:29:42

we do them here and whatnot. I feel

1:29:45

like that would be a better balanced approach

1:29:47

than what we've been doing. I have tried

1:29:49

to write style guides. I've tried to do

1:29:51

this. It is so tedious. It's not easy,

1:29:53

man. And I gave up. I kudos to

1:29:55

you for being able to do it. I

1:29:57

gave up. And that's why I went with

1:29:59

the pairing approach. So anybody working on the

1:30:02

code bases that I'm working on. You pair

1:30:04

with me until we get to a point

1:30:06

where I can trust that you don't need

1:30:08

the pairing. And even sometimes... something really complex.

1:30:10

Florn and I was working on something complex

1:30:12

last week. I'm like, dude, we need to

1:30:14

work on this together. And I'll be honest

1:30:17

with you, I know I couldn't have solved

1:30:19

it alone and I know he couldn't have

1:30:21

solved it alone. Yeah. So I'm a, I

1:30:23

love, you know, what would be my favorite

1:30:25

job if I can ever retire from Arden?

1:30:27

And retiring might be selling or just handing

1:30:29

over the range to some other people and

1:30:32

I don't have to deal with it day

1:30:34

and day and day out. I would love

1:30:36

a job. There's just no way to, there's

1:30:38

just no way to. There's no I would

1:30:40

be fired at the end of every year

1:30:42

because my performance review I'd have nothing for

1:30:44

performance review. Okay, but at least for a

1:30:46

year this would be fun I'm gonna go

1:30:49

to a company and every week pair with

1:30:51

a different team My job on every day

1:30:53

is to sit next to somebody and help

1:30:55

them code whatever it is they're working on

1:30:57

I get no credit for it because all

1:30:59

the commits are in their name And after

1:31:01

a few days with this person, I'll go

1:31:04

to another person and maybe stay with a

1:31:06

team for a month and get them all

1:31:08

up to speed on everything that we've talked

1:31:10

about and then move on to the next

1:31:12

company, move on to the next team. See,

1:31:14

the problem is when the code of you,

1:31:16

when the... performance review comes up and they

1:31:19

say, what did you do? You're like, well,

1:31:21

I didn't really do anything. Look, I got

1:31:23

no commits. Like, I have no responsibility. That's

1:31:25

the best part, right? The things breaking at

1:31:27

three in the morning ain't calling me. But

1:31:29

I would love that job. I would love

1:31:31

somebody just pay me to come to work

1:31:33

every day and pair with somebody and just

1:31:36

depart all my thoughts and what I'm thinking

1:31:38

about and get up to speed. I mean,

1:31:40

companies hire me to do that remotely. Right,

1:31:42

and I do some of that already, but

1:31:44

that to me would be the most amazing

1:31:46

job. Yeah, I mean, I feel like that's

1:31:48

a thing that companies need more than they

1:31:51

think they need. I feel it's a failing