Episode Transcript
Transcripts are displayed as originally observed. Some content, including advertisements may have changed.
Use Ctrl + F to search
0:00
Welcome to changelog and
0:03
friends, a weekly talk
0:05
show about Jamie Tana's
0:07
website. Thank
0:21
you to our partners at
0:24
fly.io, the home of changelog.com.
0:27
It's your app and you're your users all around
0:29
the world. Find out how
0:31
at fly.io. Okay,
0:33
let's talk. We
0:37
are here with Jamie Tana, longtime
0:40
listener, community member, blogger. I've
0:42
been publishing your stuff
0:45
in changelog news over the years, it seems like
0:47
at this point. And one
0:49
of our voice mailers for last year's
0:51
state of the log, Jamie, welcome to
0:53
changelog and friends. Thank you.
0:56
Not that I'm saying he's not a friend Jared, but I
0:58
was thinking maybe he wasn't really a friend. It's
1:00
more like this is changelog and friendly. I
1:02
don't know. Because he'll become a friend during
1:04
the show potentially, right? Like I was thinking about that on the
1:06
way here. What do you think about that? I feel like he's been
1:10
like a community member, hanging out in
1:12
the slack. We haven't held hands yet
1:14
though. So you know, well, he lives across the
1:16
pond. It's not easy. You haven't with
1:18
Gerhard for ages. That's true. It
1:21
took us a long time to shake his hand and give him a hug.
1:24
So true. But change log and friendlies, you
1:26
know, it's got a ring to it. Maybe we could have a,
1:28
a sub show. I guess you're looking for
1:31
a new subtitle. Maybe it's turning friendlies
1:33
into friends. There you go. Oh, I like
1:35
that. That's actually not that bad. Well, I
1:37
would like to hear Jamie is how it
1:39
feels to get your very own break master
1:41
cylinder remix. And because you're, you
1:43
are a recipient of a remix. I've never got
1:45
one for myself, but I
1:47
feel good. Doesn't it? How did I feel
1:49
hearing your voice all remixed? It was very cool. I
1:51
mean, the fact that for the second year running, I
1:54
ended up submitting my voice note like state
1:56
of the log was just impressive given
1:59
nothing. It's like a month
2:01
before the deadline when is the deadline
2:03
and I write an update So
2:06
yeah, I'm just still surprised that you managed to
2:08
squeeze me in that was very cool I'm
2:10
gonna get my own what remix was pretty
2:13
cool. I haven't yet said it as a ringtone, but maybe
2:16
that's the next step He should
2:18
is Jamie's the sweet robot
2:20
dance makeup music. That's Jamie's right. I
2:22
was the other Jamie. That's Jamie, Kurno
2:24
Okay, Jamie's was the very last one
2:26
cuz he submitted late remind me what
2:28
it says We'll see if we can
2:30
play it you want the remix or the message
2:33
the remix I
2:35
am My podcast update I'd absolutely
2:37
say that I've often prioritized change
2:39
log over cooking over my
2:41
life I don't have time to
2:43
walk the dog this year so
2:46
it is a lot I've
2:48
listened to change log for five days
2:50
at a time in a room just
2:53
laughing having a great time and Really
2:58
looking forward to what's happening over
3:00
next year Thank
3:06
You Jamie that was amazing so I
3:09
bet your dogs pretty sad I know my
3:11
dog like absolutely needs the
3:13
Adam tank filled up every single day the
3:16
dog begs me to go to work with me because you know
3:18
for now I go to a separate office and My
3:21
dog could not not be walked and
3:24
then at the same time I need to cook and
3:26
I've never sat in a room for five days straight laughing at
3:28
cheese on podcast So
3:31
you've got me beat but I appreciate
3:33
that. I'm just dedicated Yeah,
3:35
not okay. I take it back
3:37
your friend now There we go.
3:39
Just like that. We already did it We haven't even gotten
3:41
into the show yet and we've already turned a friendly into
3:44
a friend. So too easy I
3:46
was digging into my site
3:49
history and I found at least the first recorded
3:52
time I Tracked a
3:55
listen to change log was 2016. So I've
3:57
been a lot of a good chunk
3:59
of time So you track your
4:01
listens? I do. So it's
4:04
something I started two years
4:06
ago was starting to track
4:09
like every time I listen to a podcast, I'll
4:12
publish a listen to my website.
4:15
And as part of that, I then have a history
4:17
of all the things I've listened to. But
4:19
because I was only doing that from like that
4:21
point of time, I also had like a
4:23
load of history. So I found a
4:26
way to take podcast, add it SQLite
4:28
database, reverse engineer some of the
4:30
data out of that and backfill
4:32
that to my site as well. In
4:35
fact, I'm just now trying to charge up some of my
4:37
old phones to see if they've got
4:40
any history on them. I can go further
4:42
back. I can also backfill. Yeah. Interesting. That's
4:44
dedication. I love it, but I also wonder
4:46
why. Like what drives you to want
4:48
to do that even so much so that you'll go
4:50
into the history and like dig it out of a
4:52
database. Why? In 2018, I
4:56
found out about the Indie Web Community. The
4:59
community is all around like owning your data, not
5:02
relying on other platforms. So
5:04
some of it is like Indie Web is a thing
5:06
is quite cool. But it's
5:08
also just like interesting for me to have that
5:10
data. So coming up to today,
5:12
I was able to easily go
5:14
back and say the first time I definitely
5:17
know I listened to a Change. I listened to a new podcast. It
5:20
was the 27th of August and it
5:22
was request for commits number four. And
5:25
that was actually the first podcast of the change log
5:27
I listened to. It's
5:29
also interesting because that gets
5:31
syndicated to the Fedverse. So
5:34
if people are interested, they can see the sorts
5:36
of things that I'm listening to. They
5:39
can find out about new podcasts as I've listened
5:41
to them and stuff like that,
5:43
which I find quite cool as just a
5:46
passive way of sharing it. Because
5:48
I have all that data or I can go in and I
5:50
can work out like, actually, in the
5:52
last year, I have listened to five days
5:54
worth of change log. Yeah, exactly. Yeah.
5:58
And you can make sweet robot dance make out music.
6:00
But was it with that information, you know, it's amazing.
6:03
So okay, that is cool. I admit
6:05
it and that makes sense
6:07
the dedication the syndication so
6:10
indie web this is the idea of own your
6:12
own data published through your own website and Syndicate
6:15
elsewhere or like blast out. I guess there's a
6:18
couple different ways of doing it, but that seems
6:20
to be the main one. Yes So
6:23
there's a really good article Titled one year in
6:25
the indie web and this is the one that
6:28
I usually recommend people read because there's
6:30
a really good distinction There between
6:32
indie web with a capital I and
6:34
capital W Versus indie
6:36
web or lowercase or with a space between
6:39
indie and web So there's a lot of
6:41
people who think like indie web is I
6:43
need to do these particular things To
6:45
be part of the indie web whereas
6:47
like at its base
6:49
It's just you own a URL
6:52
and you publish content to it It
6:54
could be that URL is in
6:56
front of sub stack or medium could
6:58
be in front of WordPress calm or it could
7:01
be infrastructure that you are hosting in your
7:03
home lab and You've
7:05
got a load of stuff for that But
7:07
then as well as that it has
7:09
also morphed into things like owning
7:11
data across different services So
7:14
one of the things you references posse
7:17
publish on your own site syndicate
7:19
everywhere else or elsewhere and
7:21
so that is a lot of what I do I
7:24
Try to do which is in
7:26
the before times before Twitter's API
7:28
got taken down. I would reply
7:31
to people's tweets from my website
7:33
and So I would actually
7:35
go through a lot more pain and effort
7:37
to reply to someone's post which kind of
7:39
showed that I actually did A little
7:42
bit more because I would like
7:44
using so the Twitter app I would
7:47
share With an indie web app.
7:49
I would write my reply or then
7:51
hit send that would go to my website In
7:54
about 10 minutes. My site
7:56
would have rebuilt it was sent a
7:59
notification to an app called Bridgie. Bridgie
8:01
would then say, oh cool, this is a reply
8:03
to something in Twitter. So I'm now going
8:06
to send an API request to Twitter. I'm
8:08
going to syndicate that post. And that will then
8:10
appear on twitter.com. And in
8:12
all that time, like someone's probably not
8:15
bothered by my reply anyway. They're
8:18
going to lose an argument. By the time
8:20
your reply comes in, they're like, they moved
8:22
on. They don't know what you're talking about.
8:24
But I've also found it quite good because
8:26
it means that, A, those sorts of arguments
8:28
do calm down a little bit. Because, yeah,
8:31
I'm going to be taking hours to reply
8:33
to you. In fact, there was a
8:35
time where a couple of friends of mine messaged me and
8:37
was like, why have you just replied to
8:39
my tweet from six months ago?
8:43
I had a bug in my site software that
8:45
it wasn't actually sending those notifications called
8:47
web mentions for a while. So
8:50
suddenly I was replying to posts from six months
8:52
ago. Yeah,
8:55
not great. That's a good response though, that was a bug
8:57
in my software. This reminds me a
8:59
little bit, somewhat, so go
9:01
with me here, of back
9:03
when I was religious about updating the
9:05
song titles in iTunes, like back whenever
9:07
you used to manage your own MP3
9:09
library. And it's not exactly
9:11
a one to one, but I wonder if, you know,
9:14
this indie web, which I totally agree with, I love
9:16
the indie web, you know, like RSS feeds, like that's
9:18
what we do for a living. We ship MP3s via
9:20
RSS somehow to the world, and we
9:22
don't care about tracking. I mean, like we do to some
9:24
degree, but just enough to know that there's signal, not noise.
9:27
But it kind of reminds me of like, this world
9:30
where you pay attention to the details so
9:32
closely, but others may not. And so the
9:34
world doesn't really get populated with others like
9:36
you, who care so deeply about
9:38
all the layers. Do you feel like that?
9:41
Like, do you feel like that the indie
9:43
web has a possibility of
9:45
being more thriving because there's more folks who are
9:47
so purist like you are? And
9:49
am I wrong on my iTunes naming? Cause I mean,
9:51
I used to be crazy about it before they would
9:53
like do it for me. And now just
9:55
there's Spotify and other things. So yeah,
9:58
I do see it as a, of
10:00
a parallel. So I definitely wouldn't say I'm
10:02
one of the purists. There are
10:04
a number more in the community who will do a
10:07
lot more work, some of this stuff. And
10:09
for me, some of it is just like
10:11
the muscle memory of doing it. And
10:14
a lot better like, well, I've been doing it
10:16
for this long. So sunk cost fallacy, I have
10:18
to keep doing that for sure. Yeah, that's why
10:20
I'm about I've been doing this rename you for
10:22
so long, I am going to keep doing
10:24
it. So
10:27
for instance, like on websites like
10:29
lobster and Hacker News, and even
10:31
things like LinkedIn, I
10:33
don't have RC setup. So I
10:35
will just reply and it is kind of
10:37
nice to sometimes just reply and
10:39
it'd be done. But also I, I
10:42
find it's really useful to look
10:44
back on things. So for instance,
10:47
I have ADHD. And as part of
10:49
that, my memory isn't great. I blame
10:51
the ADHD for that. And
10:53
there's something in 2017, I started
10:56
doing school block implementation.
11:00
So it's blogging as a form of
11:02
documentation. And I very
11:04
recently started a personal website.
11:07
And as far as block implementation, it was a
11:09
case of I've just learned a thing, or I've
11:11
just done a thing. And it
11:13
was kind of awkward to like Google around for
11:15
and find a solution for. So I'm
11:18
going to be that person who writes a blog
11:20
post about it. And even if it
11:22
doesn't help anyone else, it will have
11:24
helped me because I would have written about it. And next
11:27
time I can go back to it. And
11:29
let partially or I,
11:31
I don't see this as like a really big
11:34
positive for me as someone with ADHD
11:36
is I now don't have
11:38
to try and remember how to do the
11:40
thing. I can now just vaguely
11:42
remember, have I written about the
11:44
thing? And then I can go and find that
11:46
in my website. Or I
11:48
don't even need to do that far,
11:50
I can just search through my site. And I
11:52
can say, I'm sure recently,
11:55
I was like talking to someone about this,
11:57
like how to do contract testing with open APC.
12:00
And I can search through my
12:02
site and then I can find a reply
12:04
that I posted somewhere. And I find that
12:06
really useful because part of
12:08
the blogmentation was a group by a
12:10
discussion with Scott Hanselman who
12:13
had a site called Keys Lift, which
12:16
counts down the number of
12:18
keystrokes you possibly have based
12:21
on how many keystrokes you have
12:23
left until you die of natural
12:25
causes. And in
12:27
this podcast he was talking about, I don't want
12:30
to waste those keystrokes on something that, like,
12:32
if he gets a dozen emails about how
12:34
do I do this thing, he wants to
12:36
be able to write that reply once and
12:38
then save the keystrokes and put
12:40
them towards something more useful. And
12:44
I very much internalize that of,
12:46
let me try and reference things
12:48
that I've already written. So,
12:50
in fact, you can see behind
12:52
me on the video, I have
12:54
my URL in the background because it's
12:57
very much like my personal brand. One
13:00
of my friends for my birthday,
13:02
the year before last, got
13:05
me a t-shirt which says, I have
13:07
a website, jvt.me, because I
13:09
would just talk about it so much
13:11
that it was like that sort of fun thing. And
13:14
another friend of mine from
13:16
a previous company, he built
13:19
a website called didyouknowjamiehasawebsite.co.uk because
13:22
it was just so common that I would talk about
13:24
my website and I would share links to things and,
13:26
like, a lot of it was coming from the place
13:28
of, I've written about this thing,
13:31
I have helped solve the thing, so here
13:33
you go. I have
13:35
a lot of thoughts about that. I
13:37
agree absolutely on the blogumentation front and
13:40
on the, as a completionist,
13:42
it appeals to me to have everything
13:44
always. But
13:46
some pushback. Well, actually I
13:49
pushed back on you about the same thing, Adam, at
13:51
that conference and we were talking about letting the
13:53
LLMs, like, know all of our thoughts, you know.
13:55
Just, like, write everything down. And
13:58
I think at that time I said, is there Is there
14:00
no value in that which is ephemeral and
14:02
I used to be like always
14:05
be recording always be publishing if it's not
14:07
forever It's not meaningful and I'm
14:09
starting to just like change my perspective on that Maybe
14:11
the kids are changing it for me because you know
14:13
the younger generations value ephemerality more
14:15
than we do because they've
14:17
grown up where like everything they do is Documented
14:20
and tracked by themselves or
14:22
by their parents or by the NSA or
14:24
whatever it is, right? And
14:27
I just feel like life has there's room
14:29
in life for both things Where
14:32
there are some conversations that
14:34
don't need to be documented. There are things that
14:36
a reply on LinkedIn
14:38
can just disappear into LinkedIn's
14:40
database somewhere and doesn't necessarily need to
14:42
be cataloged for you or for
14:45
your language model to remind
14:47
you of later and I
14:49
think the picking and choosing what I
14:51
should keep and what can just go away I
14:53
think is part of the process But I'm just
14:55
kind of like letting go a lot more than
14:57
I used to I was like does this
14:59
conversation really need to be Published or
15:02
was it just you know, I
15:05
ran into somebody in the street and we talked for
15:07
five minutes about our families And we moved on for
15:09
lives and that was just a really nice human thing
15:11
And maybe I said something there like I have a website
15:14
or I wrote about this But
15:16
beyond that I don't think
15:18
it needs to have artifacts to be valuable. Yeah
15:21
Well, I'm with Jamie on one front because I just
15:23
had a conversation with somebody for the first time We've
15:25
been looking forward to meeting each other. We met each
15:27
other at a mutual party for the Super
15:29
Bowl it's a very popular thing in the US and The
15:34
conversation was around nerdy things and I guess
15:36
this person is a lot like me where
15:38
they don't meet other people that can go
15:41
Deep nerd on Linux and Unix and BSD and
15:43
like ZF and whatever whatever the fun tinkering things
15:45
are There's not many home libraries out that you
15:48
can meet just like randomly in the street. So
15:50
that's the point and This
15:52
person is a professional in IT around
15:54
the Windows system primarily and
15:57
through conversation. We're talking about our set up
16:00
at home lab and whatnot. And I'm like, well, I
16:02
really don't do this, but I did
16:04
podcast with Matt Arons, one of the
16:06
co-creators of ZFS. He named dropped
16:08
on it. A couple years back and a lot of people liked it.
16:10
So you should go check it out if you're really wanting to go
16:12
deep on it. It helped me out a lot. Yeah. And
16:15
then like a minute or two later, we're talking about
16:17
something else. And like, well, I also talked to, so
16:19
you're the, by the way, I have a podcast. We
16:21
also talked to Alan Jude recently about free BSD and
16:24
all the BSDs and like the whole history there. And it was
16:26
really cool. So you probably enjoy that too. So
16:29
I kind of felt myself like
16:31
self-referencing things I've done, because we do document
16:33
a lot of things that are, we nerd out about, right?
16:35
That's just natural for us, but I get you on
16:38
that point. I think the point you're trying to make Jared is
16:40
that there is value in just, just simple
16:42
humanity, not journalistic
16:45
or must be recorded to then
16:47
share later aspects.
16:49
And that's, I don't
16:51
disagree with that. Right. I fully agree with
16:53
that. And having a filter, you know, I have lots of thoughts throughout
16:56
the day, as you guys do, I'm sure. Most
16:58
of them aren't worth keeping, you know? And
17:01
we make those choices. You know, Adam, I know you have lots of
17:03
ideas. Some of them you write down, other
17:05
ones you don't. Yeah. You
17:08
know? And knowing where to draw those lines is just
17:10
something I'm more, I guess, cognizant of than I have
17:12
been in the past, where I was more like, if
17:14
it doesn't live forever, it's not of any value. You
17:16
know, like you, you have, I didn't understand Snapchat. I'm
17:18
like, why would you want to take a picture that
17:20
just disappears? Like, it doesn't make any sense to me.
17:22
And now I get it. I guess I've just evolved
17:24
a little bit. I don't get that either. I
17:26
just can't get it. I
17:29
was never there, so I don't get it. Fair enough. But
17:31
yeah, we should get Adam a t-shirt. By the way, I
17:33
have a podcast, you know? It's in a little bit of
17:35
a rush. I have a website. Yeah, that
17:37
would be cool. And I don't like
17:39
doing that, though, necessarily. When I meet people, I'm not like,
17:42
in this case, it was proper. It wasn't like
17:44
boasting or self-promoting. It was like literally, hey, we're
17:46
talking about this, and
17:48
this happened. So if you are curious at all
17:50
about ZFS, this is a great
17:53
resource of many you should check out.
17:55
And it enlightened me. So you should check it out if
17:57
you are some client. It
17:59
wasn't like. Let me like slack in the
18:01
face of my podcast and I guess it is hard
18:03
to like drive that balance of my
18:05
you both have talked To loads of interesting people
18:08
you've got loads of interesting thoughts and people can
18:10
learn a lot from that But also
18:12
does it feel like I'm just pushing you
18:14
to consume all my content? Yeah
18:16
So I would say if Jared and I
18:19
were in quotes influencers Then
18:21
then when we did that it would be
18:23
disingenuous But because we are
18:25
definitely not in quotes influencers. We
18:27
might be influencers with a lowercase I Just
18:30
because we tend to think we're also in
18:32
the web with a lowercase I right precisely
18:36
So I think if we were the true version
18:38
of what people call influencers, then it would be
18:40
icky But because we're not we're
18:42
just normal human beings who just happened to have done
18:44
something so long and so consistent and
18:46
I suppose Over time so
18:48
well that more and more people have
18:50
joined our tribe All right That's like I read Seth
18:52
Godin's book way back in the day and like everything
18:55
we've done has been a version of what he prescribed
18:57
In that book is around tribes.
18:59
It's such a good book. I love stuff go and he's
19:02
got such amazing writing You know the
19:04
dip is amazing tribes is amazing. I think it's
19:06
called tribes If I
19:08
can recall correctly, but it was around that like hey
19:10
There's there's a subset of people out there who find
19:13
what I think is interesting and over time what I
19:15
find interesting Evolves and I bring more and more things
19:17
into there like my hat meat church, right? I love
19:19
I love barbecue. I geeked out with a
19:21
sponsor the other day. We were done with our conversation We
19:24
hung out for another 30 minutes talked about
19:26
Weber Kettles We talked about how we tweaked
19:28
our stuff and like the conversation was totally
19:30
over right, but we were just there now
19:33
talking about barbecue and our favorite recipes
19:35
and Methods etc and that was
19:38
what it was and that could have totally been
19:40
a podcast But it was just it
19:42
was just a conversation which was cool. Yeah, there
19:44
you go There's some value there. I
19:46
do like the indie web thing. I guess I mean
19:49
Jamie is any web I know you said there's
19:51
like a proper now and there's also like a lowercase version of
19:53
it Is the proper now in a community
19:56
is it like a group of people that identify and
19:58
self-identify do they get together? I feel
20:00
like there was an IndieWeb conference of some sorts. Can
20:02
you tell us about that? Have you been part of
20:04
any of those events and stuff? So,
20:06
short answer, yes. End of focus. Yeah,
20:09
so short answer, yes. So,
20:11
indieweb.org is the main page
20:14
for the IndieWeb
20:16
community. So, it's more
20:18
along the lines of like capital
20:20
IndieWeb, but the idea is
20:22
it's meant to be the
20:25
community for people doing IndieWeb
20:27
related things. So, there
20:29
are IndieWeb camps, which are the
20:31
conferences. In fact, in
20:34
just over a month, or just under a month, there's
20:36
one in Brighton in the UK, but
20:38
they have them all over the globe.
20:40
I've been to a couple and
20:42
they're quite good because you basically come in
20:44
and you don't have to have built your own website.
20:47
You don't have to know anything about it. There's
20:49
like a few sessions at the beginning of
20:51
the day where it's like, well, actually, is
20:54
the IndieWeb, why is it actually important to
20:56
even just like own your URL and
20:59
not constantly be on twitter.com, slash whatever,
21:01
or now x.com, slash whatever, then
21:04
people go around and like share their websites. So,
21:06
you have like demos of like, this
21:08
is what my website looks like. And then
21:11
there are breakout sessions where you can talk about different
21:13
things. So, some people
21:15
are very interested in like owning their
21:17
location data. So, for instance, instead
21:19
of using something like Foursquare, they will use their
21:21
website. And then there's also
21:24
like interesting conversations around like privacy,
21:26
things like that, because often a lot of
21:29
people who are sharing their
21:31
location data in real time
21:34
are cishet white pan who
21:36
are not generally affected by,
21:40
oh, someone knows my location, they're now gonna
21:42
stalk me. So, there's like
21:44
a lot of interesting things around that and
21:46
trying to make that data less
21:49
invasive. Also interesting, because you may want to
21:51
be able to know, here are all the
21:53
places I checked into in the last week, because
21:55
you may want to like draw a map of places
21:57
you've been, stuff like that. That's cool stuff.
22:00
I feel like I've always been I guess
22:02
I identify with the indie web it seems
22:04
like that community everyone that I've
22:06
met so far are very inclusive they're not like
22:08
drawing lines like you are or are not indie
22:10
web you know like that's not the point the
22:12
point is like to promote the stuff that you're
22:15
talking about. And I'm a
22:17
longtime believer in own your own domain published
22:19
on your own website syndicate
22:22
etc. And like Adam said that's what we've
22:24
been doing for years here
22:26
on change log com we have a website
22:28
as well Jamie change log com. And
22:32
yeah it's cool stuff how
22:34
far do you go in with so you
22:37
got Fediverse implementation or integration do you do
22:39
web mentions those are always something that
22:41
any web people bring up. We
22:44
have like we've been adopting some of
22:46
the new podcasting standards podcasting
22:49
2.0 spec which integrates
22:51
stuff that's kind of like web mentions
22:53
like cross app comments which is very
22:55
much using. Is it just
22:57
mastodon or is it the Fediverse I don't know it's
22:59
using these kind of things to
23:02
try to create for podcasting some
23:04
of the stuff that I think activity pub is starting
23:06
to create a little bit for social
23:08
and for syndication. What
23:11
all does your website do so
23:14
my website is a static website to build
23:16
with Hugo and as I mentioned it takes
23:19
like 10 minutes to deploy because I just
23:21
want to be big. I've just
23:23
got so much content in it and I
23:25
also say for instance like if I'd like
23:27
to post I want to
23:29
mock that up with what
23:31
that post actually is so instead of just having
23:33
like I looked change log
23:36
com slash friends slash whatever. I
23:38
want it to be like listen
23:40
to change on friends the episode
23:42
with Jamie Tanner sort of thing
23:44
so as part of that I also need
23:46
to fetch like information around what that post
23:49
is so in the bank. I'll
23:52
be looking at does it support
23:54
micro formats which are an indie
23:56
web markup around HTML which lets
23:58
you. Your Html
24:00
elements basically say like misses the
24:03
post content. This. Is the title of
24:05
the post. Soul. Of things like that
24:07
there will also look for things like
24:09
open graph must say so and for
24:11
backed by so can do a few
24:13
things to try and gauge what others.
24:16
Thought. San so much takes
24:18
a long time to build as
24:20
as lol content and at it's
24:22
yeah got a law files and
24:25
I'm also using Galaxy eyes. Hosted.
24:27
Runners because I've gone through like
24:30
pieces of trying to host of
24:32
myself and speed up. On.
24:34
Them to spend like. This is still
24:36
costly. And time cause these
24:38
will seven playroom will have to
24:41
for options. And then yes,
24:43
are you about women's And so yes
24:45
that is for what mentions. So.
24:47
Once my sight deploys Finally the son
24:50
of weapons that some to one of
24:52
my services.run. On Thought then
24:54
looks at the site map for the
24:56
site. And says which things
24:58
have changed recently A man l
25:00
guess reach those pages, find any
25:02
links and somehow Mcmansions to the
25:05
world which than on most my
25:07
posts will have something called pretty
25:09
Fed. To. Bridge he is Service
25:11
one by Robert said the ended
25:13
up an who tries to bridge lotta
25:15
different social networks and so pretty
25:17
fat is that said about session
25:19
of off and so I don't
25:21
have to manage your understand any
25:23
the front of us stuff. I.
25:26
Can just send a what mention? I'm Brian
25:28
does have a hard work with a
25:30
community building bridges. Had to understand what
25:32
my website looks like. He isn't think
25:34
someone for months. And then translate
25:37
that to an activity pump. Is
25:39
com time on that will then get some town
25:41
to all the people who follow me elsewhere. So.
25:44
People directly from their masters on a
25:46
coma wherever they can follow my sight
25:48
natively get the same content as if
25:51
he were just. Browsing, The
25:53
web which pray to. Now
25:55
you're riding. He. has some sort of
25:57
alternate set up where you don't have to rebuild Are
26:00
you just rebuilding the current blog post? Because I know
26:02
that I used to use Jekyll quite a bit and
26:05
it was a minute or two to build the
26:07
whole site. And yet while
26:09
I was writing, there was times where I
26:11
would just rebuild the same file. Other times I
26:13
want to see how that affected some
26:16
sort of aggregation page or tags page or those
26:18
other stuff. And it got to
26:20
be very annoying at one to two minutes just
26:23
to work on it. And with
26:25
Hugo, which is a much faster generator, your
26:28
site must be very, very large. Does
26:30
it ever get annoying while you're writing
26:32
to have all this machinations or
26:34
is that just not happening? Oh
26:37
well. So I also started with Jekyll
26:39
and I got to a couple of minutes and was
26:41
like, this is really painful. We've moved to Hugo and
26:43
over time it's got worse and worse. So
26:46
my solution for writing is don't try
26:48
and preview the post. Just
26:51
do not run my site locally is a
26:53
pretty horrible solution to it. But
26:55
I'm fairly used to what markdown
26:58
rendering looks like for my site.
27:00
So if I'm doing significant changes
27:03
on the theming, I will then run it locally. What
27:06
I will do is I will delete most
27:08
of the site's content or get Hugo to
27:10
exclude building most of the site's content, which
27:13
that means that it is a lot quicker, a
27:15
lot easier to run. Yeah, it's definitely not a
27:17
good experience most of the time. Over
27:20
the last 18 months, I keep going to
27:23
rebuild my website from scratch and make it
27:25
a dynamic site. But that's a
27:27
massive thing to do. I don't really
27:29
have time. Or I
27:31
can't make time when I have so many other things I
27:33
want to be doing and need to be doing, but
27:36
that then makes it really tough. I
27:38
don't know how you can write without previewing it. I
27:41
just have to see what it looks like. I
27:43
kind of know what it's going to look like. I
27:45
still want to see what it looks like, like the
27:47
final version. There's satisfaction in that. Yeah, and I'm an
27:49
editor. I don't do like a draft and then edit
27:51
it. I literally edit as I
27:53
write, which is one of the reasons why this guy
27:55
is writing, because that's just a painful way of doing it,
27:58
but I have never been able to just write. and
28:00
not be like, well that's a bad sentence. I just can't do
28:02
that. People can do that. I don't know if you can. They'll
28:04
just write their thoughts. They'll spit all their thoughts out and they'll
28:06
be like draft one and then they'll go back and edit it.
28:10
I've never been able to do that because I
28:12
can't leave the crappy stuff in there. So
28:14
I'm editing the whole way. I'm also previewing like almost each
28:16
step of the way. So
28:18
even knowing what Markdown is going to look like
28:20
on the website, I'm still like, nah, I just
28:23
want to see it. It's kind of funny because
28:25
you just talked about the ephemeral idea, which crappy
28:27
writing, unedited stream of thought is
28:29
kind of like embracing ephemeral in a way.
28:31
Sort of, yeah. So I would encourage you
28:33
to revisit that idea considering you're new to
28:35
change a thought. I should
28:38
try it. I should try to write an
28:40
entire blog post without stopping and just
28:42
see if I can actually get that done. I think so too.
28:44
I mean because I had to, while you were talking there, the
28:46
audience couldn't see it, but I was like shaking my head yes.
28:48
I was like, nah, no, maybe not because
28:51
there's definitely times I throw things down and I
28:53
appreciate it going back lit. I'm like, wow, this
28:55
reads terribly, but I get the idea and
28:57
it's not published, but it gives
29:00
me new position because it's old
29:02
Adam's thoughts that I've forgotten that
29:04
I go back and revisit that lead me
29:06
to new thoughts. But I was bold enough
29:08
to at least write it down even though
29:10
it sucked in the moment, so
29:13
I can appreciate it about my past writing
29:15
when I've written. So I know a
29:17
number of people who write draft posts
29:19
on their websites, so it's
29:22
a publicly visible, but it has a little
29:24
banner that says, this is a draft post.
29:27
It may contain errors. I may not have fleshed
29:29
out ideas. And that's quite a
29:31
nice way because you are very much
29:33
saying like, I have written this thing. I
29:35
know it's not great, but I'm happy
29:38
hitting publish. That
29:40
takes some guts. It does. And
29:42
like one of the problems with writing
29:44
and writing publicly where lots of people
29:46
are going to read it is
29:49
that, oh, is someone going
29:51
to complain about this comment or
29:54
how many people am I going to get well actually
29:56
by. So I kind of have like
29:58
two. of writing. So my
30:01
first mode is like blockamentation style,
30:04
which is I'm just getting the
30:06
thoughts out of my brain so I can forget
30:08
about it. And in the future I can come back
30:10
to it and learn about it again. Or I
30:12
have posts that I definitely do
30:14
want, or I'm planning on other people
30:16
reading. And some of those posts
30:19
take like weeks of going back and
30:21
reediting and everything. And
30:23
yeah, I'm fortunate that I
30:26
don't mind too much about
30:28
hitting publish. And people's thoughts on
30:30
it, but it still does take a while for
30:32
some of the big moose. So
30:34
I can't wait to see what's going on. I
30:36
know I can be particularly spicy. People can have
30:38
big thoughts. Sure. And
31:04
I'm here with Lee Robinson, VP of Product. Lee,
31:07
I know you know the tagline
31:09
for Vercel, develop preview ship, which
31:11
has been perfect, but now there's
31:13
more after the ship process. You
31:15
have to worry about security, observability,
31:18
and other parts of just running
31:20
an application production. What's the story
31:22
there? What's beyond shipping for Vercel?
31:24
Yeah, you know, when I'm building my side
31:26
projects or when I'm building my personal site,
31:29
it often looks like develop preview ship, you
31:31
know, I try out some new features. I
31:33
tried a new framework. I'm just hacking around
31:35
with something on the weekends. Everything looks good.
31:37
Great. I ship it. I'm done. But as
31:39
we talk to more customers, as we've grown
31:41
as a company, as we've added new products,
31:43
there's a lot more to the product portfolio
31:46
of Vercel nowadays to help pass that experience.
31:48
So when you're building larger, more complex products,
31:50
and when you're working with larger teams, you want
31:53
to have more features, more functionality. So, tangibly,
31:55
what that means is features like our Vercel
31:57
firewall product to help you be safe. and
32:00
to have that layer of security features
32:02
like our logging and observability tool so
32:04
you can understand and observe your application
32:06
in production, understand if there's errors, understand
32:08
if things are running smoothly and get
32:10
alerted on those. And also then really
32:13
an expansion of our integration suite as
32:15
well too because you might already be
32:17
using a tool like a Datadog or
32:19
you might already be using a tool
32:21
at the end of this software development
32:23
lifecycle that you want to integrate with
32:25
to continue to scale and secure and
32:27
observe your application. And we try to
32:29
fit into those as well too.
32:31
So we kind of continue to
32:34
bolster and improve the last mile
32:36
of delivery. That sounds
32:38
amazing. So who's using the Vercel platform like
32:40
that? Can you share some names? Yeah,
32:43
I'm thrilled that we have some
32:45
amazing customers like Under Armour, Nintendo,
32:47
Washington Post, Zapier who use
32:50
Vercel's running cloud to not only help
32:52
scale their infrastructure, scale their business and
32:54
their product, but then also enable their
32:57
team of many developers to be able
32:59
to iterate on their products
33:01
really quickly and take their ideas and
33:03
build the next great thing. Very cool.
33:05
With zero configuration for over 35 frameworks,
33:07
Vercel's running cloud makes it easy for
33:10
any team to deploy their apps. Today
33:12
you can get started with a 14-day
33:14
free trial of Vercel
33:16
Pro or get a
33:19
customized enterprise demo from
33:21
their team. Visit vercel.com/change
33:23
log pod to get
33:26
started. That's vercel.com/change log
33:28
pod. Well,
33:32
we know you're not afraid of being in
33:35
the public because another thing that you've done
33:37
famously is publish your salary, right
33:39
over the years, which has gotten
33:41
a lot of attention, helps some people
33:44
tell us a bit of that story and then we'll talk about it.
33:47
Yeah, so I've been posting my
33:49
salary publicly since 2021 and
33:52
part of it was I was leaving a
33:54
job and I had just
33:57
gone through the job hunt and
33:59
I've been talking to you. to a number
34:01
of people privately about salary
34:04
and what the job market was looking like. It was a
34:06
particularly good time, not the most
34:08
recent ridiculous job market, but it was
34:10
a pretty good set of roles out
34:12
there. I was very happy
34:15
with the salary increase I was getting at the
34:17
time. I
34:19
very impulsively was like, you know what? I'm going
34:21
to post about this. I
34:24
did not check whether legally I
34:26
could do, which thankfully I was
34:28
fine. I was like, you
34:30
know what? I'm talking to enough people about this
34:32
and back to like number of keys left and
34:35
everything. I want to make sure that this
34:37
isn't a federal thing and this is something that
34:39
people can refer to and people
34:41
don't need to come up and ask me
34:43
about it because like in the UK, people
34:45
aren't happy talking about money. I
34:47
think it's a fairly global phenomenon where
34:50
people are just not going to
34:52
talk about it and therefore
34:54
companies get away with massively
34:58
underpaying people being very
35:00
unfair or even
35:02
just people not realizing that
35:05
they're actually doing pretty well for themselves.
35:08
Yes, I posted my salary. I
35:10
went back through my history. I
35:13
had fortunately only had a couple of jobs, so I
35:16
had all of this stuff available. I was
35:18
still in the job I was at that
35:20
had been five years. I had five years
35:22
of finances that I could just
35:24
download and upload. It's
35:27
been very good because it has helped a
35:29
lot of people. I've directly had
35:32
people I do and do not know message
35:34
me to say thanks for posting it. But
35:36
also in the last
35:39
few years, I've had at
35:41
least about 45 to 50,000 hits
35:43
on that page. I
35:47
say at least because a lot of people strip
35:49
things like analytics, which is good. So
35:51
it's probably a lot more than that. That's
35:54
cool. I can definitely see how it would benefit people and
35:56
I See absolutely the
35:58
benefit. Inside the same org?
36:01
Like people talking about their salaries? Inside the
36:03
same or two. Then you realize like wait
36:05
a second, You know, You're making that
36:07
much. They're paying me half. That would have the same
36:09
job only. Ill. Let. Me: Go talk
36:11
to the boss in a like. That's an
36:13
obvious when I still think it's take a lot
36:15
to put yourself out there. I know that
36:17
I personally have always been like. How
36:20
much money I make? it's nobody's business but
36:22
my own and. Is. How much money you
36:24
make is? It's not my business either. That's just. Been.
36:26
My default stance on life. And
36:28
so it's It's interesting to seize
36:30
you willing to do that and.
36:33
As anybody, follow your lead. You.
36:35
See, Since I first mine, I
36:37
know at least five or six
36:39
people who pursue it directly off
36:41
the back to be. First.
36:43
In mind So that's been pointless
36:45
because it's been a mix of
36:47
different people have rangers levels, there's
36:49
some be billie work on what
36:52
consulting sort of a slightly different
36:54
income stream to to salaried employees
36:56
on man even since posting it
36:58
on found people who had already
37:00
been posted in as. Difference.
37:02
In the yeah so she's been posting
37:04
house for some time. Developing are good
37:06
because it gives. A. Little bit
37:09
of difference whereas like. The.
37:11
People I know or or Uk based
37:13
and around the same sort of area.
37:16
So. Things like cost of living
37:18
everything's fairly standard The for lot
37:20
of other people it's interesting. The.
37:22
Other thing or found what sickly
37:24
interesting is from my previous job
37:26
deliveries so I joined with salary
37:29
of nice thousand pounds and this
37:31
was during the period where companies
37:33
were going wild for and today's
37:35
they were. Willing. To throw a
37:37
lot of money. And about a
37:39
year or so later, Those
37:41
around a promotions and people who
37:44
were promoted to the same level
37:46
as I was when I joined
37:48
were getting ten percent less. On.
37:50
Salary. and is things got
37:52
out that if you didn't have
37:54
this data very publicly of people
37:56
told mouth you miss out on
37:58
some of that A lot of
38:01
companies do have salary bonds and
38:03
that makes it at least a little bit better. But I've
38:06
also seen job adverts which are this
38:08
salary range is anywhere between 60 grand
38:11
and 250 grand. That's
38:13
not a helpful range. That's
38:16
wildly different. What do
38:18
you think, Adam? Would you post your
38:20
– let's say if you and I
38:23
went back to being W2 employees and just had
38:25
an annual salary, sharing that publicly. Would
38:28
that feel like something you'd be willing to
38:30
do? Would you hesitate? No way? What are
38:32
your thoughts? I guess it depends
38:34
on how frequently I move jobs because I know
38:37
that that's one thing when you move a lot
38:39
or even I suppose a lot is not always
38:41
an accurate way to describe it because a lot
38:43
could be every six months to a year or
38:45
so. And that could be used against you in
38:47
some ways like when you're a new way in. I
38:50
think one of the questions that's a meme
38:52
on TikTok which is, okay,
38:54
so what was your previous salary? And
38:56
the person responds, I
38:59
signed an NDA. It's
39:02
a funny schtick they do. It's a comedic thing in a way. I
39:05
feel like I'm of your camp, Jared,
39:07
where I don't – I feel like what
39:09
I make is my own business and what you make
39:11
is your own business. And I
39:13
think the reason why is obvious societal. I
39:15
think it begins with that
39:18
when people know certain details about
39:20
your person, they can begin to assume. They
39:22
begin to – not so much
39:24
cash judgment but they begin to sort
39:26
of determine who you are and what
39:29
you can do based upon what they think
39:31
they know about you. And then the biggest
39:33
thing it really is, your finances because that's
39:35
like one of the number
39:37
one resources that we all leverage to
39:40
progress in life in
39:42
some way, shape, or form. And so that's the
39:44
resource for which many people are judged by. How
39:46
much money do you make? Oh, well, that's why
39:48
you drive that car. Oh, that's why you have
39:50
that home or oh, that's why you value these
39:52
things or oh, that's why X, Y, or Z.
39:55
So I'm like maybe no, but
39:57
then I also see the benefit in it. I
40:00
think if it was, I just don't know if I could be in that
40:02
camp. I see the benefit, but I don't know
40:04
if I can participate in the camp that says, okay,
40:06
let me do this. I
40:08
wanna contrast this though, because you mentioned this in
40:10
the preparatory stuff, and this is something
40:13
I knew about Oxide, is they, Brian
40:15
Cantrill and Steve Tuck, when they founded the
40:17
company, they were like, okay, we wanna sort
40:19
of pay ourselves a goodness
40:21
salary, but we don't have to stress so much about
40:24
money. It's not so much that we're making, our
40:27
company may be valued at X, and on paper,
40:29
we're worth more, because over time, Oxide begins to
40:31
have a higher and higher valuation, but their salary
40:33
is 175 a year USD. And
40:38
when I met some folks recently
40:40
at All Things Open, I
40:42
was surprised to learn, this is when I first learned,
40:44
I learned face to face that everyone in the company
40:46
makes the same amount of money as the two founders.
40:49
And so they said, we wanna pay everyone the same.
40:52
And so I think in that context, Jared, that
40:54
might be interesting for us, you and
40:56
I, this change, I'll just say, this is how much money we make,
40:59
that would really be amazing if
41:02
we paid everyone else the same thing we pay ourselves,
41:04
but that's not exactly true. So
41:06
that might be embarrassing to be like, okay, there's
41:08
that divide there, so to speak. At
41:11
the same time, we're also the ones taking all the
41:13
risk and riding the wave of down
41:15
years, up years, et cetera. So I get that, but
41:17
we're not building an Oxide company. In
41:19
that light, if I were building an Oxide,
41:22
and I had similar possibilities and ambitions that
41:24
Brian and his team do, then
41:26
I would probably do that. That sounds like
41:28
a good thing for the
41:31
morale and a good thing to market. Like we're
41:33
talking about it in this moment because it's cool,
41:35
right? So there's a marketability to it. I
41:38
think as well, so Buffer in the
41:40
last few days, they've also announced that
41:42
they're doing similar. The
41:45
way Buffer does it is they have
41:47
everyone's salary as public, which
41:49
I think I'm less a fan of. So
41:51
I really like the way that Oxide
41:54
does it, where everyone gets the same, and
41:56
it's very, very stable
41:59
and nice and... equal or equal.
42:01
Yeah, equal, not equal. The
42:03
buffer you have the risk that you can
42:05
go and look someone else's salary up. Like,
42:09
I have absolutely made the choice that
42:11
I am happy with for the
42:13
rest of my life, because I'm not planning on undoing
42:15
this. For the rest of my life, people
42:17
will see what I'm paid. But to be
42:19
able to know that one of your friends works
42:21
buffer and just look up their name, and
42:24
look up how much they're getting. I'm not
42:26
as much of a fan. And I believe there
42:28
is a not out for employees. It's
42:30
the sort of thing that you know, I don't
42:33
recommend everyone try and post their
42:35
salary publicly, because very few
42:37
people can do it is a very privileged
42:39
thing to be able to do, and to
42:41
be happy doing that if
42:43
people are at least talking about
42:45
it privately with their friends and
42:47
family, that is really the important
42:49
thing because companies get away with
42:51
underpaying people and less
42:55
and things are generally because people aren't talking
42:57
about things. People are feeling that they're being
42:59
like pitted against each other. When really it's
43:02
the workers who do the work to make
43:04
the company what it is. It's
43:07
really important for people to get fair
43:10
pay, fair share. And
43:12
yeah, I absolutely think more
43:14
people should be talking about it.
43:16
delivery, we didn't really talk about things like
43:19
this. So aside from
43:21
me joining and sharing my salary and a few people
43:23
talking to me after the vote, we
43:26
didn't really talk about salaries until redundancy's
43:28
opened. And at that point, people
43:30
were much more happy to talk about things much
43:32
more happy to talk about, here's
43:34
what I got on my performance management reviews,
43:36
and things like it's a shame that we
43:38
had to go through such a horrible position
43:41
to feel that we could actually share
43:43
those things that make everyone
43:45
more equal and make it easier
43:47
for us to be better people.
43:50
So two questions. One,
43:53
did anybody at Deliveroo, especially
43:55
upper management, take offense
43:58
or were they upset that you that. So
44:02
legally, we're protected in the UK.
44:04
So fortunately, there's nothing they could
44:06
officially do. However, it's one
44:09
of those things that I
44:11
am probably not quite tarnished. But
44:14
I'm sure there will be companies who will look at
44:16
me and be like, No, not that guy. So
44:19
comments from Tina Fey, over the last couple of
44:21
days about not saying what you
44:23
think about films, because you never know when you're
44:25
gonna work with that director stuff like that. I
44:28
am sure that there are going to
44:30
be impacts of that, where
44:33
people know that I will be willing
44:35
to share my salary because I am
44:37
protected to do so. And it is,
44:39
in my opinion, the right thing to do for me. Also,
44:43
I think, yeah, I speak out
44:45
a lot. So in the meetings
44:47
and stuff, I was happy voicing
44:49
the opinions that I would
44:51
be happy lending my voice to us asking
44:54
a question that I knew other people would
44:56
be wanting to know. And
44:58
I know that not just a delivery, but
45:01
a capital one where I was when
45:03
I posted my salary, I know
45:05
that it did make things difficult for me, because I
45:08
had handed in my notice, I have
45:11
been posted that I have missed categorize
45:13
something as a counter offer when
45:15
it wasn't. And I know that definitely
45:17
did ruffles and feathers. And
45:19
that was on me misunderstanding the
45:22
wording of something. But those things
45:24
like that. And I'm sure there
45:26
will be companies who look at me, like,
45:28
I don't think we want someone
45:30
like that. And that's fine. But I'm happy
45:32
if that's what they want to do. Hopefully,
45:34
it doesn't mean that I can work literally
45:36
nowhere. Because then I then I will definitely
45:38
regret it. Well, that was gonna be my
45:41
second question was if there are any regrets
45:43
whatsoever. And it sounds like
45:45
there's some trepidation about potentially becoming
45:47
persona non grata, utterly, at which
45:49
point you're like, well, that wasn't
45:51
very smart. But beyond that, I
45:53
mean, you're going to continue it for the rest of
45:55
your life. So you can't regret it very much
45:57
if at all. Yeah, so it's still I
46:00
absolutely plan on carrying on with it.
46:02
I did recently add on-call pay as
46:05
well. Is that a separate page on my site? Just
46:08
because it's interesting, because different organizations have
46:10
different things. I
46:13
feel it's something that I may regret it
46:15
in 30 years, holding myself
46:17
to it for the rest of my career. He
46:22
has the blowback, I think. Being
46:24
too open, there's not so much anonymity, but
46:28
a certain privacy level of life
46:30
has its benefits. Especially,
46:32
you don't know what the future's gonna be, so if you're
46:34
open now, and I know you're sort of
46:36
pre-committed to this conversation, and now
46:39
here on the call, you're like, for life, I'm
46:41
gonna share my salary. You've laid
46:43
that down. And in
46:45
two years from now, something you do
46:47
could make you super famous. Maybe
46:50
that's cool, and maybe that's part of your brand at that
46:52
point, you'll just embrace it. And that's
46:54
cool, but maybe somebody, I'm
46:56
actually, Jared knows this, he makes
46:58
fun of me. I say absolute
47:00
things, because I feel so
47:03
strongly in the moment, and I make a long-term,
47:05
forever commitment in a way, with
47:07
something I'll say, and then later on, I'm
47:09
like, well, my idea around
47:12
that change, so now what I've said in the
47:14
past is no longer accurate, so I try to
47:16
not be so absolute about things. My
47:19
wife and I have a funny thing between us, we always say always
47:22
and every time, because
47:24
early in our relationship, we would say you always do this,
47:26
or every time, just these absolute things that
47:29
are just totally not true. It's just
47:31
our irrational, in the moment selves, saying things that
47:33
are not accurate and true about the other person.
47:36
And then that's now become our love language
47:38
in a way, because we make fun of
47:40
old selves, essentially, by saying you
47:42
always do this every time, we always say it back
47:44
to just jokingly, but this
47:46
absolute idea that you've done it
47:48
forever, I wonder how
47:51
it'll play out for you long-term, I suppose, because you're
47:53
pretty young, right? Like how old are you? Willing
47:55
to show your age? Wouldn't
47:58
that be funny if he's like, I don't want to share my age. So I mean,
48:00
you got a lot of years left in your life, right? Keys left
48:02
for you is pretty big. Probably
48:04
in the, you know, I would
48:06
say probably 400 million keys is my guess
48:09
for you. Oh, that's a lot. Well,
48:11
I just did my own keys left and I knew my
48:13
age and I just did some division there. So I assume
48:15
that his is probably double-ish mine or
48:18
at least one-third-ish mine. You got a
48:20
lot of typing to do. Yeah, it's a lot
48:22
of typing. But I guess on the topic of
48:24
like, regressing things. So I, since 2020, have
48:27
been posting weeknotes. So
48:32
every week on Sunday evening, I write
48:35
a blog post about what has happened in the week.
48:38
And I sorted it just
48:40
like before COVID. And it
48:42
was interesting because certain things
48:44
changed. I'm looking back at
48:46
that. But it wasn't until
48:48
about six months, a year
48:51
ago, when I started hearing about
48:53
like, some of my extended family reading
48:56
my weeknotes that I was like, oh, that's
48:58
a bit weird. And like your diary
49:00
in a way. And I was like, why
49:02
am I finding this weird? I have been
49:04
posting publicly on the internet things about
49:06
my life. So why do I
49:09
care that my extended family have been
49:11
reading it, let alone like someone,
49:13
several countries away who I've never met. And
49:16
that was like an interesting inflection point of
49:19
rethinking what sort of things I have
49:22
been sharing publicly. And so there's
49:24
definitely some things that I also have like private
49:26
journal, that I do post a bit
49:28
more stuff into that. And I
49:30
keep my weeknotes a little bit lighter on
49:32
some things because I
49:35
have a fair few readers, I don't
49:37
know exactly how many, but there's definitely
49:39
people who have like, messaged
49:41
me, like, ah, I've also watched that
49:43
film recently. And
49:46
I was like, that's weird. But yeah,
49:48
but yeah, also like, I'm posting it
49:50
publicly. Like, right. Yeah, it should not
49:52
be unexpected, because you are literally
49:54
publishing it for people to read. But
49:56
yet sometimes it still is unexpected,
49:59
especially when. when you know them. It's
50:01
actually, there's a weird disconnect
50:04
when it comes to strangers and when it comes to people that you
50:06
know. Like, I
50:08
get more self-conscious about
50:10
something that I make, especially
50:12
if it's like my work, when
50:15
somebody that I know or that I care
50:17
about is gonna consume it or judge it or
50:19
whatever, but I ship stuff out
50:21
to thousands of people every
50:23
day to do the exact same
50:25
thing, you know? And to me,
50:27
like them judging it is fine. It's like, okay,
50:29
like it or not, hopefully you like it, but
50:32
if my wife thinks something sucks, you
50:34
know, I was like, well, I didn't
50:36
make it for you specifically, but if you don't like
50:38
it, that hurts, you know? And so
50:41
it's just weird. It's like we disconnect strangers from
50:43
people that we know, yeah. Can we go
50:45
a personal layer deeper on this one? Whose
50:48
person, Jamie? For you, Jamie, yes, for you.
50:51
It's up to them. Our guest, our friend. It's up to them. Not our
50:53
friend, our friend. So you mentioned you have ADHD
50:56
and you mentioned that you have the inattentive type
50:59
and you mentioned that you have memory
51:01
challenges. I think it's mostly
51:03
in short-term, not your actual like long-term memory. And
51:07
what I know because I was
51:09
the host of a show called Brain Science for a Bit and
51:11
I've studied neuroscience to some degree, not
51:14
at all a clinical psychologist nor a doctor
51:16
of any sort, but just a layman who's
51:18
curious. And that, you
51:20
know, this is common for someone that
51:23
has your diagnosis. It's called memory breadcrumbs.
51:26
And so I think you do it one as a, I'm
51:29
assuming this, I wanna hear your response. I
51:31
think you do it one for therapeutic reasons. It's
51:33
helpful for you to have this
51:35
habit to do this. It's part of
51:37
your ritual and routine in
51:40
life that gives you peace to sort
51:42
of put this out there. Then as a
51:44
technologist, you value the exhaust, the output of
51:46
that for future. Because you can,
51:48
again, I can vaguely remember I wrote about this X,
51:50
Y, or Z and then pull it up and don't
51:52
have to leverage your memory
51:55
anymore. You can sort of leverage your handicap
51:57
in a way. And
51:59
then also the... personal journals probably super therapeutic
52:01
for you. There's certain things you probably learned
52:03
over time you can't involve in these weekly
52:05
notes. But that this
52:07
is essentially like your response,
52:10
kind of like a coping mechanism or
52:12
a fight or flight kind of response in a
52:14
way to your diagnosis that this
52:16
is actually memory breadcrumbs for you. This is a
52:19
way for you and how it's helpful for someone
52:21
like you to put down
52:23
what they learn and remember. And
52:25
I'm just curious how that how you feel about
52:27
that. Like is that pretty accurate? Have I pegged that?
52:30
Yeah, so I've never heard the term memory breadcrumbs. So
52:32
thank you for that. I'm going to go and search
52:35
around on that. And yeah, it absolutely sounds
52:37
like what I'm doing. So
52:39
on the technical side is yeah, leaving
52:41
those breadcrumbs for myself to try
52:43
and remember what was doing
52:45
for the week notes. It was
52:48
more just I've seen some friends
52:50
doing this on I see what it's like. But
52:52
it has definitely become this therapeutic
52:54
thing where every Sunday night
52:56
is week notes night. And it gives me
52:58
a chance to reflect on the week. It's
53:02
like a lot of the time it's it's interesting looking
53:04
back and I don't often read
53:06
my old ones which is also kind of
53:08
ephemeral, but I'll read them while I'm writing
53:10
them and then I won't look at them
53:13
for years. But something interesting
53:15
is like looking back and seeing
53:18
okay, so the things that I was most
53:20
bothered about that week was like a technical
53:22
problem that who cares in the big
53:24
scheme of things or I'll do
53:26
things like I'll note the different media
53:28
I've watched in the week. And
53:31
that's interesting because I can see every so
53:33
often, oh, I've actually only
53:35
watched three things this week. So
53:37
be the not really watching a lot of TV
53:40
or I've binged something a lot. And
53:43
that's also interesting to like look back on that.
53:45
Yes, to go back to your original
53:47
question, because he is sidetracking me. Yes,
53:50
you did diagnose that fairly well. Do you
53:52
think it gives you a position of gratitude
53:55
better because I feel like the reason why
53:57
journaling is helpful is not just in the
53:59
therapeutic process of espousing
54:01
your own thoughts and putting them
54:03
down on paper and or digital paper
54:06
whatever the medium really is. But
54:08
it also maybe helps you reflect like you said reflect
54:11
on what happened this week and you
54:13
may not go back and read the other ones but
54:16
you're probably taking track each week mentally okay this week's
54:18
different than last week where you see
54:20
progress in the moment as you write it down.
54:23
Does it give you a chance to sort of have
54:25
a deeper gratitude for
54:28
what's going on or does it give you
54:30
content like what does it help you feel
54:32
better or less about? So
54:34
I guess when I said journaling it's
54:37
not like proper journaling so it's more
54:39
just like talk about things that
54:41
have happened in the week and
54:44
there's not always like a lot of reflection
54:46
sometimes it is just these are the things
54:48
that happened and that were of
54:50
note for me to have remembered
54:53
during the week to write down but
54:55
yeah sometimes like there are definitely friends
54:57
and there's things like oh
54:59
I've been very busy
55:02
with work this week I've been trying to deal
55:04
with a complex problem and therefore
55:06
like my personal life is maybe a little
55:08
bit more chaotic. I think it was
55:10
last week I was at the State
55:12
of OpenCon conference in London and
55:15
in the lead up to the conference I spent a lot
55:17
of time mentally blocked to write
55:19
my slides for the talk that I
55:22
was presenting there and so
55:24
a lot of my free time was
55:26
kind of like oh I should really
55:28
be doing my slides but I don't really have the
55:30
brain power to do it so I'm just going to
55:32
do nothing and I'm not going to do like a
55:34
load of other life admin I'm not going to
55:36
do a load of other personal projects I'm just going to not
55:39
do anything because if I had the brain power
55:41
I would do that and yeah
55:43
so it does give some reflection on things like that
55:45
as well. Are you a big movie fan?
55:47
You watch a lot of movies? Yes. Are
55:49
you a Christopher Nolan fan by any chance? The
55:52
director Christopher Nolan? Yeah. There's
55:54
a movie he had done called Memento. Does this ring a bell
55:56
to you? I haven't seen it.
55:58
I've heard of it. Well, I won't
56:00
ruin it for you But
56:05
it definitely involves memory and it
56:07
kind of involves time because it really
56:10
It goes about the plot of the
56:12
movie differently than any other movie you've ever seen before
56:15
It's very groundbreaking in the way it tells the story
56:18
But loosely this person has memory challenges
56:20
and loosely this person uses
56:23
their own notes To
56:25
guide them through their next steps and
56:27
sometimes these notes aren't really accurate Because
56:30
they have memory challenges and so they they they
56:32
remember them one way they when they write them
56:34
down for the future selves These memory breadcrumbs so
56:36
to speak but then in the
56:38
moment because they have short-term memory loss they
56:40
saying too much man You're saying I'm giving
56:42
you some stuff like it's really challenging You're
56:45
like dude. This is one of the best movies at
56:47
all. Did I give it did I give it away
56:49
too much? I may not remember. It's pretty true. That's
56:52
pretty that's very vague. I think I was pretty vague
56:54
You should just go watch it It's different than ADHD
56:56
for sure But it's it's still the idea of memory
56:58
breadcrumbs and like leveraging the things you write down to
57:00
remember the things in the future We're gonna have
57:02
to blow the whole the spoiler horn before that
57:04
whole session though. Remember that you think so God
57:07
Oh, whatever just in case didn't try to know
57:09
I know you were trying not to but you
57:11
just kept going I'm like, uh, oh, he's starting
57:13
to get more and more on the floor pull
57:15
it back. Yeah reel it in It's
57:17
a good movie I would definitely recommend watching
57:20
it and I think given this conversation You
57:22
will appreciate the movie even more and we'll know
57:24
about it when we read your weak notes. That's
57:26
right Like Adam was
57:28
wrong. This movie is terrible. I hope he doesn't
57:31
read this. That's right. That's right I
57:52
Know friends this episode is brought
57:54
to you by our friends at
57:56
Sanadia. Sanadia is helping teams take
57:58
NASA the next level via a
58:01
global multi-cloud, multi-geo, and extensible
58:03
service. Fully managed by Sanidia,
58:05
they take care of all
58:07
the infrastructure, management, monitoring, and
58:10
maintenance for you. So you
58:12
can focus on building exceptional
58:14
distributed applications. And I'm here
58:16
with VP of product and engineering, Byron Ruth. So
58:19
Byron in the NATS versus
58:21
Kafka conversation, I hear a
58:24
couple of different things, one I hear out there,
58:26
I hate Kafka with a passion. That's quoted by
58:28
the way on Hacker News. I
58:30
hear Kafka is dead, long live Kafka.
58:33
And I hear Kafka is the default,
58:35
but I hate it. So what's the
58:37
deal with NATS versus Kafka? Yeah.
58:40
So Kafka is an interesting one. I've
58:42
personally followed Kafka for quite some time
58:44
ever since the LinkedIn days. And I
58:46
think what they've done in terms
58:48
of transitioning the landscape to event
58:51
streaming has been wonderful. I think
58:53
they definitely were the sort of
58:55
first market for persistent data streaming.
58:57
However, over time, as people have
58:59
adopted it, they were the first
59:01
to market, they provided a solution,
59:04
but you don't know what you
59:06
don't know in terms of you
59:08
need this solution, you need this
59:10
capability, but inevitably there's also all
59:12
this operational pain and overhead that
59:14
people have come to associate with
59:16
Kafka deployments based on our experience
59:19
and what users and customers have
59:21
come to us with, they would
59:23
say, we are spending a
59:25
ton of money on spend on a
59:28
team to maintain our Kafka clusters or
59:31
managed services or something like that.
59:33
The paradigm of how they model
59:35
topics and how you partition
59:38
topics and how you scale them is
59:40
not really in line with what they fundamentally
59:42
want to do. And that's
59:45
where NATS can provide, as we refer
59:47
to it, subject based addressing, which
59:49
has a much more granular way of
59:51
addressing messages, sending messages, subscribing
59:54
to messages and things like that, which
59:56
is very different from what Kafka does
59:58
and the second that we... introduced persistence
1:00:00
with our Jetstream subsystem as we
1:00:02
refer to it a handful of
1:00:04
years ago. We literally
1:00:06
had a flood of people saying,
1:00:09
can I replace my Kafka deployments
1:00:11
with this NATS Jetstream alternative? And
1:00:13
we've been getting constant inbounds, constant
1:00:15
customers asking, hey, can you enlighten
1:00:18
us with what NATS can do?
1:00:20
And oh, by the way, here's
1:00:22
all these other dependencies like Redis
1:00:24
and other things and some of
1:00:27
our services based things that
1:00:29
we could potentially migrate and evolve over
1:00:31
time by adopting NATS as a technology,
1:00:34
as a core technology to people's systems
1:00:36
and platforms. So this has
1:00:38
been largely organic. We never from day
1:00:40
one, you know, with our persistence layer
1:00:42
Jetstream, the intention was never to say
1:00:45
we're going to go after Kafka. But
1:00:47
because of how we layered the persistence
1:00:49
on top of this really nice pub
1:00:51
sub coronats foundation, and then we promoted
1:00:53
it and we say, hey, now we
1:00:55
have the same, you know, same semantics,
1:00:57
same paradigm with these new primitives that
1:01:00
introduce persistence in terms of
1:01:02
streams and consumers. A flood gate
1:01:04
just opened and everyone was frankly
1:01:06
coming to us and wanting to
1:01:09
simplify their architecture, reduce costs, operational
1:01:11
costs, get all of these other
1:01:13
advantages that NATS has to offer that Kafka
1:01:15
does not whatsoever, or any of the other
1:01:17
similar offerings out there. And you get all
1:01:19
these other advantages that NATS has to offer.
1:01:22
So there's someone out there listening to this
1:01:24
right now. They're the Kafka cluster admin, the
1:01:26
person in charge of this cluster going down
1:01:28
or not. They manage the team, they
1:01:31
feel the pain, all the things,
1:01:33
give a prescription, what should they do? What
1:01:36
we always recommend is that you can go
1:01:38
to the NATS website, download the server, look
1:01:40
at the client and model a stream. There
1:01:42
are some guides on doing that. We also
1:01:45
have Cenadia provided a basically
1:01:47
a packet of resources to
1:01:49
inform people because we get again, so many
1:01:51
inbound requests about how do you compare NATS
1:01:53
and Kafka? And we're like, let's actually just
1:01:56
put a thing together that can inform people
1:01:58
how to compare and contrast. them.
1:02:00
So we have a link on the
1:02:02
website that we can share and you
1:02:04
can basically go get those set of
1:02:07
resources. This includes a very like lengthy
1:02:09
white paper from an outside consultant that
1:02:11
did performance benchmarks and stuff like that
1:02:14
and discuss basically the different
1:02:16
trade-offs that are made and
1:02:18
they also do a total
1:02:20
cost of ownership assessment between
1:02:22
people who are organizations running
1:02:24
Kafka versus running NATs for
1:02:26
comparable workloads. Well there you
1:02:28
go, you have a prescription. Check for
1:02:31
a link in the show
1:02:33
notes to those resources. Yes,
1:02:35
FaceTech is not cutting it.
1:02:37
NATs powered by the global
1:02:39
multi-cloud multi-geo and extensible service
1:02:41
that is fully managed by
1:02:44
Cineadia. It's the way of
1:02:46
the future. Learn more at
1:02:48
cineadia.com/change log. That's s-y-n-a-d-i-a.com/change log.
1:02:54
Well we would be remiss to let you
1:02:56
go and not talk about DMD because this
1:02:59
is a tech podcast after all. Is this
1:03:01
a tech podcast? I don't know. I mean
1:03:03
in a category in RSS feed it says tech
1:03:06
or something. I can't remember what I actually put. Ostensibly
1:03:10
it is. DMD, dependencies,
1:03:13
one of my favorite things. Managing
1:03:15
dependencies, not one of my favorite things. DMD
1:03:19
management data, a set of tooling
1:03:21
for knowing all about your dependencies,
1:03:23
like querying,
1:03:25
researching. Tell us
1:03:28
more. Yeah, so dependency management
1:03:30
data DMD for short and
1:03:32
very poorly named. I
1:03:35
now have like a t-shirt with
1:03:37
it on so I'm kind of like
1:03:39
a t-shirt with it on. I'm kind
1:03:41
of like a t-shirt with it on.
1:03:43
I like it because it's kind of
1:03:45
a reflection of my inability to name
1:03:47
things well and think of
1:03:50
catchy stuff. But yeah, so dependency
1:03:52
management data is a project I was building while
1:03:54
I was at Deliveroo and it
1:03:56
was coming up to October 1st and
1:03:58
I was a big pro. and Silam, a
1:04:01
big proponent of open source, of give
1:04:03
them back. And aside
1:04:05
from issues that maintainers have
1:04:07
during HEXO first, I
1:04:09
was still very pro. Let's get
1:04:12
some more delivery engineers contributing because
1:04:14
everyone should be trying to at least contribute to
1:04:16
that. And one of the questions I
1:04:19
was asked a number of times was,
1:04:21
well, which project should I contribute to?
1:04:23
Because it's hard to know where
1:04:26
to start with open source. Instead
1:04:29
of sending, go for a
1:04:31
good first issues list, I was like, you
1:04:34
know what? Let's look at the dependencies that
1:04:36
Deliveroo has. And let's say, of
1:04:38
all of these, these are the ones we use the
1:04:40
most. So maybe have a look at these ones. And
1:04:42
Deliveroo's a very good data-driven organization.
1:04:45
So I was like, okay, I'm gonna get
1:04:47
some data to actually show that. So
1:04:50
I ended up building
1:04:52
some very horrible hacky
1:04:54
scripts which would take
1:04:56
output from GitHub's dependable
1:04:58
graph and it would
1:05:00
then parse those and spit
1:05:02
it out into an SQLite database. Over
1:05:05
the next few months, I realized that actually, this
1:05:07
is quite a useful thing. Like being able to
1:05:09
have a database of all your
1:05:11
dependencies and being able to query them, being
1:05:14
able to add like additional insight is
1:05:16
actually really powerful. So
1:05:18
I started playing around
1:05:20
with endoflife.date, which
1:05:22
is an API for looking up common end
1:05:25
of life or deprecation dates. So
1:05:27
for instance, how long is Node.js supported?
1:05:30
When is, I don't
1:05:32
think they have any like physical hardware. There's
1:05:34
lots of like programming languages. They are
1:05:36
very open to supporting new things. And
1:05:39
so I was starting to add this data and
1:05:41
realized that this actually shouldn't be
1:05:44
a thing that just sits within Deliveroo. It's
1:05:47
something that I can absolutely see other
1:05:49
people get benefit from. I
1:05:51
was spending a lot of evenings and weekends on
1:05:54
the idea anyway. So I started
1:05:56
an open source project for this called DMD.
1:06:00
it has just passed its anniversary
1:06:02
of its first birthday. So I've
1:06:04
been spending just over a year, I think one
1:06:07
in three days of the
1:06:09
last year I have been
1:06:11
committing to the project. So I've
1:06:14
spent a lot of time working on it. And
1:06:16
as well as being able to just
1:06:18
have a list of those dependencies, being
1:06:21
able to say things like, where are
1:06:24
we using Terraform in the organization?
1:06:27
Or which go HTTP servers
1:06:29
are we using? And how
1:06:31
many the different teams using? You
1:06:34
can also do things like looking up
1:06:36
OpenSSF school card data to
1:06:38
say things like, of the dependencies, my
1:06:41
most critical application in our business
1:06:43
support relies on how many
1:06:45
of those do zero
1:06:47
code review? And
1:06:49
how many of them just yeeting things into prod? And
1:06:52
it turns out it's quite a lot more than you would
1:06:54
think. So having a lot of
1:06:56
that data is really useful. And
1:06:58
both my time at Deliveroo and
1:07:01
now at Elastic, we've been
1:07:03
spending a lot of time using
1:07:05
that data to make some really
1:07:07
interesting decisions and interesting understandings of
1:07:09
our data. Yeah,
1:07:11
you can find some very interesting things.
1:07:13
And I've said interesting about a dozen times
1:07:16
there. It's very interesting.
1:07:18
So interesting. Yeeting things
1:07:21
into prod, I like that. Well,
1:07:23
on the name, so DMD, DMD
1:07:25
is cooler than what it
1:07:27
stands for, dependency management data. If
1:07:30
somehow you could fit a recursive acronym in
1:07:32
there, then that would be really cool. If
1:07:34
you could still maintain the DMD. But free
1:07:36
piece of marketing advice, there is
1:07:38
a excellent ACDC
1:07:41
song called TNT. Because
1:07:46
I'm TNT. And
1:07:48
I know I'm nice. TNT.
1:07:52
And I will not fight. I'm
1:07:56
a power load. Who
1:08:00
wants me to be here? What
1:08:02
do you think? Repurpose. Because they have
1:08:04
this big... Because DMD. TNT.
1:08:08
You could attach yourself to that. Don't
1:08:10
change the name. Just give
1:08:12
it a little bit of excitement. A little bit
1:08:14
of explosion. A little bit of hard rock to
1:08:17
round out the soft edges. What do you think, Adam?
1:08:19
I think it's dynamite. You know this song? Yeah.
1:08:22
Thank you. Thank you. Sorry
1:08:25
about that. I had to bring it in.
1:08:28
I loved it. I loved it. Does it
1:08:30
matter what stack you're using or anything? Is
1:08:32
it agnostic? Yeah, so that's a really good
1:08:34
question. So first of all, the
1:08:36
changelog's data is all in there. Oh, no. They
1:08:38
can actually go and have a look. Oh,
1:08:41
can I? Yeah. So... OK.
1:08:43
I like that. So would
1:08:45
you like to see which
1:08:47
of your dependencies are end-of-life
1:08:49
deprecated? Have security issues? Absolutely.
1:08:51
Please, yeah. It would be very interesting. We're
1:08:53
shifting left here, don't you know? Yeah. I
1:08:55
put the brakes on hard. I don't know
1:08:57
what shift the brake card left. So
1:09:01
for instance, in this list, like you can see,
1:09:04
you've got a couple of non-standard
1:09:06
licenses, which just means that Deps.dev,
1:09:09
the API, or one of the
1:09:11
APIs that dependency management data users,
1:09:14
couldn't understand the licensing. We've
1:09:16
got a few packages that have vulnerabilities. And
1:09:19
then you're also using Go 120, which
1:09:22
has been end-of-life for seven days. How
1:09:24
are you not already on the latest version? I'm with us.
1:09:29
But yeah, so one of the good choices I
1:09:32
made around dependency management data was
1:09:35
trying to do as little work as possible. So
1:09:38
instead of me trying to understand
1:09:40
the different ecosystems and all
1:09:42
the data that they use and how the different
1:09:44
package managers work, I'm outsourcing
1:09:46
that to someone else. So the
1:09:48
original version uses DependaBot. But
1:09:51
the data that I was getting from that didn't
1:09:53
include a lot of things that we used at Todavaru.
1:09:56
And we're just starting to roll out
1:09:58
renovate bot for dependency upgrades. And
1:10:00
as part of that, I noticed that Renovate had
1:10:03
a just like a load of support for
1:10:05
a load of different things, but it
1:10:07
also spat out a big blob of
1:10:09
JSON, which contained all the packages it
1:10:11
knows of. So what I've done
1:10:13
is I've added a very small wrapper on top of
1:10:15
Renovate to basically dump out that
1:10:17
dependency tree. So for
1:10:20
instance, you can... So
1:10:23
another link I've just dropped in there, Google Doc.
1:10:26
So yeah, one of the reasons for using Renovate
1:10:28
meant that I could support dozens and
1:10:31
dozens of different things out the box
1:10:33
for very little work. So
1:10:35
for instance, because Renovate
1:10:37
understands Mix, the Alexa dependency
1:10:39
management tool, I
1:10:41
actually have all the details of all the
1:10:43
versions of tools that ChangeLog
1:10:46
is currently using. And from
1:10:48
that data, you can then look up things
1:10:50
like which version of the OAuth libraries we're
1:10:52
using. Do any of our dependencies
1:10:54
have some pretty poor repository
1:10:57
health? Things like that. So out
1:10:59
of the box, it supports a lot of things depending
1:11:01
on which tools you use to extract
1:11:03
it, but then you can also
1:11:05
plug it in with things like software,
1:11:07
bit of materials, S-bombs. So
1:11:09
if you're using a platform like Snyk, you
1:11:12
can export the data from Snyk. You can
1:11:14
still leave Snyk managing that, and
1:11:16
you can just import that data into DMD. So
1:11:19
you still get the visibility. You just
1:11:21
don't have to do the data collection part. Super
1:11:24
cool. You're using Simon Wilson's dataset
1:11:26
underneath the hood, huh? I've seen that on
1:11:28
the homepage. It is very useful. And
1:11:31
yeah, user interfaces, user experience,
1:11:33
just like design in general, is not
1:11:35
my passion, which a lot of
1:11:37
people have seen over the years. So
1:11:39
having the dataset project as the database
1:11:41
browser was amazing because it didn't
1:11:44
have to do anything. It's really useful.
1:11:46
And as you've seen, you
1:11:48
can share URLs with people, and it pre-fills
1:11:50
the query. So this is really useful internally
1:11:52
where you want to send it to someone.
1:11:54
You want to say, hey, your
1:11:56
service uses a load of insecure dependencies. You
1:11:58
can just send someone. on a link and they
1:12:01
can get that instead of sending them like
1:12:03
300 lines of SQL. And
1:12:05
it's like, go to this URL, then paste it in. Uh,
1:12:08
yeah. Yeah. Super cool, man. Have
1:12:10
you gotten much pickup? Have you gotten people using
1:12:12
it? So I've definitely got three
1:12:14
companies who are using it right now,
1:12:17
two of which I've worked out. So it's not
1:12:19
as fun. It's grassroots marketing, you
1:12:21
know, just one, one job at a
1:12:23
time. Yeah. There's definitely at least one
1:12:25
company who I have not been there
1:12:28
to set it all up. So that's
1:12:30
been cool. Um, I have
1:12:32
talked with a number of other people, uh, here and
1:12:34
there. So
1:12:36
I know there are people trying it out who
1:12:39
maybe haven't told me they're using it. Yeah.
1:12:41
Let's just get more people using that. Um,
1:12:43
understand things and interesting things. Cause
1:12:46
in my opinion, this is better than any
1:12:48
of the other offerings that other platforms have.
1:12:50
But also this is coming from a very
1:12:52
specific view and I don't have a lot
1:12:54
of VC funding, so I can't do as
1:12:57
much as they do. Right. Do you have any VC funding? I
1:12:59
do not. Okay. Just
1:13:02
to be clear, right now a lot of means none. Well,
1:13:04
now you have a grassroots marketing idea, you know, that's free
1:13:07
by the way. I'm not going to charge you for that.
1:13:09
The TNT ACDC tie-in is really
1:13:11
good. I didn't take this over the top. That's
1:13:15
brute force. Actually not grassroots.
1:13:18
We'll see. There's opportunity here, right?
1:13:20
I mean, like, I think so. Could
1:13:22
you turn this into a, a product?
1:13:25
So it's absolutely something I'm thinking of.
1:13:28
For instance, I was listening to change
1:13:31
log 577 with nausea and a lot
1:13:33
of things I
1:13:35
was like, Oh, this is an interesting
1:13:37
idea. And, Oh, and so
1:13:39
like recently I have been listening
1:13:42
to a number of podcasts around building
1:13:44
open source businesses. So there's the business
1:13:46
folk and sauce with Emily
1:13:48
O'Mear, which is really interesting around like different
1:13:51
companies who've tried to do it. And
1:13:54
it's something I'm actively thinking about.
1:13:56
I haven't yet done anything
1:13:58
on it aside from. a bit of hacking around
1:14:01
with things because like some of it
1:14:03
I think Organizations
1:14:05
probably want to be in control So
1:14:08
saying hey attach this github app for
1:14:10
this thing that some random developer you've
1:14:12
never met and just give them access
1:14:15
to your Source code for everything a
1:14:17
lot of organizations probably aren't gonna be
1:14:19
up for that So I'm
1:14:22
trying to look at making it easy for
1:14:24
self hosting Yeah for organizations to bring it
1:14:26
in and it kind of just to
1:14:28
work Yeah, so
1:14:30
there's a few options there that I'm looking
1:14:32
at around how to make that easier. Yeah
1:14:34
Yeah, well to be clear at
1:14:37
one point source graph was just some random
1:14:39
developers and at what point
1:14:41
socket dev was just some random well-known
1:14:45
developer frost book the DJ and They
1:14:48
both have github apps. That is part of
1:14:50
the install process. I know that that was
1:14:52
one thing I talked to Quinn
1:14:55
slack about on founders talk was how did you
1:14:57
get these enterprises to trust you with their data?
1:14:59
He's like that was the hard part. That was
1:15:01
like the biggest hurdle, but
1:15:03
you know once they did It
1:15:07
wasn't that challenging to prove the value So
1:15:09
I think if you can prove that
1:15:11
value and prove the trust then I think it
1:15:14
gets a little easier That's the that's the true
1:15:16
hurdle the dip really for that one Or
1:15:19
a different way around it really, you know,
1:15:21
well good luck to you on that I mean, I want
1:15:23
to hear more for sure a bit. Good luck to you on that. I mean, that's
1:15:25
a It's a fun battle,
1:15:27
you know, it's a worthwhile battle and it
1:15:29
is one of those things that like the
1:15:32
the sorts of use cases It's unlocks have
1:15:35
been like honestly game-changing for at
1:15:37
least me I delivery one
1:15:39
of our platform teams Kind
1:15:41
of pivoted what they were doing because they
1:15:43
could suddenly do so much more with this
1:15:45
data But they were able to
1:15:47
understand say which terraform modules
1:15:49
the teams using and at what versions
1:15:52
they were able to easily see but
1:15:54
they were able to run a Fairly
1:15:56
straightforward SQL query because not
1:15:58
everyone's gonna find SQL straightforward, that
1:16:01
they were able to write that query with the data
1:16:03
set UI, and just see
1:16:05
across the board which teams are doing
1:16:07
different things. Yeah. And then I'd also
1:16:10
ingested things like ownership information. So it
1:16:12
would actually say like, Oh, and it's this team that
1:16:15
owns that. And then teams could break
1:16:17
it down by different things. I've
1:16:20
over like the last month or so I've written
1:16:22
a number of case studies on the website around
1:16:24
like different use cases that has
1:16:26
been used for things like we're
1:16:28
in the middle of an incident delivery.
1:16:30
There wasn't like, wasn't any problems, but
1:16:33
it was a proactive
1:16:35
incident raised. And
1:16:37
we've noticed that there was one
1:16:39
of our duck sidecars for Kafka
1:16:41
consumption was susceptible
1:16:44
to a race condition. And
1:16:47
one option would be to try and
1:16:49
just like grep through lots of different
1:16:51
code bases and find what was affected.
1:16:53
Or it was like a seven line
1:16:56
SQL statement that we could just
1:16:58
do because we had all the data there. And
1:17:00
things like that. Until you have that data, you
1:17:02
probably can't see some of that benefit. But once
1:17:04
you have it, and you can start interrogating those
1:17:07
sorts of things. And as you say,
1:17:09
like the source graph, it's being able to
1:17:11
actually think up those sorts
1:17:13
of questions when you didn't have anything there before
1:17:16
is difficult. But then you get access to it and you're like,
1:17:18
Oh, actually, yeah, I can now do a load
1:17:21
of things that would never available. That's
1:17:23
why I think those examples you have and
1:17:25
just like the sample questions are
1:17:27
a useful thing. Because
1:17:29
you help the person dream,
1:17:31
you know, realizing, because otherwise, like,
1:17:34
I'm kind of like, Yeah, I got data
1:17:36
on lots of stuff, you know, it's like,
1:17:39
I kind of know my dependencies. And there's tools
1:17:41
built into hex, for instance,
1:17:44
that will say what's outdated, you know, it's like,
1:17:46
that's probably good enough. That's like, yeah, but can
1:17:48
you ask this question? And you answer this, can
1:17:50
you know this about your start
1:17:52
to think, Oh, okay, you open you
1:17:54
broaden the imagination of
1:17:56
the potential user. And so I
1:17:59
think the more of that you can do as well
1:18:01
as the TNT tie-in, which is gonna
1:18:03
be just groundbreaking. You're
1:18:07
already gonna. Give it up, man. It's
1:18:09
becoming not Dyna. I
1:18:12
don't know, that might have been my best. I'm guessing
1:18:14
you got math. I'm just razzing you. I
1:18:16
think there's like a, it's
1:18:18
like a version of a subset of
1:18:21
what Sourcegraph offers, like is answering questions
1:18:23
about your code base. Yeah. Right,
1:18:25
and you've gotta have certain, you know, queryable knowledge, which
1:18:28
isn't always on everybody's purview.
1:18:31
And there's a version of this, it's like in the
1:18:33
camp of Socket, which is kind of why I mentioned
1:18:36
both of them, because they both required to have the
1:18:39
user's trust of their data and their code
1:18:41
base, which is obviously part
1:18:43
of the battle. But I think you're a version,
1:18:46
you're like a love child in a way, but like a smaller love
1:18:48
child. And I don't even know the full tool, but it's a version
1:18:50
of maybe a hate child.
1:18:52
Who knows, Jared? I don't know, man. I'm
1:18:54
going with love. Okay. It's a version of
1:18:56
what both of those tools do, but
1:18:59
not to their full circumference, because Sourcegraph has obviously
1:19:01
done some really cool stuff with Insights and their dashboard
1:19:03
they do with that, which I think is pretty cool
1:19:06
too, which is kind of what
1:19:08
you've done here in a way, but this is more of
1:19:10
a hacker's version of it, which
1:19:12
is command line and whatnot. But I think
1:19:14
there's something here. I would encourage you to pursue
1:19:16
what a good next step might be to productize
1:19:19
this. Maybe there's people who are like, you know what, Sourcegraph is
1:19:22
just way too big. I need something that's simpler. And
1:19:24
maybe there's people who's like, you know what, I don't need the security aspects
1:19:26
of Socket, or like some of the things it does, because
1:19:30
Socket is all about dependencies, but from a
1:19:32
security lens. And even
1:19:34
from a, is this
1:19:36
repository active? And answering certain questions
1:19:39
about the repository, the dependency. Has
1:19:42
the committers or core
1:19:44
team members changed recently? Things
1:19:47
like that. That's kind of sort of what
1:19:49
Socket does in deeper levels, and more on
1:19:51
learn and awareness. But that installs as
1:19:53
a GitHub app. That's like one of the most popular ways.
1:19:55
I know that because they're a sponsor, and that's sponsoring this
1:19:57
one here, but they've sponsored this before. That's
1:20:00
how like one of the main ways people interface
1:20:02
with socket is via the get a bat. Yeah,
1:20:05
I guess I'm also
1:20:07
trying to avoid Going
1:20:09
after what I don't want to try and
1:20:11
compete with people like mend and sneak and
1:20:14
the big Companies doing
1:20:16
that sort of thing. Why not because I think
1:20:18
for some of it It's I
1:20:20
believe they're still gonna get what business
1:20:22
it's that's always true in competition But
1:20:26
I guess so I don't see it as like Able
1:20:29
to compete for some of the things and I don't
1:20:31
want it to try and compete some of the things
1:20:34
So on like the same a security
1:20:36
aspect for sure. I think that's like
1:20:39
a fairly saturated market and There
1:20:41
are lots of people who have a massive
1:20:43
share, but there aren't many people doing
1:20:45
things like give me a list of
1:20:49
Dependencies that haven't had code review
1:20:51
recently in the last 90 days How
1:20:54
many of those tendencies have used stuff
1:20:56
into prod without any code review? There's
1:20:59
things like that. So I can add some
1:21:01
interesting views on that data while also
1:21:05
Given the ability to add your own
1:21:07
stuff into it while also being able
1:21:09
to play nicely with things like
1:21:11
s-bones produced by other tools You
1:21:13
can use existing platforms. Yeah.
1:21:15
Yeah, keep doing it man. Don't
1:21:18
stop Even if you have
1:21:20
to overlap a little bit I think there's
1:21:22
certain things that obviously like sneak is just
1:21:24
a Goliath in the industry But at the
1:21:26
same time and I'm not trying to poo
1:21:28
talk by any means I think what I've
1:21:30
heard negatively about sneak is not they're bad
1:21:32
but that they report on CVS,
1:21:35
it's like known vulnerabilities. It's
1:21:38
not predictive. It's reactive
1:21:41
So I think if you can get you're kind
1:21:43
of like teetering in the predictive realm because you're
1:21:45
able to give insights that not They're
1:21:47
not easily foreseen otherwise
1:21:50
like has there been PRs in
1:21:52
the last 90 days on X number of my dependencies
1:21:55
They may not say that you have eight
1:21:57
bad dependencies and may just say these may
1:21:59
be one that you should question or
1:22:01
look into personally or use other tools
1:22:03
that give you more insights beyond this
1:22:05
because it's like it's a smell. You're
1:22:07
identifying code smell or dependency smell of
1:22:10
sorts. There's definitely something there. Keep
1:22:12
going. I am definitely planning
1:22:14
on. I'm not putting it
1:22:16
down anytime soon. It's been, yeah,
1:22:19
massive passion project. I'm
1:22:21
hoping to, yeah, go a lot further
1:22:24
with it. Yeah I think maybe Jared
1:22:26
might be onto something though with this TNT stuff, you know. Because
1:22:29
you can like bring in more
1:22:31
iconography and stuff like that around,
1:22:34
you know, like if you have bad dependencies,
1:22:36
that's a blow up moment, right? That's a
1:22:38
dynamite situation, right? It's a bad
1:22:40
thing for your code base to have sad
1:22:43
dependencies. So maybe
1:22:45
you could pivot the TNT versus
1:22:47
DMD. I don't know. I'll
1:22:50
have a look at the licensing fees for
1:22:53
using TNT as a song. Well,
1:22:55
there's always BMC and renditions, you know that right?
1:22:57
Great artist steal, Jamie. Great artist steal. That's right.
1:23:00
You can always do a rendition of the song.
1:23:02
It could be like the song, not
1:23:04
the actual. It's an homage, you know.
1:23:07
It's fair use. It's an homage. And
1:23:10
ask for forgiveness, not permission. I mean,
1:23:12
that's the indie web way, isn't it? I just made that
1:23:15
up. This
1:23:17
message is not vouch for by the
1:23:19
indie web. Yeah. Well,
1:23:21
I'm glad we're friends now versus friendlies. This
1:23:23
is so much cooler than the friendly aspect.
1:23:25
Yeah. It's been cool because
1:23:27
I've been listening to you both for years.
1:23:30
Although one thing that is interesting is listening
1:23:32
to you at one time speed. So
1:23:34
I usually listen on 1.7 times. So
1:23:38
I have actually listened to a couple of
1:23:40
podcasts at one time to prepare. Sounds weird.
1:23:42
We wouldn't disappoint you, you know, how slow
1:23:44
we are in real life. Oh,
1:23:47
man. Wow. There's so many
1:23:49
people who listen to podcasts that beyond 1X. I
1:23:52
have honestly tried several times
1:23:54
to just taste the sauce and see if
1:23:56
I like it, which I do not. Because…
1:24:00
You're talking about the speed of the
1:24:02
sauce? Yeah. I'm like, what am I…
1:24:04
am I missing something? Is something wrong?
1:24:06
And maybe I was telling Nick this
1:24:08
actually because we were hanging out when
1:24:10
he was here for
1:24:12
that conference recently. Nick Nissi,
1:24:14
by the way. And I
1:24:16
was like, maybe your brain
1:24:19
operates at a different frequency, like a
1:24:21
higher frequency than mine does. And so
1:24:23
it allows you to listen
1:24:25
at a higher speed because
1:24:27
you comprehend slightly faster than I do.
1:24:30
Not that we comprehend differently, but that the
1:24:32
frequency for which we comprehend is
1:24:34
different because I'm an
1:24:37
artist myself and I like the pure nature of
1:24:39
the art. So I can't
1:24:41
change the way – like
1:24:43
I'm not watching movies at 1.5, right? Like
1:24:45
I'm just not going to do it. It just – it
1:24:48
ruins the whole thing. I just can't do
1:24:50
that with anything else, whether it's a book
1:24:52
or a podcast or anything. I'm going to
1:24:54
listen to it at the director's intended speed.
1:24:57
And as close as I can get to
1:24:59
what the director intended, cinematically even if it's
1:25:01
a film, you know, like if
1:25:03
someone says you've got to watch Oppenheimer in the theater for
1:25:05
the first time, then you can watch it anywhere else you
1:25:07
want to afterwards. I'm going to do
1:25:09
my best to get to an IMAX theater with the rockin'
1:25:12
sound system or I'm just not going to
1:25:14
do it at all. I'm a teetotaler in that respect. So
1:25:16
that's my mode. So maybe that's –
1:25:19
you just listen and comprehend at a
1:25:22
higher frequency than I do. I have
1:25:24
both experiences because I listen
1:25:26
to our stuff sped
1:25:29
up 1.75 to 2X just
1:25:31
because I'm listening to clip. I'm not listening to
1:25:34
enjoy the conversation necessarily. And
1:25:37
what's funny about that is that I clip
1:25:39
out of Overcast and so
1:25:41
I'm listening at 1.75X, but when
1:25:43
I hit the clip button, it opens
1:25:46
up a brand new little clip
1:25:48
widget thing which does not have
1:25:50
the speed applied to it. And
1:25:53
so I actually get the fast and the
1:25:55
slow version multiple times throughout a single episode,
1:25:57
and that's when we sound really dumb.
1:26:00
I'm like, this is good. Then I hit the button. It's like,
1:26:02
yes, I know what I'm talking about. I'm
1:26:05
like, oh man, I'm really
1:26:07
slow. Ovenheimer. That's
1:26:09
a very strange experience. I
1:26:12
would not, I would not advise it. Yeah.
1:26:16
Back to back speed is probably in, wow, that's gonna be jarring.
1:26:20
It is jarring. So Jamie, sorry to disappoint
1:26:22
you with our regular speeds of
1:26:24
our brains, you know? That's how we operate. Hopefully
1:26:26
we didn't lose your attention. Too many times throughout
1:26:29
this conversation. We do appreciate you hanging out with
1:26:31
us. Yeah. And the
1:26:33
work that you're doing, your radical transparency
1:26:35
with your salary, definitely interesting
1:26:38
and helpful for many people, very cool. Your
1:26:41
commitment to, you know, syndicating and, and, you know, your website,
1:26:44
your commitment to
1:26:47
your website that you have is laudable. And this is a
1:26:49
cool project. DMD, check it out. Sounds
1:26:52
like TNT. So you know it's cool. And
1:26:56
you know, it's cool. And what else?
1:26:59
Anything else before we say goodbye to our friends?
1:27:01
No, yeah. Thanks very much for having me. It's been really fun. It's
1:27:05
been nice to speak. Shout out to the multiple
1:27:07
people who have requested to have Jamie on
1:27:09
the show throughout the years. Most recently it
1:27:11
was Dan Moore. That's right. Dan Moore. He
1:27:14
was a recent guest. And didn't you, you pointed
1:27:16
us to Dan, didn't you? Yeah, I did. Okay, did you
1:27:18
guys cut a deal? Did you cut a deal? No,
1:27:20
he messaged me after. Yeah, he
1:27:23
messaged me after, Dan, like. I'll request you. You
1:27:25
request me. He's not the first one though. Somebody
1:27:27
else used the request form and
1:27:29
submitted you a little
1:27:31
while back. And so it was percolating. The timing was
1:27:33
percolating. I had multiple requests. Plus
1:27:35
obviously known your work for many years. So I
1:27:37
thought, yeah, it's time. It's time to get him
1:27:39
on the show. I think the issue with that
1:27:41
one request though was they were recommending the TNT
1:27:43
project he was working
1:27:46
on. And they were wrecked and we
1:27:48
couldn't find it. Right. It was actually
1:27:50
DMD. Exactly. So we couldn't
1:27:52
find you. You need to redirect on that sucker. CNT.dev.
1:27:57
Check it out. Oh, that'd be cool. Can
1:27:59
you get that? We're over
1:28:01
here renaming his project. He's
1:28:04
like, I don't even like... He's like, I got a t-shirt
1:28:06
already, guys. I'm being polite. I'm not renaming it. I got
1:28:08
a t-shirt. I'm
1:28:11
checking it for you real quick. Let's close this loop, at least, and
1:28:13
then we'll let everybody go. Check domain.
1:28:15
TNT.dev is already taken. How about a different
1:28:17
one? Crying shame. That was the only challenge
1:28:19
we had. That was it. That was it.
1:28:22
Now we can't do it. All right. Goodbye,
1:28:24
friends. Bye, friends. Bye, friends.
1:28:27
Bye, friends. Bye, friends.
1:28:29
It just so happens that
1:28:31
Jamie is a changelog++ member.
1:28:38
So he was super forward to
1:28:40
stick around for a bonus 15
1:28:42
minutes and chat with us about
1:28:45
them, Zed, his reluctance to use
1:28:47
A.I. code tools, and more. If
1:28:49
you're not a plus plus subscriber, now
1:28:51
is a great time to directly support
1:28:54
our work with your hard-earned cash. Or
1:28:56
even better, your employer-sponsored education
1:28:59
budget. As a thanks, we
1:29:01
make the ads disappear, send you some
1:29:03
free changelog stickers, hook you up
1:29:05
with bonus content like this, and
1:29:07
more. It's better. But
1:29:09
don't take our word for it. Here's Jamie's. It
1:29:12
is better. It's been better for years. Join
1:29:14
Jamie and hundreds of your fellow listeners who
1:29:16
have already got in on the better at
1:29:20
changelog.com/plus plus. Make
1:29:22
your dreams come true. Thanks
1:29:27
again to our friends at fly2io
1:29:29
and to Sentry. Use code changelog
1:29:31
when you sign up to save 100
1:29:33
bucks on their team plan. And of
1:29:35
course, to our mysterious friend, Breakmaster Cylinder.
1:29:38
That's all for now, but let's talk again
1:29:40
real soon.
Podchaser is the ultimate destination for podcast data, search, and discovery. Learn More