Linus Torvalds, the creator and maintainer of the Linux kernel, joined me at the recent Open Source Summit in Vienna for an insightful conversation covering a wide range of topics. He discussed his ongoing role in the Linux kernel’s development, emphasizing his focus on merging and maintaining stability. Torvalds pointed out a significant shift in the open-source community, where development is now driven more by large companies with specific needs than by individual contributors. This change, he said, has resulted in a more structured and hierarchical approach to development. Torvalds also highlighted the importance of security, criticizing long embargoes on hardware security issues. He expressed skepticism about the current AI hype, preferring to focus on practical applications. When it came to programming languages, Torvalds also mentioned his preference for C over Rust and his dislike for synchronous meetings, opting for asynchronous communication via email. He wrapped up by expressing his continued interest in the intersection of hardware and software development.
Расшифровка видео
0:00
hi this is your host BH and we are here at open source Summit in Vienna Austria and once again we have with us after a
0:05
very long time Li worlds first of all it’s great to have you on the show again nice to be here again I think last time
0:11
was in several years ago I foret where it was it could be Vancouver it was
0:17
before Co time you know and then Co Disturbed literally everything so which brings me to the very first one what are
0:24
you doing these days I’m doing all the same things uh very little development
0:30
uh this week I’m doing a lot of merging because the merge window and
0:36
uh it’s been very similar for the last at least 10 years where the actual
0:43
coding is done by others and and I act as
0:49
a just a connection point and the occasional bad guy who says no this was
0:57
not right I can’t accept this in this form but um it’s Colonel development has
1:05
actually been very smooth I feel I mean we have technical issues but technical issues are things that we’re also very
1:11
good at solving so we we haven’t had any huge blow up problems in the last
1:18
several years by now yeah ecosystem has matured you know and adoption has grown how have you seen how are you happy how
1:25
much happy you are with the adoption of Linux across it it doesn’t even matter
1:30
well the thing is I’ve never cared right I mean to me I want to have I want Linux to be
1:37
widely used not because I care about like market share numbers or anything like that but because use is what finds
1:46
bugs and uses what also finds new interesting use cases that that then
1:51
impact the de development itself right but at the same time to me how widely
1:58
the colonel is used has never been the primary issue it’s always been about just the technical side so I don’t try
2:05
to I don’t do statistics I don’t try to follow what big companies use Linux and
2:12
and things like that uh it’s I have a big enough ego without having having to
2:17
try to figure that out I think you once said that you you have a small bigo as big as a planet right I’m not wrong
2:25
maybe yeah and is that still the case yeah okay but when we look at the world
2:31
around us you know almost everything is like powered by Linux without realizing which also like I was talking to D
2:37
before that the early days you know we were talking about why people should use Linux and now everybody is losing so
2:43
when you interact with the community how what kind of contrast you see today versus them I think the
2:49
biggest contrast is that the early days we had a lot
2:55
more uh individual contributors and people people had a lot of like personal things
3:03
that they needed to have fixed and these days I think largely because we’ve fixed
3:09
all the lwh hanging fruits uh a lot of the development comes from like big
3:16
sites or like huge companies that have very specific needs and uh usually the
3:24
needs are also better described just because the it’s it’s it’s more
3:31
professional I mean slightly in some respects it’s more boring it was more
3:37
wild west back when and uh and we were able to do more crazy things when when
3:43
we didn’t have quite as many big companies depending on the end result uh but at the same time I mean it is it’s
3:52
maybe not as crazy in a good way as it was I mean 25 30 years ago so do you
4:00
miss those Wild West days more or you like this you know more stable calm I like the more stable calm I mean it’s at
4:07
the same time I kind of miss the craziness I miss the fact that we could
4:14
do things that we weren’t as serious about I miss the fact that people didn’t take Linux as or me personally as
4:21
seriously as they do today I think today I have to be much more
4:27
careful about what I say because people take me too seriously right and and the
4:33
early days in many ways we’re more freewheeling and and more fun in that
4:38
respect at the same time I mean from a technical angle I much prefer the whole
4:44
we finally actually figured out how to do things right and it took us a long
4:50
time and and I do enjoy that fact that I think we have a much better process than
4:56
we used to have so you lose you lose a big of excitement you lose a little bit
5:02
of of Personality maybe but you need to lose it at some point and that some
5:08
point was years ago yeah that’s that’s true also as you’re saying though that early days a lot of contributors were
5:14
like individuals now it’s mostly company how does that how has that changed the colel community because at one point we
5:20
were also talking about uh that the colel community is you know growing older we need fresh how how just talk
5:28
about the strength and sustain how do you see at the Kel Community well I I
5:33
think the big change has not so much been like the serious commercial big
5:38
companies versus individuals the big change has really been that we’re just so much bigger right I mean 30 years
5:47
ago there were like tens of core developers and hundreds of people
5:53
involved and now it’s hundreds and thousands right uh and and that’s a big
6:00
change it it means that it’s you have less of the like knowing people
6:09
personally and you have to be able to scale more you have to scale your trust
6:15
so you can no longer trust everybody around you you have to find the people that are kind of your group that you
6:21
know and and the the way you scale development is that they then have more
6:27
people that they work with and Trust and that has changed the kind of
6:34
psychology of the thing in that it used to be that it was more one to one and
6:41
now I for example have to just rely on the top 50 or 100 maintainers that I
6:47
have interacted with a lot and and the Thousand or more other developers that
6:55
are involved in every single release I don’t have a one to one relationship with anymore and I mean
7:01
it’s not just I this is true of everybody that that we’ve had to kind
7:07
of become more organized and have more of a hierarchy
7:14
of developers that we didn’t have long ago and
7:19
um the commercial side I think happened coincided with this
7:26
change but I don’t think it’s because they they ended up being involved when it comes to Tech we often talk about
7:34
tech shortage companies I mean of course there a lot of layoffs are happening so but when it comes to the kernel Community do you feel that they are not
7:40
enough maintainers and contributor you’re like no we are very good shape now well we’re in good shape uh that
7:48
said I mean we do have I think more developers than maintainers I mean I
7:53
know we have more developers than maintainers but I think it would be healthier if we had more maintainers so
8:00
that the maintainers we have wouldn’t be as overworked sometimes it depends a bit
8:06
on the area uh but there are certainly different parts of the kernel and and
8:11
some of them are happier than others because they they don’t have the same I
8:17
mean some areas have more maintainers per developers and that kind of spreads
8:23
out the workload and the pain and the stress I mean being a maintainer in many
8:29
ways can be very stressful because now you have to you have you’re the one who in in
8:34
many ways are like the buck stops with you being an end developer who only
8:40
works on one small driver is less stressful because you only care about one small thing uh being
8:48
a maintainer of a big subsystem can be very stressful and and I’d like us to have
8:54
more maintainers but it’s not always a fun job right so I mean it’s not like
9:01
the the Cal Community is not like a organization where you can have strategies to bring in more M but do you
9:06
have anything things to kind of uh create like grow this maintainer well I mean the thing that
9:13
grows our maintainer base is that it may not always be the most Pleasant job but it is interesting right I mean
9:21
if if you’re looking for a technical challenge being a colonel maintainer is
9:27
I feel still after for 30 years is is one of the more interesting things you
9:33
can do and the good news is also I mean it is something that is appreciated not
9:38
just by other kernel maintainers but if you’re a kernel maintainer finding a job
9:44
in in the industry tends to be fairly straightforward so so it does have its
9:50
upsides are you seeing any kind of you know more interest for you know the or
9:55
you’re like this this might become a problem in 5 to 10 years from now when the current breed of Maintenance will
10:01
grow older or they Veer out or in another way what is the challenge that you face feel is there
10:09
for you so I’m not it’s not particular areas it’s uh it’s sometimes
10:17
particular people and workflows that I find worrying for example I used to I
10:22
used to worry about Greg a lot because he just took on so much work and and I
10:28
talked to him like several times and said hey you’re just spreading yourself too thin you’re doing so many things and
10:35
you’re being so important in so many areas are you sure this is healthy and
10:41
uh he always says that he actually really enjoys what he’s doing and he has
10:47
automated a lot of the work he does to a degree that he does not feel it’s that
10:52
stressful uh there are certain other subsystems where there’s only one person
10:58
who really does a lot of the work in one subsystem and I mean some people feel
11:04
that me saying that is a bit silly because I’ve been the one top level maintainer for a long time but at the
11:11
same time it’s like I feel like there are number of people who could do what I
11:17
do they may not want to do what I do it’s but but there’s no VI specialized knowledge it’s more
11:25
about being trusted in the community right we do have few areas in the kernel
11:31
where there is very specialized knowledge and there are very few people
11:36
that have that knowledge and I think that’s more dangerous than than my kind of situation where okay I am trusted in
11:43
the community but at the same time there are many other people who are trusted in the community too if you remember the
11:49
early days of arm or Nvidia that those were the new challenges in the modern
11:54
world what are the new use cases where you feel that the potential is there but there is some struggle or you’re like no
12:01
lytics has conquered everything so actually I don’t even care anymore well to some degree it is true that we are so
12:08
good at so many different areas that I don’t it’s not that I don’t care I don’t
12:13
worry so much uh there are I think to some degree the one use case where I
12:20
feel Linux is not necessarily doing as well
12:25
as some people wish it to do tends to be the microcontroller world and and that’s
12:31
where I do see projects where people do small kernels for smaller workloads
12:38
right and I think that’s natural and I don’t think that’s an area that Linux
12:44
will I’m not saying grow into because it’s really shrink into people have tried to make like a smaller footprint
12:52
Linux and we’ve tried that for a long time and it never really works out I
12:58
mean when you have a certain amount of basic infrastructure that everybody
13:03
takes for granted you have grown past a certain size and uh and so I accept the
13:13
fact that Linux won’t be everywhere because if you want a really small tiny device you are probably better off with
13:20
a really small tiny operating system too right uh the good news is usually the
13:26
small tiny devices tend to grow up and then they want to have Linux after all but but the end of moris law may change
13:34
that rule right yeah and it’s zipper and all those there a lot of R are there which are you know they’re but good news
13:40
is they’re all open source so they’re always and it’s not R it’s always and Z for and I do think that people have
13:47
realized so much that when you’re doing basic infrastructure and whatever operating system you’re doing it’s
13:53
always going to be basic infrastructure open source is the only way to do it there’s no there’s no money trying to
13:59
make a like a custom commercial small operating system there’s no point to
14:05
that so yes whatever comes and replaces Linux on the low end I feel is going to be open source and what I see today is
14:13
that almost everybody is using op elment no matter where we look everybody is using open source the challenge is that they don’t know that open source and
14:20
they don’t become good open source citizens so what are your thoughts on that well I mean that’s always been true
14:25
that there’s been a lot of people who are companies areas where people want to
14:33
use open source but they don’t want to participate and I I
14:38
feel okay that’s never going away uh some companies some people just
14:47
want to take what is offered and not give back and um that’s their choice
14:55
right and it means that they will not be part of future newer development and they can’t
15:02
like push the direction of the project in in the direction they might want and
15:08
that’s kind of the tradeoff and I don’t worry about I some people think that hey there are there are leeches on the
15:15
community and I don’t agree right I mean uh it’s it’s one of those things where
15:22
not not necessarily wanting to partake in development is fine too and sometimes
15:29
like they don’t have technical extra resources and they don’t do the they don’t have the melice sometimes they
15:35
just and also a lot of things has to do with awareness also you know well yeah uh there are there is bad behavior and
15:42
sometimes it’s literally because you’re a bad person or a bad company and I can’t fix that nobody can fix that uh we
15:51
do try to encourage people to do the right thing and companies to do the
15:57
right thing and ultimately there are obviously also groups that in the extreme case will actually look at legal
16:04
options right but how happily I have to say it is rare it is really rare and uh
16:12
and it’s not something I personally find worry some if you look at the very early days ofn you know you like interacting
16:18
with the developers you know the the young kids from the school colleges do you still interact with those young or
16:24
you’re like no you are just working with this mature I mean do you miss that that working with because they are the way
16:30
this young people saw I mean I’m not saying that you and I are old but still we are uh I don’t interact that much
16:37
with new developers anymore and part of it is just personality
16:43
issues I’m not a people person I’ve never claimed to be a people person I
16:50
even when I interact with other developers I really prefer to interact on a technical level and uh and the
16:59
kinds of problems we have now tend
17:04
to tend to be more about maintenance than about
17:10
some really revolutionary new feature right so the kind of people I interact
17:16
with tend to be other maintainers that’s not to say that occasionally somebody I
17:22
mean it happens I people know what my email address is and I ignore 90% of my
17:27
emails that I don’t feel I need to reply to but it still does happen that somebody sends a an interesting email an
17:35
interesting question a question that I feel like I’m actually one of the few people who can answer uh and and then I
17:42
will answer uh it’s that said exactly because I feel like I will only answer
17:50
emails that where I feel like I am the authority it tends to be very Oddball
17:58
specific area is right if you send me an email and you just ask me a random Linux question I’ll ignore it I I I won’t even
18:05
think very much about it I get too much email I will I will not delete it I will
18:11
just archive it and go on to the next one but occasionally you get like high
18:17
school students who get really interested in open source
18:22
and in Linux and they they still have interesting questions and I occasionally
18:28
answer those most questions aren’t interesting that’s the the colel development is still
18:33
through the Mals or uh yeah I’m almost entirely still email based I mean we
18:39
still we obviously use git for the actual code transfer but for all the
18:46
real like discussion for me it’s all email I I and
18:53
part of it is I I’m very good at ignoring the email that I don’t need to
18:58
interact with I absolutely abhor Zoom calls or any kind of
19:04
video thing or phones are even worse right uh and and they really work very
19:12
badly for konel development anyway because everybody’s in different time zones uh so email the whole asynchronous
19:20
nature of email really suits me much better than
19:26
some people like using chat right and real time interaction with people and
19:31
I’m like that’s not me but uh I the fact that I use all email doesn’t
19:39
necessarily mean that everybody else does I mean we do have clearly people who are still using IRC or more modern
19:49
versions of chat quite often right Discord whatever and uh and email ends
19:56
up being what I do it’s not necessarily what every other Colonel developer does which brings me to a different question
20:03
I will go back to the kernel is that do you not like to get on Zoom call because you still work bathroom because you want
20:09
say B bring you the point of how do you what is the work life look like it’s I
20:14
mean partly it’s the bathrobe issue although I say that jokingly much more than that it’s actually I hate the
20:21
synchronous meetings because then you have to you have to set it in your calendar and you have to be there at a
20:28
particular time and that so often for me breaks my flow right if I’m in the
20:34
middle of something and I’m getting stuff done the worst thing that can
20:40
happen is that my calendar alarm goes off and says okay meeting in five
20:45
minutes right and uh so I have actively avoided meetings I
20:52
mean basically forever even when I was at transmeta and working not on the carel but on on like as a developer
21:03
inside a company even when I was like in the office and the meeting was just walk
21:09
over a couple of rooms and sit in with people even then I absolutely hated it
21:15
just because if I happen to be working on something that when it breaks your concentration it just it takes you out
21:23
of the moment so I’ve never liked any of these communication models where you have to
21:29
synchronize with other people because it just doesn’t work well for me and and
21:36
that’s so that’s why I don’t like Zoom that’s why I don’t like phone calls but that’s also why I’m not a huge fan of
21:43
in-person meetings that’s said I like in-person meetings way more than Zoom calls I do not envy all those people who
21:52
who ended up doing eight hours of Zoom calls a day during the covid years I feel like if I have to have a meeting
21:58
I’d much rather have a meeting with a person in front of me right that’s what there is joke also this meeting could
22:05
have been just an email right that the whole thing uh what I’m also asking is what is your work life balance look like
22:10
what are the other and also no before we go there since you mentioned get you create a Linux you create a g then you
22:16
create a subsurface is there any other projects that or we should create a problem so that you can solve it I hope
22:22
not right I hope somebody else does them I all the projects I’ve had have been
22:28
issued issues where I felt nothing else was available right
22:34
get very much so I mean there were other source control projects obviously right
22:39
but none of them worked for my workflow at all right uh
22:46
so right now I actually feel like my needs are being met very well I mean
22:53
people people still laugh about the whole year of the desktop for Linux but
22:58
the fact is as far as I’ve been concerned the
23:05
last almost three decades have been the year of Desu for me because Linux has
23:10
been doing everything I want with some exceptions get and subsurface being
23:16
those but um um I don’t see anything else coming up but never say never yeah
23:23
no since when you brought the point of Linux desktop the fact is that it’s steam de you know this is a Linux you know I mean I have a steam deck and I
23:30
also have the Aus one you know so I play I’ve never actually even tried steam I’m just not a gamer at all I’m heavy gamer
23:37
so I have a steam deck I have this also you know like and you know of course I mean Android and you know it’s like all
23:43
the I mean the modern EVS you know they all run on Linux so it’s the desktop and
23:48
Microsoft with wsl2 Linux is you know it comes by default so I mean Microsoft has
23:54
ironically become the biggest distributor in yeah but I mean even before all the like real Linux desktop
24:02
things I think I mean looking back 20 years just
24:07
the web and using the web browser for so much change an enormous amount of of
24:14
situations where 20 plus years ago you had to have a Windows application to do like basic
24:21
stuff and now a lot of those random oneoff applications are they’re not
24:28
applic anymore they’re web pages so that’s I think being the big change
24:34
for the Linux desktop thing is that for a lot of people all you really end up
24:40
needing is pretty much a web browser right I mean that’s kind of the background I mean Chromebooks and things
24:47
like that right and then you do have Gamers who want to have something else right we don’t talk about these anymore
24:54
so in early days we used to talk about this a lot because the way we consume software has also changed Abol so how do
25:02
you see that word has changed around you and us I actually I mean I think we’ve
25:07
mostly plateaued but I do think that we’re going to see another Spurt of that
25:14
especially with now windows and the whole arm ecosystem and and that will
25:21
again push any remaining apps that try to be OS Centric and try to be like
25:28
Windows only it just gets so painful I mean it’s uh much easier to generally
25:36
try to support a web presence or a more
25:41
portable model than than the traditional software model that is already almost gone I think it’s it’s just getting less
25:48
and less relevant all the time which obviously for for Linux and especially the fact
25:55
that we support so many different architectures is a good thing right it
26:00
uh it used to really be I mean I used to not just run Linux but I run Linux on
26:07
Alpha and on power PC and and then arm well arm
26:12
64 and uh it’s gotten so much easier over the years because there’s so much
26:19
less of this one OS one architecture model left
26:26
in the software industry even Apple they move to their own arm chips you know M1 M2 M3 chips so everybody is moving to
26:35
arm but they’ve done it before they did I think they did a great job at the transition again uh I mean partly
26:43
because they just control their whole ecosystem much more than than
26:49
like certainly much more than the Linux ecosystem which is so fragmented and
26:54
spread out and hundreds of different companies is doing distributions and and
27:01
things like that so so Apple in many ways is a much easier
27:07
model uh because they control the hardware and they control a big portion
27:12
of their software stack too now I’ll go back to the the colonel Point once again
27:18
uh rust was there what are the languages that you are embracing or what are the
27:25
language that you look are quite popular you’re like ah not right now you know
27:30
what I’m still a sea person at heart I’m old it’s hard to teach an old dog
27:36
nutrics uh I mean what made me like rust was mainly that there was not something
27:45
really obnoxiously horrible about it like I’m famously not a huge fan of
27:50
C++ uh because I think that language made so many bad design decisions that I
27:57
it’s not that I can’t imagine programming in it and in fact subsurface these days is almost
28:04
entirely C++ but it’s not a language I
28:09
like uh rust is not a language I love but at the same time it doesn’t have it
28:17
hasn’t reached the point where I actively dislike it either and I think it’s interesting and I think it’s uh
28:24
something that can bring in new ideas and new people uh so rust is not going to be something
28:31
that I personally am going to use a lot I I it’s something
28:36
where hopefully as it becomes more common in the kernel I will get better
28:41
at reading it because right now I have to think a lot when I look at rust code right uh because it’s not natural it’s
28:48
not automatic like it is for code I’ve just looked at so long that I don’t even think about it I look at it and I see
28:55
what it does and and rust code for me is not there and maybe it will never be
29:00
there that’s fine too right and what about the rest of the main like Greg how do they look at these languages oh there
29:08
seems to be a wide variety some people really don’t want to have anything at all to do with rust uh Greg seems to be
29:16
he’s always very positive and and open
29:21
to different people and different idea he seems to be fine with rust uh I’m not
29:27
going to start naming names but some people are certainly not not nearly as open to the whole issue uh and that’s
29:35
fine I I I find it more amusing than anything else all
29:40
the flame Wars that have happened and honestly I mean if you look online and
29:45
look at a lot of the like negative discussions around rust
29:53
seem to not really be core Kel developers it it’s it’s a lot of random
29:59
people who just for one reason or another have a very strong opinion so
30:05
what are the other either languages or projects that are trying to get into the Linux and either you’re pushing back or
30:10
you’re excited oh yes let’s get them there I don’t think we have any other languages one I mean
30:16
the the other thing that is kind of similar to rust and that took
30:21
forever uh it took over a decade to make uh
30:26
clang uh compil the colonel so llvm I don’t I can’t even remember when
30:35
the whole llvm Linux project started but it’s a long time ago and uh I actually
30:43
got very personally pretty frustrated because having two different compilers
30:48
was actually something I really wanted to have and it took much longer than I I was hoping for but now for the
30:56
last couple of years we’ve had GCC and and clang
31:03
coexist in the colonel Community pretty widely so you have big companies like I
31:09
think Android is almost entirely clang compiled these days and has been for a
31:16
while so you have big companies that use clang even though GCC is still obviously
31:21
the traditional compiler and there are differences uh we do have like code
31:28
which ends up saying if clang then this kind of things but they are
31:35
fairly small and and uh and it hasn’t been a big issue in the
31:41
last year or so I mean not to say that I saw a patch
31:47
just very recently that had an ifdef clang versus GCC but uh but it’s it was
31:55
another project with smaller ramification that than rust but it it took a long time a
32:02
surprisingly long time so that’s why I think that okay we’ve only worked on this whole rific for a couple of years I
32:09
mean there was some effort before that too but I’m like the fact that it’s not
32:16
even close to ready I doesn’t worry me at all because I’ve seen these other projects that were much less intrusive
32:23
that took took a decade or long longer so uh I don’t see any other huge pushes
32:30
I mean people mention other languages but it’s uh nothing that looks all that
32:37
likely right now earlier you also used to mention that we never break the user space you know in a way have your
32:43
philosophy changed over time because the no I mean we still have a very hard rule that we never break user space but then
32:50
at the same time never is one of those black and white words that sometimes reality inter
32:58
comes in and and bites you in the ass um so we do end up having
33:05
like if there’s a major security issue which haven’t hasn’t actually happened
33:10
recently but we’ve had situations where we’ve just said we have to make this
33:15
change because our old Behavior was just it wasn’t that we that it was buggy it
33:23
was buggy in a way where we could it was a security Hass right uh there’s also at
33:31
some point like we have gone to come to the point where we just
33:39
can’t reasonably maintain certain old architectures so we stopped supporting
33:46
the original 386 a long time ago just because the maintenance burden
33:54
was too high for what it gave us so I mean in a sense that is breaking user
34:00
space because if you had a 386 based machine at that point and the carel said
34:06
we can’t support that anymore obviously we did break that situation so I mean
34:14
you have to always have some kind of balance between okay enough is enough we
34:19
can’t support that anymore but on the whole uh we still have the no you’re not
34:26
adding a new feature you’re not trying to fix a bug and if you break something else that’s not a bug bug fix that’s
34:33
that’s needs to get done some other way and since you mentioned security I’m also kind of curious that how the
34:40
security has changed over years this used to be someone else problem now there are a lot of cultural shifts happening like the death SEC offs and
34:46
the whole shift left movement uh what are your thoughts on security not just for the colonel because you folks patch
34:52
things but they don’t get patch or never applied in a way or you used to say you know the whole heartbe happen you
34:58
famously said you know bucks are part of software development process you know which I still quote these days how do
35:03
you see Security in the colal space so I mean security is always important but
35:10
what has been a problem uh the secrecy that goes along with some
35:18
of these security problems H is a huge pain I mean it it fundamentally doesn’t
35:26
work with the open source development model and uh so we’ve the big issues
35:33
with security to me tend to be with the security companies and the
35:41
people involved when they want to have 90-day embargos or especially when it
35:48
comes to the hardware security issues that we’ve had over the last five plus
35:54
years some of the embargos have been not not 90 days but more like a year and a
36:00
half where the hardware company says we can’t fix this and we don’t want anybody
36:07
to know about this so you can’t talk about it you can’t you can’t publish the
36:13
fixes for this until the end of next year and that has been a huge pain uh
36:21
and and it I pushed back pretty hard on that and
36:28
and uh I say that hey security bugs are just bugs and the
36:34
other side of that is pretty much any bug can be a security bug it’s I mean
36:40
some bugs are so simple that it’s really hard to misuse them but but quite often
36:46
we end up doing fixes to some silly small bug and nobody even realized until
36:52
a year or two later that the silly small bug could be misused and trivial fix
36:59
ended up later being a huge security fix right and then you have to go back and backport this thing that you didn’t
37:05
think was so important to to much older Colonels too so security can be painful
37:14
but as far as I’m concerned from a development standpoint I’m trying to make sure that
37:20
we treat them as just important but still regular bugs as far as
37:28
possible right now let’s just talk about some of the general trends that are happening and I want to hear your thoughts on that and one of the you know
37:34
hottest topic these days is geni uh how do you look at it and if the of course
37:39
you folks on everything on the mailing list so it’s not even about embracing some of those Tech but what are your thoughts on these some of these modern
37:45
technology that are happening these days just be specific let’s let’s say
37:51
geni for example yeah and that was what I was assuming you were aiming for
37:57
I you know what I think AI is really interesting
38:03
and I think it is going to change the world and at the same time I hate the hype cycle so much that I really don’t
38:11
want to go there so my approach to AI right now is I will basically ignore it because I
38:19
think the whole tech industry around AI
38:24
is in a very bad position and it’s 90% marketing and 10%
38:32
reality and uh and in 5 years things will change and and at that
38:39
point we’ll see what of the AI is getting used every day for real workloads instead of just
38:46
chat GPT makes great like uh demonstrations and it’s
38:53
obviously being used especially for I mean in a many many areas but I mean
38:58
Graphics design things like that
39:03
but I really hate the the hype cycle and it’s not just AI it’s I think it’s an
39:12
industry problem I mean before AI a couple of years the only ago the only thing people talked about was
39:19
crypto and and I I just don’t I don’t like the hype cycle and
39:26
part of it is I fundamentally am the kind of person who I find something interesting and I
39:33
stick to it right and this is why I’m doing the carinal 30 years later right
39:38
and then I find it so in the the people who are like Social Butterflies and that
39:44
fly from one problem to another and talk about what is cool today instead of what
39:50
is cool over a decade I I just it’s not my thing even I as a writer I use it for
39:57
like lowlevel you know work but you have to on top of that you have to create value on it’s like but you’re absolutely
40:03
right the hype is the problem not the actual technology so if I flip the question what are the modern technology
40:08
that you are excited about you like hey this is something really interesting even it’s catching
40:15
your interest I don’t know I so I used to say that I I thought that if you
40:21
actually wanted to do something interesting you shouldn’t be in it at all you should be in like bioscience es
40:27
or something like that and uh but I used to say that like over a decade ago and it
40:35
doesn’t seem to be all that much more interesting now right so I I think it’s
40:42
not that there is like the one interesting area it’s it’s whatever is
40:49
interesting to you and to me it’s still Colonels I I’ve always been it’s not
40:54
that it’s Colonels per se I’ve always been interested in kind of
41:00
the the era between hardware and software right which is why I do Kels
41:06
but I also like compilers and that has always been my interest but the fact
41:11
that it’s my interest doesn’t and shouldn’t matter for anybody else and I
41:17
think the same is true of new technology it’s not that there is this one new
41:23
technology that everybody should be excited about and that’s actually like again what annoys me about AI it’s
41:32
that people should try to find what they’re interested in and then new technologies can be the thing that makes
41:40
it interesting I for example I mean just looking back I I thought Raspberry Pi
41:45
was great not because the Raspberry Pi Hardware is all that interesting but
41:51
because it made these like microcontroller Hardware so widely
41:56
available and and by making it widely available you found a whole new audience
42:01
of people who could do new things right it’s not
42:06
that Rasberry Pi was revolutionary in itself but I think it made the whole
42:12
like people are still playing around with wearall Computing and doing these random small projects and I think that
42:19
was great right and and that allowed people to find not the one new thing but
42:27
the thing to them right uh so and I don’t know what to me there’s nothing
42:35
new on the horizon because I’m still interested in that hardware and software area I
42:45
mean I just got the fifth generation of Rasberry by5 from Microcenter couple of weeks ago because it’s always out of
42:50
stock you can never find and I have like dozens of Rasberry pie uh what but I mean the the RP 5 now it’s fairly
42:59
expensive it’s like 100 bucks it used to be $35 you know I think
43:06
the in to some degree rber Pi started this whole cheap Hardware thing but now
43:13
there’s a lot of other small cheap Hardware that you can get for your own embedded projects so so now the
43:21
raspberry pies have obviously gone like they’re trying to go slightly more upscale and uh and that’s good for the
43:28
people who want more of a real computer but I think you will find a lot of the
43:34
people who were excited about the original RPI are now using these new
43:40
random small embedded platforms like yeah we wanted something tiny really tiny we can just shove it anywhere in
43:46
small box and that’s what I do a lot of projects to that’s why uh what are your thoughts of EVS because EVS are all
43:53
software mostly I mean it’s like computer on Wheels you know okay the driving part is the secondary the more
43:58
any thoughts you have on that self-driving EVS or whatever does that excite you I like EVs and I we switched
44:07
most of our cars over I we have a Volvo all electric EV is our main car these
44:13
days I like them because I don’t like combustion engines uh the instant torque
44:20
at low RPM just makes them more fun to drive to me I’m
44:25
not that interested in the whole self-driving I
44:31
mean I we had a 20-year-old car that we replace with a modern EV and I like the
44:38
fact that okay it does Lane following and it does all the basics that our old car did not do but at the same time yeah
44:45
it’s like a small detail in in the big picture right uh but at the same
44:53
time I the reason people think of eeve as Electronics on Wheels is because the
45:01
electric motor itself is so much simpler than the combustion engine right if
45:06
you’re doing a traditional combustion engine you need to have Decades of history with combustion engines to make
45:13
a good engine and you can see from the EV market today that hey you don’t need
45:18
a decade of experience to make a a a good electric motor right because they are so much simpler I mean I don’t the
45:27
number of moving Parts in a traditional engine and an electric motor is like two
45:32
orders of magnitude different right what I like about EV more is less about as you said the self driving but that is
45:39
allof like new features everything is software driven so you can fix things patch things so is your Volvo running on
45:46
your own pel or you’re still running the Volvo’s Os Oh no have you tinkered with it at all or you don’t touch pretty sure
45:51
it runs Linux in there but I don’t touch it right it’s uh yeah know it it has the
45:58
it’s like an appliance to you yeah it’s absolutely an appliance to me I’m not I’m not a car person and I say that as
46:05
somebody who actually had a sports car for 20 years I had this small two
46:10
theater convertible that I really like driving but even then I was never really
46:16
a car person so a car to me is just get convenience yeah yeah and I I like it to
46:24
be like easy and comfortable but I don’t care
46:31
that much apart from that it’s one of the reasons I actually ended up getting rid of my
46:38
sports convertible was that I decided that hey one of the advantages of EV was
46:44
this whole smooth acceleration and not not having to
46:50
change gears to me made the car so much better while a lot of like serious car
46:56
people they hate that part they want the rumble of the V8 they want I drive a Mustang so
47:02
I can understand they like whole manual gear changing and like yeah I learned to
47:09
drive a car that way no manual is not great I’d much rather have
47:15
no gears at all right so yeah I’m not a serious car person any new hobbies that
47:21
you picked no I I’m still I I’ve never had a lot of hobbies my my main hobby is
47:29
reading and uh scuba diving is still like when I travel I try to do scuba
47:36
iing not this trip not so much but uh but I like we have our family vacation
47:44
in a couple of weeks and I’m going to Hawaii and I’m going to do scuba diving
47:50
and what kind of readings do you do crap oh absolute disgusting crap um I used to
47:57
do hard science fiction like the traditional Asimov hind line kind of I mean
48:04
uh I got a Kindle I we had a shelves and shelves of books at home
48:13
to the point where it took up a lot of space and I got a Kindle many many years
48:18
ago and noticed that I what I actually like doing is getting the really cheap
48:23
like completely forgettable books I read not to inform myself or to become a
48:30
better human I I read as a Pastime right and so I these days I read random
48:38
fantasy still read science fiction um nothing serious at all anything else
48:44
that you like to talk about or you think we are because want some no I don’t have a message I mean I’ve said that before
48:51
that it’s it’s like it’s not like I try to push my world be on others that that
48:56
I I think I think open source is interesting but at the same time I was
49:02
never one of the free software people who thought that it was what everybody else should do it’s something I want to
49:08
do but I I don’t have a message I don’t have anything I want
49:14
to tell other people to do or tell about myself it was great you know to sit down
49:19
and talk just to kind of refresh the memories also walk through the memory line so thank you so much and you know
49:25
as usual I would love to you know talk to you whenever you are interested again but I thank you so much at a conference
49:31
in a couple of years again when there’s something new that’s going on
49:39
[Music]
49:56
d