End-to-end testing of mobile and web-applications with Paloma Oliveira from Saucelabs
Cloud CommuteJuly 05, 2024
1
x
19
00:26:4524.5 MB

End-to-end testing of mobile and web-applications with Paloma Oliveira from Saucelabs

In this week's episode of simplyblock's Cloud Commute podcast, host Chris Engelbert welcomes Paloma Oliveira, a Brazilian growth engineer at Sauce Labs and active member of the Python Software Verband and PyLadies Berlin.

In this episode of Cloud Commute, Chris and Paloma discuss:

  • The role of Sauce Labs in end-to-end testing for software and devices
  • The importance of cross-platform testing in mobile and web applications
  • Balancing complexity and simplicity in software development and testing
  • The future of AI in responsible development and ethical usage

Interested to learn more about the cloud infrastructure stack like storage, security, and Kubernetes? Head to our website (www.simplyblock.io/cloud-commute-podcast) for more episodes, and follow us on LinkedIn (www.linkedin.com/company/simplyblock-io). You can also check out the detailed show notes on Youtube (www.youtube.com/watch?v=1yDKIGhnTCo).

You can find Paloma Oliviera on X @pcultural?lang=en and Linkedin: /discombobulateme.

About simplyblock:

Simplyblock is an intelligent database storage orchestrator for IO-intensive workloads in Kubernetes, including databases and analytics solutions. It uses smart NVMe caching to speed up read I/O latency and queries. A single system connects local NVMe disks, GP3 volumes, and S3 making it easier to handle storage capacity and performance. With the benefits of thin provisioning, storage tiering, and volume pooling, your database workloads get better performance at lower cost without changes to existing AWS infrastructure.

👉 Get started with simplyblock: https://www.simplyblock.io/buy-now

🏪 simplyblock AWS Marketplace: https://aws.amazon.com/marketplace/seller-profile?id=seller-fzdtuccq3edzm


01:00:00
"I don't want AI to write my text so

01:00:03
I can have more time for

01:00:04
laundry. I want AI to do my

01:00:06
laundry so I can write more texts."

01:00:12
You're listening to simplyblock's Cloud Commute Podcast,

01:00:15
your weekly 20 minute

01:00:16
podcast about cloud technologies,

01:00:18
Kubernetes, security,

01:00:19
sustainability, and more.

01:00:21
Hello, everyone. Welcome back to

01:00:22
this week's episode of

01:00:24
simplyblock's Cloud Commute podcast.

01:00:26
Today I have with me another

01:00:27
incredible guest, Paloma. I hope I

01:00:30
pronounced that correct.

01:00:32
But she can actually say a few

01:00:34
words about herself. Hey, Paloma.

01:00:37
And, whoops, happy to have you.

01:00:40
Just spilling my water all over

01:00:43
the place. Happy to have you.

01:00:46
Maybe just start by introducing yourself.

01:00:49
Yeah, hello. It's a

01:00:51
pleasure to have my coffee with

01:00:54
Chris and you all.

01:00:58
I'm Brazilian but I'm now a person

01:01:03
living in the countryside, which I

01:01:05
appreciate very much.

01:01:07
It's sunny, which is also

01:01:09
appreciated. Summer's almost

01:01:11
here. I am waking up,

01:01:18
coming back from a conference from

01:01:20
PyCon, and PyData in Berlin.

01:01:25
That's an introduction because

01:01:27
that conference is, wraps up a lot

01:01:30
of what I do. I'm currently part

01:01:33
of the Python Software Verband. I am

01:01:35
part of PyLadies Berlin. We're

01:01:37
very connected to PyLadies

01:01:38
Germany. And on top of that, I

01:01:43
talk to a few humans. I love

01:01:45
talking to humans and I'm a

01:01:47
growth engineer at SARS-CoV-1. All

01:01:50
right. Cool. So

01:01:51
where are you right now?

01:01:52
You're certainly not

01:01:53
in Berlin. I see that.

01:01:54
I'm in a small city close to Munich.

01:01:59
It's called Erlangen, here in Germany.

01:02:02
All right. And

01:02:03
there's weather like that.

01:02:07
God. I'm in the wrong place. All

01:02:10
right. You mentioned Sourcelab.

01:02:13
Tell us a little bit

01:02:13
about Sourcelab. What do you guys do?

01:02:16
Yeah. Sourcelab is, and it's

01:02:20
funny because the name

01:02:21
it gave us the opportunity to

01:02:24
create the best swag if you ever

01:02:26
need us in a conference. Please

01:02:28
ask for it because it is

01:02:29
delicious. We give away sources.

01:02:31
We are a platform for testing

01:02:33
and it allows you to do loads of

01:02:35
testing. The whole idea is to

01:02:39
allow you to test your whole

01:02:42
pipeline and whatever you need, a

01:02:46
real device, a specific device, a

01:02:49
browser that you cannot find

01:02:51
anymore. Basically, we guarantee

01:02:56
that your product will be

01:02:58
delivered really flawless.

01:03:02
And to me as a growth engineer,

01:03:04
that you're impactful and we take

01:03:06
for granted a lot.

01:03:09
I have this vision of, and I said,

01:03:10
I like humans because of that

01:03:12
intersection of engineering and

01:03:14
humans for me, it is very

01:03:16
important. It is what we're doing

01:03:18
this for. Why am

01:03:19
I spending so many

01:03:20
hours in this issue? Why am I

01:03:22
doing that? But then I remember

01:03:24
who's going to use it and they

01:03:25
will facilitate their lives or

01:03:27
then improve their lives a little

01:03:28
better and it

01:03:29
makes a lot of sense.

01:03:31
And this is what he wants to

01:03:32
guarantee that whatever you're

01:03:34
doing, whoever is using has a

01:03:37
really quality and they're happy

01:03:38
and they're really meant to do

01:03:39
whatever they're supposed to do.

01:03:42
Right. I think you kind of

01:03:44
mentioned to that it's like about

01:03:46
the whole pipeline.

01:03:47
A lot of people think about when

01:03:50
they say, okay, we need testing,

01:03:52
we need unit tests,

01:03:53
we need integration tests. They

01:03:54
often think about all the backend

01:03:56
technologies and like, yeah,

01:03:58
I mean, we need to make sure that

01:04:00
every query works, we need to make

01:04:02
sure that the interaction

01:04:03
between the different

01:04:04
microservices work. But I think

01:04:06
SourceLab is more

01:04:08
specifically trying to do

01:04:09
the whole end-to-end testing,

01:04:11
including, as you said, browsers

01:04:12
and mobile phones and that.

01:04:15
Right. Why is that important? I

01:04:19
mean, I'm a backend engineer. Why

01:04:20
would I care about front-end

01:04:22
testing? Mean question, I know.

01:04:28
You can do API testing with them

01:04:30
as well. You can do, I mean,

01:04:34
testing is a fun

01:04:35
thing. And one of the

01:04:38
main mentors who introduced me to

01:04:40
the testing board was Christian

01:04:42
Brouman. That is one of the,

01:04:44
it is the person behind WebDriver.

01:04:46
And I admire him deeply. And

01:04:51
whenever talking about

01:04:52
testing, I like his approach. I'm

01:04:54
always like, you do something and

01:04:55
you make sure it works,

01:04:58
the block of the things. And we

01:05:00
don't have those things separated.

01:05:03
Whatever you ask from your

01:05:05
client is coming from a backend.

01:05:07
If anything breaks, the endpoint

01:05:11
is not going to work. And

01:05:13
I think it's easier, like all the

01:05:14
front-end of the, what is directly

01:05:17
at the end of the user?

01:05:20
That's my end thing. This is how

01:05:22
the user, so I think it's, that's

01:05:27
the last point of contact.

01:05:29
The backend will come and I will

01:05:31
be seeing here, but maybe it's

01:05:32
easier when you think backwards.

01:05:35
It's a very strikeable thing. I

01:05:38
don't know if it makes sense, but

01:05:40
I like to do this backwards

01:05:42
perspective all the time. It's

01:05:44
like, okay, but what it really

01:05:46
matters for me

01:05:47
particularly is what the

01:05:48
user is seeing. If there is

01:05:50
something, and this is quite funny

01:05:51
because sometimes we have this

01:05:53
engineering meetings and it's

01:05:54
like, of course, all the engineers

01:05:57
love to create challenging things.

01:06:00
We don't like to always create a

01:06:03
CSS that fix, but most of the time

01:06:05
the CSS will fix.

01:06:08
Because it's like, yeah, maybe we

01:06:09
don't need to reinvent the wheel.

01:06:11
Maybe if you just leave a

01:06:12
notification or a tote for a

01:06:15
better color to see the button

01:06:16
would solve the issue. And for me,

01:06:19
this is why it's so important. Not

01:06:22
the complexity of the thing, but

01:06:23
how this person that is using,

01:06:26
it would feel happy using.

01:06:28
It would make sense.

01:06:29
I think you mentioned something

01:06:30
very important. And that is, as I

01:06:32
said, I'm coming from a background

01:06:34
engineering perspective, but also

01:06:37
as I try to implement an API or as

01:06:41
I try to design an API,

01:06:43
I always try to look at it from a

01:06:44
user's perspective. You

01:06:46
start with the contract or

01:06:48
interface, whatever you want to

01:06:49
call them, and you design the API

01:06:52
in a way that you want it to be

01:06:54
used or as a user you like it to

01:06:56
use. And you don't think about it

01:06:58
from the implementation

01:06:59
perspective yet. And that

01:07:03
basically giving you the hard time

01:07:05
afterwards, but

01:07:06
making it easy for the

01:07:07
user. And I think with front end,

01:07:10
as you said, it's kind of the same

01:07:11
thing. It's the first thing

01:07:12
when we use an app or a web

01:07:14
application or anything like that,

01:07:16
it's the first thing we see.

01:07:18
And if that thing completely

01:07:19
breaks or falls apart, and

01:07:21
especially with mobile phones,

01:07:23
I mean, we all love to laugh about

01:07:26
browsers and how different they

01:07:27
show the same thing.

01:07:29
That kind of, well, that thing is

01:07:32
mostly solved because it's either

01:07:34
chromium based or it's

01:07:36
webkit based. So that kind of

01:07:39
thing solved itself. But we have

01:07:41
the same thing with mobile phones,

01:07:42
especially in the Android

01:07:44
ecosystem where it's basically all

01:07:47
Android. But because everyone's

01:07:49
trying to be different now based

01:07:51
on the fact that they have their

01:07:54
own UI overlays or whatever you

01:07:57
want to call those things, apps

01:07:59
still behave very different. And I

01:08:02
mean, especially when you're on

01:08:05
the smaller brands, you see a lot

01:08:07
of phones and a lot of apps that

01:08:08
don't really work the way it

01:08:10
should be because they thought

01:08:12
like, oh, like dual screen is

01:08:15
amazing. And none of the apps

01:08:17
understand how dual screen works,

01:08:18
except for their own, basically.

01:08:22
So yeah, I kind of get it.

01:08:24
I think it's important to really

01:08:27
have this like cross platform

01:08:29
testing. And I think to do that,

01:08:31
you probably have to have a

01:08:33
platform either like SourceLabs

01:08:35
or I know there's a few more,

01:08:37
especially for mobile phones,

01:08:39
because you have to have this like

01:08:41
massive range of

01:08:43
devices to test on.

01:08:45
And it is, I will say that

01:08:47
I although as a developer, I find

01:08:50
it of course, annoying,

01:08:51
I want to just push it to one code

01:08:53
that works in everywhere, they

01:08:54
have to think about it. But

01:08:58
I really did. And this is why I also

01:09:02
defend very defensive of open

01:09:07
standards. So this,

01:09:08
if we use the standards, we could

01:09:11
do our own thing, but

01:09:13
it's to meet expectations.

01:09:16
That's what means why it's important. Of

01:09:17
course, that doesn't happen in

01:09:18
internet, internet is crazy

01:09:20
place and all the vendors of

01:09:22
browsers, they well, it's all a,

01:09:24
all a problematic

01:09:25
thing about finding

01:09:26
the standard. But as there's

01:09:30
another side of me that supports

01:09:33
and ferments net art or art for

01:09:35
the internet. And then is when it

01:09:38
becomes really interesting.

01:09:42
Because then you can make a

01:09:44
use of this that one would call

01:09:46
stake or a mirror and make use of

01:09:49
it in an interesting way. And

01:09:51
recently, that for me was, I mean,

01:09:54
I haven't been developing for so

01:09:56
long. So I haven't developed

01:09:58
doing together with the evolution

01:10:01
of internet. But about two years

01:10:03
ago, I had this, I created

01:10:05
this exhibition, and one of the

01:10:06
artists, which is a super young

01:10:08
artist, Fabiola

01:10:09
Lydis, she developed

01:10:11
this art piece that kind of wanted

01:10:14
to look like old time old school

01:10:17
browsers. And we have the

01:10:20
center for that school. So

01:10:22
the center for that art where we

01:10:24
fermented the net art project.

01:10:27
And she developed, we have this

01:10:29
really old Mac. It's actually not

01:10:32
that old, but you know,

01:10:33
those colorful ones.

01:10:35
Yeah, the iMac thing.

01:10:38
And she developed

01:10:39
for that. And it was so fun.

01:10:42
But you had so many issues about

01:10:44
how much has changed within the

01:10:46
times, even just HTML and CSS.

01:10:49
But the funniest is how can you

01:10:51
replicate his experience? And

01:10:53
everything was so slow and keep

01:10:55
crashing. And you could not have

01:10:56
more than one tab open at the same

01:10:59
time. And you look at really

01:11:00
different than her computer than

01:11:03
it was in that computer. But it's

01:11:05
so different. There was like,

01:11:07
I should say, see, why is it

01:11:10
different? She of course was

01:11:12
developed with a MacBook Pro that

01:11:13
it's a new generation. It's

01:11:15
probably a M1. And it was a

01:11:20
quite fun experience. And then you

01:11:22
understand the importance of the

01:11:24
testing. That was playful, right?

01:11:26
This is no one would depend on it.

01:11:28
But what if it's a doctor that

01:11:32
really needs an interface to access

01:11:35
an exam? And then

01:11:37
you have something

01:11:37
that didn't supposed to be there.

01:11:39
I don't know. I'm tripping now,

01:11:40
right? I'm not sure if that's an

01:11:42
actual real case. But what is

01:11:44
something that is really

01:11:45
fundamental like automated cars or

01:11:51
control towers for airplanes?

01:11:54
Software is everywhere.

01:11:56
Or airplanes itself. We're not

01:12:00
talking about Boeing here.

01:12:05
Yeah, but I mean, Software, it's not just the standards

01:12:09
that we do software is everywhere.

01:12:11
It's called literally

01:12:12
controlls our lives. There's lots of

01:12:16
cases when Google Maps is started.

01:12:19
And if you go to some,

01:12:20
I'm originally from Brazil. And

01:12:22
there some places like in Rio de

01:12:24
Janeiro, you really have to pay

01:12:25
attention or you do when I get to

01:12:27
the wrong street, whatever that

01:12:29
cultural means, it's not

01:12:30
like we can find information on

01:12:31
the internet. And sometimes the

01:12:34
map with a guide to those places,

01:12:36
and there's several, they put

01:12:38
people either in a very dangerous

01:12:39
situation, or even when the roads

01:12:41
stop, and then people just keep on

01:12:43
following blindly there. We have

01:12:46
this connection of trust

01:12:47
and we literally guide us. Oh,

01:12:50
that was always a fun literally

01:12:51
was guiding you through our life.

01:12:55
Yeah, it's true. I think you

01:12:56
mentioned an important fact

01:12:58
software is everywhere. Even

01:12:59
even if you think of things like

01:13:01
hardware, like embedded systems,

01:13:03
there's still firmware in it

01:13:05
and firmware can contain bugs. And

01:13:08
if we if we think about the you've

01:13:11
probably seen the picture

01:13:12
I forgot her name, but the lady

01:13:15
that worked on the Apollo project

01:13:16
standing next to the source code

01:13:18
printed out in like a 1.5m to

01:13:21
meter kind of stack of papers. I

01:13:26
find this really incredible

01:13:29
how much code they've written and

01:13:32
tried to test without actual real

01:13:35
world possibilities to test

01:13:37
it and it still worked days like,

01:13:39
"hmm, looking at today's software

01:13:41
development, that is far from

01:13:44
what it is." Or not going. Yeah,

01:13:49
we're not going that that far

01:13:51
back. When games were delivered on

01:13:54
cardridges, the game pretty much

01:13:57
had to work, at least to an extent

01:13:59
that you don't have to recall

01:14:01
all the cartridges, rebuild them

01:14:03
and send them back to the owners.

01:14:05
These days, the first thing

01:14:06
you do when you buy a game is

01:14:08
downloading a patch for it because

01:14:10
it was probably printed

01:14:11
way ahead was not ready

01:14:14
was still like very bug driven.

01:14:20
Software development changed

01:14:21
somehow.

01:14:23
That's so interesting.

01:14:24
Why do you think that you

01:14:27
have my assumptions, but like, why

01:14:28
do you think we're shipping code

01:14:31
without caring that much?

01:14:34
Why I think that happens. I guess

01:14:38
it's twofold. First of all,

01:14:42
especially with games, stuff gets

01:14:44
more complex and even embedded

01:14:46
firmware and stuff, right? The

01:14:48
Apollo system was

01:14:49
really fairly simple.

01:14:52
So the complexity of software

01:14:54
changed quite massively. And I

01:14:57
guess the development times,

01:15:00
the pressure, the deadlines, all

01:15:03
that changed as well. Well, one

01:15:06
got bigger, the other one shrank.

01:15:09
So I think that is a big part of

01:15:11
it. But I guess the bigger part is

01:15:12
probably the complexity,

01:15:13
to be honest. And that is why it

01:15:16
is so important to actually test

01:15:17
your source code. Which is funny,

01:15:20
because a lot of startups

01:15:21
initially don't because you're

01:15:23
iterating too fast to

01:15:24
actually make anything

01:15:25
testable for real. So anyway,

01:15:30
looking back at SourceLab, how do

01:15:33
I have to think about that as

01:15:35
a developer? I saw on the website,

01:15:38
you talk a lot about low code. So

01:15:41
I think you're basically

01:15:43
having building blocks to build

01:15:45
together your testing pipeline.

01:15:47
How do I have to think about that?

01:15:49
Where the platform and you use it,

01:15:55
whatever suits your needs, we

01:15:58
allow you to feed your needs

01:16:00
within so you can test it in

01:16:02
scale. And so you don't have to

01:16:04
keep pushing the

01:16:05
button. So we will

01:16:06
give you like, for example,

01:16:07
insights, or you can just ship it.

01:16:10
Oh, there's a really pile of alerts.

01:16:14
There's a soon there will be an

01:16:17
integration with JetBrains where

01:16:20
you can test your real devices

01:16:21
directly on their IDE. That's a

01:16:23
very cool one. We're open source. I'm proud of

01:16:27
the team. We facilitate from that,

01:16:31
because especially in scale, it's

01:16:33
very hard. It is also interesting.

01:16:37
Like I knew the biggest

01:16:38
case anything that we most use for

01:16:40
bigger sales. But me as a

01:16:44
developer, when I have to do

01:16:46
smaller projects, and I really

01:16:48
want to test all the rows, I

01:16:49
don't have all those devices. So I

01:16:51
really use it for testing, make

01:16:53
sure the quality is in small

01:16:54
devices. So that's

01:16:55
like a case in there.

01:16:57
I'm ok that all the

01:17:00
smaller case, right? Meaning like

01:17:02
me as a tiny developer, just want

01:17:04
to make sure my website works,

01:17:05
but it's really useful when you're

01:17:08
a big company, then you see a huge

01:17:11
numbers in speed, performance

01:17:15
and catching up. Yeah, that's it.

01:17:20
I find it really interesting when

01:17:22
you said that the complexity or

01:17:24
that with the complexity of the

01:17:26
software today, it's not just the

01:17:29
software everything needs to go with it.

01:17:34
And one of the types that I

01:17:37
find most important that for that

01:17:38
we do have integration,

01:17:39
we take you, it is accessibility

01:17:42
testing, we need to think not only

01:17:45
the complexity, it's not just

01:17:47
doing the software itself. But

01:17:49
finally, we as human kind,

01:17:51
starting to

01:17:52
include other humans that

01:17:53
are not part of what we call

01:17:55
normal, which is very massive and

01:17:59
very bad. Forgetting the words

01:18:02
here, it's a very bad notion of

01:18:04
how a human should be a human

01:18:07
within a variety

01:18:08
of human kind, how

01:18:10
people receive the world. And that

01:18:12
for me is the most important that

01:18:14
it is so funny, but because

01:18:16
we gather for all the complexities

01:18:20
of cloud, it's super complex, it's

01:18:23
such abstract concepts on

01:18:25
top of abstractions. But somehow

01:18:27
our softwares cannot even match

01:18:30
with the basic of the basic

01:18:31
color contrast, which is like a

01:18:33
more basic. And I keep thinking a

01:18:36
lot about that, how we as humans,

01:18:38
we can go so far ahead in some

01:18:41
complexities. And this is why for

01:18:43
me, being in this position

01:18:45
between you as an engineer is

01:18:47
always kind of feet on the ground.

01:18:50
It's like, why are we doing

01:18:51
this for the needed complexity?

01:18:53
And most importantly, what is the

01:18:55
complexity we need to

01:18:56
be paying attention to?

01:19:01
It comes down to the saying that

01:19:03
software engineers or software

01:19:06
engineering is the only

01:19:08
profession that learned how to

01:19:11
make sure their jobs are

01:19:12
sustainable by

01:19:13
just introducing more

01:19:14
complexity, by hiding more

01:19:16
complexity, or at least saying

01:19:20
they hide complexity. I mean,

01:19:23
something like Kubernetes is

01:19:24
really nice, but it hides a lot of

01:19:27
complexity. As long as everything

01:19:30
works, it's an amazing piece of

01:19:32
technology. But the second something

01:19:34
fails, you better understand

01:19:36
how it works. And that is a reason

01:19:38
why a lot of people say it's a

01:19:39
platform to build platform,

01:19:40
it should never go into the end of

01:19:42
the end user in the sense of the

01:19:44
actual people running the

01:19:47
applications. Let that be a thing

01:19:50
for the big companies like Google,

01:19:52
Amazon, whatever, that

01:19:54
actually offer container runtimes

01:19:57
as a service and not Kubernetes as

01:20:00
a service. We'll see.

01:20:03
This is what...

01:20:05
No, go ahead.

01:20:07
No, it's just like, because I think in

01:20:10
this podcast will reach a lot of

01:20:11
developers have been ingrained a

01:20:13
bit of seed of corn to them that

01:20:14
there's something that really

01:20:18
worries me. It is thinking about,

01:20:23
we have this mix, right? Nowadays,

01:20:26
software development perspective

01:20:27
is that rush for the novelty

01:20:30
without the worry for the well

01:20:32
done, the bigger

01:20:34
complexity hidden,

01:20:36
hidden in a way that makes you

01:20:37
think you can do it enough until

01:20:40
the thing really breaks.

01:20:42
And put that together, which is

01:20:44
happening right now. And

01:20:45
companies trying to lower

01:20:47
exploitation systems, we live in

01:20:49
this now-liberated system that

01:20:53
says you have to do

01:20:54
most with Google. And what we're

01:20:56
having, and I think that explains

01:20:57
a lot, is a lot of people that

01:20:59
do not have enough knowledge being

01:21:01
pushed to do a lot of really

01:21:02
complex things that would require

01:21:04
very specialized knowledge. And

01:21:06
then it's been called really new

01:21:09
professions. And I see that

01:21:12
raises so much anxiety about

01:21:14
everyone, because you're not

01:21:17
allowed to do one thing properly,

01:21:19
you just have to know the next

01:21:20
product that you have to feed in

01:21:22
and know

01:21:22
everything, it will break.

01:21:25
Then the question is, how

01:21:27
sustainable is that in a medium

01:21:28
term, not even a long term?

01:21:31
I would like to have more

01:21:33
conversations with developers

01:21:35
about that, about how can we start

01:21:37
pushing back and thinking about

01:21:40
it. And I mentioned that because I

01:21:43
think there's already a

01:21:44
movement pushback with HTMX, for

01:21:50
example. I read, I listened to

01:21:52
this amazing interview with the

01:21:55
developer. I'm sorry, I forgot

01:21:56
that name. Please add it to the links.

01:21:58
Just send me the link

01:22:00
afterwards and we'll add it to the

01:22:01
show notes.

01:22:03
It was this beautiful interview

01:22:04
saying, hey, I, developer,

01:22:07
like at this superpower, the HTML,

01:22:10
to be added, because you mostly

01:22:12
you don't need a complex,

01:22:13
server side render. Most of the

01:22:15
websites do not need it. Let's

01:22:17
start going back and really

01:22:19
understanding what you're doing

01:22:20
and how to build those blocks that

01:22:22
is sustainable. Because on the

01:22:24
top of that, it becomes literally

01:22:29
unsustainable, the amount of data

01:22:30
and the amount of consumption

01:22:32
and energy consumption. It's not

01:22:34
only about, it is about

01:22:38
performance, it

01:22:39
is about doing more

01:22:40
performative software that we're

01:22:42
not trying to hide the complexity,

01:22:44
but we're using the tools,

01:22:46
things that we need to achieve.

01:22:49
Makes sense. I hope it made some

01:22:51
connection. People that are

01:22:52
listening made connection with everything.

01:22:55
I think so. I think a

01:22:58
lot of people try to think more

01:23:01
about that. It's the same as

01:23:04
the work-life balance. It's a big

01:23:06
topic. How can we reduce

01:23:08
complexity in our

01:23:09
own lives? I think

01:23:11
work-life balance is a big thing

01:23:13
trying to reduce that. We're

01:23:15
pretty much out

01:23:16
of time. One really

01:23:18
quick last question. Yeah, 20

01:23:20
minutes. I know. What do you think

01:23:22
is the big next

01:23:24
thing or the current

01:23:26
trend you see and what do you

01:23:28
think is it good? Is it bad?

01:23:33
It's definitely AI. But AI, I hope

01:23:41
I want to believe that

01:23:42
it's an ethical AI.

01:23:44
It is a way, and I will quote

01:23:47
Richard that just gave the story

01:23:49
from the Responsible AI Institute.

01:23:54
Now I've tried and I

01:23:56
keep quoting someone saying, "I

01:23:58
don't want AI to write my text so

01:24:05
I can have more time for

01:24:06
laundry. I want AI to do my

01:24:08
laundry so I can write more texts."

01:24:13
I like that. I like that.

01:24:16
I would find it really in a lot of

01:24:17
talk of people pushing back

01:24:19
saying, "Yo, this

01:24:20
is stuff. It's not

01:24:21
thinking. It's not human. Do not

01:24:23
humanize the system. What are you

01:24:25
doing?" And people thinking

01:24:27
what is the next in the future? Is

01:24:29
that just image recognition? Is

01:24:32
that what it does? Are we really

01:24:34
losing our competition because

01:24:37
we're just trusting this algorithm

01:24:40
that's not actually thinking to do

01:24:42
things for us and then we're not

01:24:44
doing anymore? So I think it's

01:24:47
there to stay. It

01:24:48
is changing a lot

01:24:49
how we're doing. He's an amazing

01:24:51
research assistant. I like to

01:24:54
define like this. And a lot

01:24:56
of things are coming with very

01:24:58
strong. There's regs or the da da

01:25:04
da, augmented

01:25:05
generated. That means

01:25:07
you have to put on top of ChatGPT.

01:25:10
And I think that's here to stay.

01:25:12
And definitely for sure,

01:25:14
the thing is how we are using our

01:25:17
knowledge to drive that. It's

01:25:19
something that will

01:25:21
not help us do our laundries, but

01:25:24
it'll do the laundry

01:25:25
for us so we can write.

01:25:27
I like that. I like that. Yeah, I

01:25:29
think that is very important. We

01:25:31
need to make sure that

01:25:32
we use AI responsibly and that the

01:25:36
training sets are actually very

01:25:38
diversified and make sure that

01:25:40
all groups are included in the

01:25:44
training set so you don't get this

01:25:46
immediate bias. And I think

01:25:47
that will be the hardest thing

01:25:49
because as humans, if we like it

01:25:51
or not, we are inherently biased.

01:25:54
So whatever we're going to train

01:25:56
the system with is probably biased

01:25:58
as well. I think that is hard

01:25:59
to overcome. Yeah, as I said, we

01:26:03
are out of time. Thank you very

01:26:05
much. That was

01:26:05
very, very, very good.

01:26:07
I loved the chat. You probably have

01:26:10
to come back at some point.

01:26:12
I'd love that. I really love talking to

01:26:14
you and definitely hope we can chat more.

01:26:18
All right. And for the

01:26:19
audience, next week,

01:26:22
same time, same place, listen in

01:26:24
again. And I hope to, well, not

01:26:27
here again, but you hear us again.

01:26:30
Thank you very much.

01:26:32
The cloud commute podcast is sponsored by

01:26:34
simplyblock your own elastic

01:26:36
block storage engine for the cloud.

01:26:37
Get higher IOPS and low predictable

01:26:39
latency while bringing down your

01:26:41
total cost of ownership.

01:26:42
www.simplyblock.io