WEBVTT

00:00.000 --> 00:11.040
My talk is rather different from the talks that have been given a so far in the

00:11.040 --> 00:12.040
development.

00:12.040 --> 00:16.880
It's not about the technology, but reflects a bit on the kind of technologies we've

00:16.880 --> 00:21.480
used and the value that are behind the technologies.

00:21.480 --> 00:32.480
It has also considered something, when you put it in your pocket, I didn't switch off anything.

00:32.480 --> 00:36.480
No, it looks like it's still on, isn't it?

00:36.480 --> 00:37.480
Yeah.

00:37.480 --> 00:38.480
Yeah.

00:38.480 --> 00:39.480
No.

00:39.480 --> 00:44.040
It's like it's off, but nothing's unique.

00:44.040 --> 00:46.040
It's inside.

00:46.040 --> 00:47.040
Wait a minute.

00:47.040 --> 00:48.040
Again?

00:48.040 --> 00:51.040
There it is.

00:51.040 --> 00:52.040
It is on.

00:52.040 --> 00:53.040
It is on.

00:53.040 --> 00:54.040
I'm sorry.

00:54.040 --> 00:57.040
I should do you want to.

00:57.040 --> 00:58.040
Okay.

00:58.040 --> 00:59.040
Better?

00:59.040 --> 01:00.040
Sorry.

01:00.040 --> 01:01.040
Great.

01:01.040 --> 01:02.040
Good.

01:02.040 --> 01:06.040
As I was saying, this is not about the technology.

01:06.040 --> 01:11.640
It's about the way we select technologies and the fact that technologies are not

01:11.640 --> 01:26.640
neutral and that are values.

01:26.640 --> 01:30.640
It's behind the visor software development work.

01:30.640 --> 01:32.640
Why is reflection?

01:32.640 --> 01:36.640
I am personally growing incredibly uncomfortable.

01:36.640 --> 01:45.640
We've added to you that there is currently a research to focus only on technologies.

01:45.640 --> 01:51.640
I think that it gives quite clearly that everyone thinks about the technology of the future

01:51.640 --> 01:53.640
of things, about the elements.

01:53.640 --> 02:01.640
I'm not convinced that visor should be the way, especially because the research becomes

02:01.640 --> 02:08.640
richer when there are multiple perspectives and if only one technology is to take the lead,

02:08.640 --> 02:14.640
visor is somehow reducing the reality of opportunities.

02:14.640 --> 02:21.640
I think that also there is a considerable difference between the way we think as human and the way

02:21.640 --> 02:28.640
LLM's work in token base analysis and the visor is not the way human mind is working.

02:28.640 --> 02:32.640
This is the reason why I wanted to propose visor reflection.

02:32.640 --> 02:34.640
It has a personal reflection.

02:34.640 --> 02:39.640
I think that it makes sense to say a couple of words about myself.

02:39.640 --> 02:42.640
My name is Judith Parolini.

02:42.640 --> 02:45.640
I work for a financial release to the museum.

02:45.640 --> 02:52.640
I work on a project on accessibility of data practices and infrastructures in biodiversity science.

02:52.640 --> 02:55.640
That is sponsored by the German Research Foundation.

02:55.640 --> 03:01.640
Before working as a data scientist, I used to be an historian of science and technology again working

03:01.640 --> 03:06.640
in research and before that I was a freelance science writer.

03:06.640 --> 03:12.640
I've been going through different jobs in my career.

03:12.640 --> 03:18.640
Floss has been part of my research work because as a researcher I started to code.

03:18.640 --> 03:26.640
As a data scientist, the Python library for data science are the main tools in my job.

03:26.640 --> 03:32.640
I'm grateful that we are open source software because the visor is allowed me to learn

03:32.640 --> 03:35.640
to use the tools without worrying about licenses.

03:35.640 --> 03:41.640
But also because the visor is allowed me to teach other people to use the visor.

03:41.640 --> 03:44.640
In research but also outside of research.

03:44.640 --> 03:50.640
I volunteer for a couple of courses as an instructor at the very school for digital integration in Berlin.

03:50.640 --> 03:53.640
That is a project for refugees.

03:53.640 --> 03:59.640
I have been so positively impressed by open source.

03:59.640 --> 04:06.640
In my work as a researcher, a couple of years ago I decided that I wanted to give back something to the community.

04:06.640 --> 04:10.640
And so I started the volunteering for open source projects.

04:10.640 --> 04:17.640
In 2024 I volunteered for a few months for the psyche-teamage project that is a Python-based project.

04:17.640 --> 04:24.640
And last year I started volunteering for another project, a good docs project that doesn't produce code,

04:24.640 --> 04:28.640
but generates templates for software documentation.

04:29.640 --> 04:35.640
Both experiences not only allow me to give back something to the community of open source,

04:35.640 --> 04:41.640
but also to learn new skills that I can use in my professional development.

04:41.640 --> 04:43.640
Open source is successful.

04:43.640 --> 04:48.640
As a data scientist, I see it all the time in data technologies,

04:48.640 --> 04:51.640
a container technology is clouded technologies.

04:52.640 --> 04:56.640
But the recipe of success is not to be found in money.

04:56.640 --> 05:00.640
Of course, money has value, but it's not to be driving force.

05:00.640 --> 05:06.640
We drive force behind the flaws communities are the humans that are within.

05:06.640 --> 05:11.640
We work in this community that volunteers and contribute to this community.

05:11.640 --> 05:14.640
And this is of course not just my idea.

05:14.640 --> 05:19.640
Carafogala has written a very interesting book producing open source software.

05:19.640 --> 05:28.640
But the deals are in detail with the human dimension of open source and the kind of skills that you need to have

05:28.640 --> 05:32.640
if you want to be a developer in open source.

05:32.640 --> 05:37.640
Among the points that the focus makes, there are a few that really resonates with me,

05:37.640 --> 05:42.640
with my experience of open source, and with what I think it can be helpful for research.

05:43.640 --> 05:48.640
These are discussing in public building inclusive community,

05:48.640 --> 05:52.640
and setting up effective governance structures.

05:52.640 --> 05:56.640
This discussing in public has been mentioned in the open ones,

05:56.640 --> 06:00.640
even today, because this is really the whole market of open source.

06:00.640 --> 06:04.640
If you want to carry on an open source project,

06:04.640 --> 06:09.640
you need to make your decision in public, you need to set your roadmap in public,

06:09.640 --> 06:14.640
because this allows you to gather the expertise of all the members of your communities,

06:14.640 --> 06:17.640
and not just of your maintainers.

06:17.640 --> 06:20.640
Of course, if you want to contribute,

06:20.640 --> 06:22.640
then you have to learn how to do it.

06:22.640 --> 06:24.640
You have to learn to make your point,

06:24.640 --> 06:26.640
but you also have to learn to listen to other people,

06:26.640 --> 06:30.640
to make compromises in the best interest of a project,

06:30.640 --> 06:33.640
and you have to be cautious at all times.

06:33.640 --> 06:38.640
With the skills that will open source developers need to learn as part of their job.

06:38.640 --> 06:43.640
And I think that the business way in which knowledge is really produced.

06:43.640 --> 06:47.640
Business is also part of building inclusive communities,

06:47.640 --> 06:52.640
because in my experience as long as you abide

06:52.640 --> 06:56.640
to the code of conduct, you have read the contributors guidelines,

06:56.640 --> 06:59.640
and you have an interest in the project that you have,

06:59.640 --> 07:03.640
the most welcome person in the world in an open source project.

07:03.640 --> 07:06.640
And I find incredibly enriching to work together

07:06.640 --> 07:08.640
with people who live in other continents,

07:08.640 --> 07:11.640
with people who do completely different things for a living,

07:11.640 --> 07:16.640
but we all are able to give a value to the project

07:16.640 --> 07:18.640
and to contribute to the project.

07:18.640 --> 07:20.640
And across projects, I think,

07:20.640 --> 07:23.640
an example of how you can get more contributors,

07:23.640 --> 07:24.640
that are your real value.

07:24.640 --> 07:26.640
You organize an archetyne,

07:26.640 --> 07:28.640
you organize a sprint at the conference,

07:28.640 --> 07:31.640
you try to talk to people at a different events,

07:32.640 --> 07:35.640
because the richness is not in the money,

07:35.640 --> 07:40.640
you have a richness in the people you can somehow bring to your project

07:40.640 --> 07:44.640
and invade ideas that these people can contribute to your project.

07:44.640 --> 07:47.640
But of course, a visa doesn't come at the cost,

07:47.640 --> 07:51.640
and the cost is to set up an effective governance structure.

07:51.640 --> 07:54.640
But it is also important to open source project

07:54.640 --> 07:57.640
and understand the value of intellectual property

07:57.640 --> 07:58.640
and what concerns the software,

07:58.640 --> 08:02.640
but I understand the value of finding funding strategies

08:02.640 --> 08:04.640
that are not the core.

08:04.640 --> 08:08.640
For finding strategies of venture capital funding strategy,

08:08.640 --> 08:13.640
but the way in which we can make things go on,

08:13.640 --> 08:18.640
despite being a community of very different people.

08:18.640 --> 08:22.640
And I think that these things can be valuable in research.

08:22.640 --> 08:26.640
Some of the problems of research have emerged in the talks

08:26.640 --> 08:28.640
this morning.

08:28.640 --> 08:32.640
I would like to just mention a few of the problems with the performance metrics

08:32.640 --> 08:36.640
that we have set up that are typically performance metrics

08:36.640 --> 08:38.640
that only focus on publications,

08:38.640 --> 08:42.640
and that privilege in publications quantity of equality.

08:42.640 --> 08:45.640
This is of course a penalizing and credibly research software

08:45.640 --> 08:47.640
because the research software needs,

08:47.640 --> 08:49.640
and instead a long term thinking,

08:49.640 --> 08:51.640
including also maintenance.

08:51.640 --> 08:54.640
And also because of course,

08:54.640 --> 08:56.640
as a little scientist you know,

08:56.640 --> 08:58.640
when you make metrics,

08:58.640 --> 09:02.640
then there is someone who finds a way to gain this metric.

09:02.640 --> 09:06.640
And so we have authors who do things

09:06.640 --> 09:09.640
and not correct in terms of citations,

09:09.640 --> 09:12.640
authors who publish paper every few days,

09:12.640 --> 09:15.640
and this is not helping in terms of promoting quality.

09:15.640 --> 09:18.640
Academic publishers are also not helping very much

09:18.640 --> 09:21.640
because they have transformed the idea of open access

09:22.640 --> 09:25.640
as a way to democratize access to knowledge

09:25.640 --> 09:27.640
in a lucrative business model

09:27.640 --> 09:29.640
that just wants to publish more

09:29.640 --> 09:32.640
without really much looking into the quality.

09:32.640 --> 09:33.640
Otherwise we could,

09:33.640 --> 09:36.640
and if the quality instead of where an issue

09:36.640 --> 09:38.640
we could, we would not be able to explain

09:38.640 --> 09:40.640
how many publications now,

09:40.640 --> 09:45.640
even basic elements of issues

09:45.640 --> 09:48.640
like recurrent and correct references.

09:48.640 --> 09:51.640
To add to that very applicability crisis,

09:51.640 --> 09:53.640
that was also discussed,

09:53.640 --> 09:56.640
I just put a remark to a book

09:56.640 --> 09:59.640
that is about very applicability crisis

09:59.640 --> 10:01.640
in the biomedical science.

10:01.640 --> 10:03.640
I think that it makes for an interesting reading,

10:03.640 --> 10:05.640
especially if you want to read

10:05.640 --> 10:07.640
also what people working in get fulfilled

10:07.640 --> 10:10.640
as a possible solution to it.

10:10.640 --> 10:13.640
The self-correcting mechanism in science

10:13.640 --> 10:15.640
are not working very well,

10:15.640 --> 10:17.640
of course, even in publications

10:17.640 --> 10:19.640
that are the output of science

10:19.640 --> 10:21.640
in the current world.

10:21.640 --> 10:24.640
It takes years before you can correct

10:24.640 --> 10:29.640
or retract a paper that has even serious issues.

10:29.640 --> 10:31.640
Then depending on where you live,

10:31.640 --> 10:34.640
politics also is limiting freedom

10:34.640 --> 10:36.640
because politicians are now deciding

10:36.640 --> 10:40.640
what makes for good research and for better research.

10:40.640 --> 10:42.640
It is good to continue,

10:42.640 --> 10:45.640
but I think that one of the problems,

10:45.640 --> 10:47.640
I don't believe that LLM's

10:47.640 --> 10:49.640
or generative AI in general

10:49.640 --> 10:52.640
are a solution to the issues of research.

10:52.640 --> 10:54.640
I think that they have just added

10:54.640 --> 10:56.640
to the problems that were already there.

10:56.640 --> 11:00.640
Now they are used to fabricate research data

11:00.640 --> 11:03.640
and images and also text of publications

11:03.640 --> 11:05.640
without a proper oversight.

11:05.640 --> 11:08.640
They are promoted as a replacement

11:08.640 --> 11:11.640
for human expertise in every sense,

11:11.640 --> 11:13.640
even the expertise of peer review,

11:13.640 --> 11:17.640
and I am really not sure how the summary

11:17.640 --> 11:19.640
or the so-called analysis

11:19.640 --> 11:21.640
that an LLM can produce

11:21.640 --> 11:23.640
can be considered an equivalent

11:23.640 --> 11:26.640
for the competence and the expertise of human reviewers.

11:26.640 --> 11:29.640
But some research institutions are

11:29.640 --> 11:30.640
particularly keen on them,

11:30.640 --> 11:32.640
and for instance,

11:32.640 --> 11:34.640
were was a discussion about the decision

11:34.640 --> 11:36.640
of the University of California

11:36.640 --> 11:39.640
to make an agreement with Open AI

11:39.640 --> 11:41.640
for making a generative AI

11:41.640 --> 11:44.640
to a backbone of their research and education.

11:44.640 --> 11:47.640
Some researchers at the University of California

11:47.640 --> 11:51.640
are also not very much in agreement with this idea.

11:51.640 --> 11:53.640
And the last but not least,

11:53.640 --> 11:56.640
in the case of research software,

11:56.640 --> 11:59.640
AI coding assistants are basically

11:59.640 --> 12:01.640
contributing to this killer

12:01.640 --> 12:03.640
rather than to obscure people,

12:03.640 --> 12:05.640
because there was a discussion

12:05.640 --> 12:07.640
now even the historian can code,

12:07.640 --> 12:09.640
because it can use the LLM.

12:09.640 --> 12:11.640
But what kind of value

12:11.640 --> 12:14.640
is very in a person who gets some lines of code

12:14.640 --> 12:15.640
from an LLM,

12:15.640 --> 12:17.640
but is not able to understand

12:17.640 --> 12:19.640
fully what this code does,

12:19.640 --> 12:21.640
and also does not understand

12:21.640 --> 12:23.640
the key principles of software engineering

12:23.640 --> 12:25.640
like that you have to test your code

12:25.640 --> 12:28.640
that your code is consuming resources.

12:28.640 --> 12:30.640
So I see this as a very,

12:30.640 --> 12:31.640
very negative thing,

12:31.640 --> 12:33.640
and I think that this LLM

12:33.640 --> 12:36.640
is a recipe for a permanent technical depth

12:36.640 --> 12:39.640
and not for enriching the resources software.

12:39.640 --> 12:42.640
With this reason why,

12:42.640 --> 12:44.640
and definitely is a thought,

12:44.640 --> 12:47.640
with what I think is a researcher.

12:47.640 --> 12:48.640
For me, the WING research,

12:48.640 --> 12:51.640
including creating research software,

12:51.640 --> 12:54.640
is a way to channel a human curiosity,

12:54.640 --> 12:56.640
creativity, intelligence,

12:56.640 --> 12:59.640
and to go into group competence and expertise.

12:59.640 --> 13:02.640
And I honestly don't see anything good.

13:02.640 --> 13:05.640
Of it in a technology driven world,

13:05.640 --> 13:07.640
in the technology driven world,

13:07.640 --> 13:10.640
but someone he is expecting.

13:10.640 --> 13:12.640
And instead I would like to see,

13:12.640 --> 13:13.640
for instance,

13:13.640 --> 13:16.640
more opportunities for public discussion within research,

13:16.640 --> 13:19.640
because it's clear that the traditional peer review system

13:19.640 --> 13:22.640
is not keeping up with what is happening.

13:22.640 --> 13:24.640
But on the other hand,

13:24.640 --> 13:26.640
we also see that the research community

13:26.640 --> 13:29.640
is a key non-contributed to public discussions.

13:30.640 --> 13:33.640
Post-publication peer review platforms,

13:33.640 --> 13:34.640
like public peer,

13:34.640 --> 13:36.640
but constantly have people

13:36.640 --> 13:38.640
contributing very insights.

13:38.640 --> 13:40.640
A very example of the fact that the community

13:40.640 --> 13:42.640
is a very much alive.

13:42.640 --> 13:45.640
But the current system of peer review

13:45.640 --> 13:48.640
doesn't an knowledge of this contribution.

13:48.640 --> 13:50.640
It's just something between the headytore

13:50.640 --> 13:53.640
and a couple of peer reviewers,

13:53.640 --> 13:55.640
and visions are really not helping the community,

13:55.640 --> 13:58.640
but should be more in terms of public discussion.

13:58.640 --> 14:01.640
Similarly, I would like to see more inclusion

14:01.640 --> 14:03.640
within a research.

14:03.640 --> 14:05.640
Because nowadays,

14:05.640 --> 14:08.640
research networks are basically networks based

14:08.640 --> 14:09.640
on geography,

14:09.640 --> 14:12.640
on the funders of your research.

14:12.640 --> 14:16.640
Sometimes you also have networks based

14:16.640 --> 14:21.640
on institutions that belong to similar groups.

14:21.640 --> 14:23.640
But at the end of the day,

14:23.640 --> 14:25.640
research is made by people

14:25.640 --> 14:28.640
who are a communities of practice,

14:28.640 --> 14:30.640
that is a people who want to do something.

14:30.640 --> 14:31.640
And so,

14:31.640 --> 14:32.640
the question you should ask

14:32.640 --> 14:35.640
to people who do research is not who you are,

14:35.640 --> 14:36.640
but what you want to do.

14:36.640 --> 14:37.640
Visits by way,

14:37.640 --> 14:39.640
flows communities work,

14:39.640 --> 14:40.640
what you want to do,

14:40.640 --> 14:41.640
not who you are,

14:41.640 --> 14:43.640
and I think that business something

14:43.640 --> 14:45.640
that research organization

14:45.640 --> 14:48.640
should start to think about.

14:48.640 --> 14:50.640
And last but not least,

14:50.640 --> 14:51.640
the vagabondance problems.

14:51.640 --> 14:54.640
These are very clear in terms of research software,

14:54.640 --> 14:56.640
but in general,

14:56.640 --> 14:58.640
in many aspects of research,

14:58.640 --> 15:00.640
there is a need to understand better

15:00.640 --> 15:03.640
intellectual property rights.

15:03.640 --> 15:05.640
That, for instance,

15:05.640 --> 15:08.640
are a huge issue in terms of using LLM,

15:08.640 --> 15:09.640
for generating code,

15:09.640 --> 15:11.640
but also for generating publications.

15:11.640 --> 15:14.640
And also to understand

15:14.640 --> 15:15.640
for instance financing,

15:15.640 --> 15:17.640
in ways of that are not only dependent

15:17.640 --> 15:21.640
on the funding provided by

15:21.640 --> 15:24.640
the grant giving institutions.

15:24.640 --> 15:26.640
I didn't cross

15:26.640 --> 15:28.640
as a linking influencer.

15:28.640 --> 15:31.640
This was really not what I wanted to do.

15:31.640 --> 15:33.640
I just wanted to raise the point

15:33.640 --> 15:36.640
that we should really

15:36.640 --> 15:39.640
wait one minute within

15:39.640 --> 15:41.640
with AI technologies,

15:41.640 --> 15:43.640
generative AI technologies,

15:43.640 --> 15:45.640
and because we never really ask

15:45.640 --> 15:46.640
this question,

15:46.640 --> 15:47.640
should we?

15:47.640 --> 15:48.640
Of course,

15:48.640 --> 15:50.640
we can do that,

15:50.640 --> 15:51.640
and we do that.

15:51.640 --> 15:53.640
And I think that the

15:53.640 --> 15:54.640
VISA website,

15:54.640 --> 15:56.640
the AI Darwin Award,

15:56.640 --> 15:59.640
or some interesting example of

15:59.640 --> 16:02.640
what can happen if you don't ask

16:02.640 --> 16:03.640
a VISA question,

16:03.640 --> 16:04.640
should we,

16:04.640 --> 16:06.640
but you only do what you can,

16:06.640 --> 16:09.640
because I think that somehow

16:09.640 --> 16:11.640
put a different perspective on

16:11.640 --> 16:12.640
the thing.

16:12.640 --> 16:14.640
And aside from the

16:14.640 --> 16:16.640
questionable quality of AI tools,

16:16.640 --> 16:17.640
I don't think that we are putting

16:17.640 --> 16:19.640
any value into research.

16:19.640 --> 16:21.640
I think that VISA technology,

16:21.640 --> 16:22.640
I mean,

16:22.640 --> 16:24.640
one should also know that VISA technology

16:24.640 --> 16:26.640
have huge environmental impact.

16:26.640 --> 16:28.640
We are floating the

16:28.640 --> 16:30.640
world of data centers.

16:30.640 --> 16:32.640
I'm not sure how much energy

16:32.640 --> 16:35.640
we will need to keep up with this.

16:35.640 --> 16:37.640
VISA tools are also based

16:37.640 --> 16:39.640
on infringement

16:39.640 --> 16:41.640
a little property rights of

16:41.640 --> 16:43.640
creators, authors, designers,

16:43.640 --> 16:44.640
developers.

16:44.640 --> 16:46.640
Is this what we want?

16:46.640 --> 16:48.640
And the cost of VISA tools,

16:49.640 --> 16:50.640
now there is a,

16:50.640 --> 16:52.640
the user of VISA tools are not

16:52.640 --> 16:54.640
paying for the cost of VISA tools.

16:54.640 --> 16:55.640
They are paying for something

16:55.640 --> 16:57.640
that is highly subsidized by

16:57.640 --> 16:58.640
venture capital.

16:58.640 --> 17:00.640
What will happen when they have

17:00.640 --> 17:02.640
to pay the true,

17:02.640 --> 17:03.640
the true price?

17:03.640 --> 17:04.640
In a word,

17:04.640 --> 17:06.640
a generative AI technologies are

17:06.640 --> 17:07.640
to me,

17:07.640 --> 17:10.640
an example of unsustainable technologies.

17:10.640 --> 17:11.640
And as a researcher,

17:11.640 --> 17:12.640
and the human being,

17:12.640 --> 17:14.640
I want the writer to say,

17:14.640 --> 17:15.640
okay,

17:15.640 --> 17:16.640
I'm not interested in using them.

17:16.640 --> 17:18.640
I don't want to use them.

17:18.640 --> 17:20.640
VISA's reason why I wanted to

17:20.640 --> 17:21.640
bring a VISA discussion.

17:21.640 --> 17:22.640
You're at first

17:22.640 --> 17:23.640
them,

17:23.640 --> 17:25.640
because if you had been yesterday

17:25.640 --> 17:27.640
at the main track,

17:27.640 --> 17:29.640
VISA films came up in

17:29.640 --> 17:31.640
more than one talk.

17:31.640 --> 17:32.640
And the

17:32.640 --> 17:34.640
VISA community has been

17:34.640 --> 17:35.640
a really a community that has

17:35.640 --> 17:37.640
spoken up against the VISA things.

17:37.640 --> 17:39.640
maintainers of open source

17:39.640 --> 17:42.640
projects have banned AI tools

17:42.640 --> 17:45.640
from a poor request,

17:45.640 --> 17:47.640
because I've simply told people,

17:47.640 --> 17:48.640
okay,

17:48.640 --> 17:49.640
no,

17:49.640 --> 17:51.640
VISA is a huge risk for

17:51.640 --> 17:53.640
ventilator property of a project.

17:53.640 --> 17:55.640
VISA is also a huge risk for

17:55.640 --> 17:56.640
hazard,

17:56.640 --> 17:58.640
because we are burning out,

17:58.640 --> 18:00.640
keeping up with

18:00.640 --> 18:02.640
too many requests that don't

18:02.640 --> 18:04.640
really had any value

18:04.640 --> 18:06.640
to the things that we do.

18:06.640 --> 18:08.640
And on GitHub,

18:08.640 --> 18:10.640
I must say that GitHub has become

18:10.640 --> 18:12.640
not only at least where people

18:12.640 --> 18:14.640
discuss technology,

18:14.640 --> 18:16.640
and technical choices,

18:16.640 --> 18:18.640
but also where people discuss

18:18.640 --> 18:20.640
the use of these tools.

18:20.640 --> 18:22.640
The link to the discussion

18:22.640 --> 18:24.640
on Reddit here is about a guy

18:24.640 --> 18:28.640
that brought together several

18:28.640 --> 18:32.640
GitHub issues of one repository

18:32.640 --> 18:34.640
that was created by Microsoft,

18:34.640 --> 18:36.640
but it is an open source repository.

18:36.640 --> 18:38.640
And on VISA repositories,

18:38.640 --> 18:40.640
Microsoft use AI agents,

18:40.640 --> 18:42.640
and the open source developers started

18:42.640 --> 18:43.640
to comment in.

18:43.640 --> 18:45.640
And the comments that there is

18:45.640 --> 18:47.640
were about the security,

18:47.640 --> 18:48.640
the quality, the licensing,

18:48.640 --> 18:50.640
and the performance of the code

18:50.640 --> 18:51.640
written by AI,

18:51.640 --> 18:53.640
and we were not very impressed about

18:53.640 --> 18:54.640
the results,

18:54.640 --> 18:56.640
but we were also concerned about

18:56.640 --> 18:58.640
the fact that the expression is

18:58.640 --> 19:00.640
not mine is one of the developers.

19:00.640 --> 19:01.640
We are cooking the planet

19:01.640 --> 19:04.640
for getting a very poor result.

19:04.640 --> 19:06.640
And so I can only say

19:06.640 --> 19:08.640
good also to the flaws community,

19:08.640 --> 19:11.640
because continues to believe in human values

19:11.640 --> 19:12.640
of people's skills,

19:12.640 --> 19:14.640
and I sincerely hope

19:14.640 --> 19:16.640
that the research community

19:16.640 --> 19:17.640
will also be an equally

19:17.640 --> 19:19.640
stronger advocate for VISA values

19:19.640 --> 19:21.640
and VISA skills.

19:42.640 --> 19:45.640
I agree with everything.

19:45.640 --> 19:48.640
But as researchers were often already

19:48.640 --> 19:50.640
short on free time,

19:50.640 --> 19:52.640
a lot of things you're recommending.

19:52.640 --> 19:53.640
We require us to put more time

19:53.640 --> 19:55.640
into whether peer reviews

19:55.640 --> 19:57.640
or working on these open source projects,

19:57.640 --> 19:59.640
or already working on others,

19:59.640 --> 20:00.640
and only do things.

20:00.640 --> 20:02.640
So where do we prioritize our time

20:02.640 --> 20:06.640
or how do we get more time out of ourselves?

20:06.640 --> 20:10.640
The question is about the time pressure

20:10.640 --> 20:11.640
that is on researcher,

20:11.640 --> 20:14.640
and the fact that by not adopting

20:14.640 --> 20:15.640
for instance, LLMs,

20:15.640 --> 20:19.640
we need to do a different prioritization

20:19.640 --> 20:21.640
of our time.

20:21.640 --> 20:24.640
I would like to answer you

20:24.640 --> 20:26.640
with a recent article

20:26.640 --> 20:27.640
that I read,

20:27.640 --> 20:29.640
because it was in the retraction

20:29.640 --> 20:32.640
which we can read.

20:32.640 --> 20:34.640
It was a researcher

20:34.640 --> 20:36.640
who had used the chat to P.T.

20:36.640 --> 20:38.640
for the past two years

20:38.640 --> 20:41.640
for doing everything from a researcher

20:41.640 --> 20:43.640
to teaching the students and so on.

20:43.640 --> 20:45.640
This person at some point

20:45.640 --> 20:48.640
and click the option of

20:48.640 --> 20:49.640
keep my data safe.

20:49.640 --> 20:51.640
What happened was that the chat

20:51.640 --> 20:53.640
to P.T. cancelled everything

20:53.640 --> 20:56.640
because in order to avoid the transfer

20:56.640 --> 20:57.640
of data,

20:57.640 --> 20:58.640
they basically can't do everything

20:58.640 --> 21:00.640
and in order to abide

21:00.640 --> 21:01.640
to data protection,

21:01.640 --> 21:03.640
low-visit is not recoverable.

21:03.640 --> 21:05.640
So this person finds

21:06.640 --> 21:08.640
himself without all of the data

21:08.640 --> 21:10.640
that he had collected

21:10.640 --> 21:12.640
for time saving purposes

21:12.640 --> 21:14.640
in the past two years

21:14.640 --> 21:16.640
and wonders whether

21:16.640 --> 21:18.640
he really had got the point

21:18.640 --> 21:21.640
in up to the end.

21:21.640 --> 21:24.640
So I would like to reply

21:24.640 --> 21:27.640
to you with another question.

21:27.640 --> 21:30.640
What do we really value?

21:30.640 --> 21:32.640
I can understand

21:32.640 --> 21:35.640
that the marketing that has been done

21:35.640 --> 21:37.640
and also I mean

21:37.640 --> 21:39.640
I mentioned explicitly that

21:39.640 --> 21:41.640
there are organizations

21:41.640 --> 21:42.640
or research organizations

21:42.640 --> 21:44.640
that have made agreement

21:44.640 --> 21:45.640
with open AI.

21:45.640 --> 21:48.640
There are several US organizations

21:48.640 --> 21:50.640
that have decided that they want

21:50.640 --> 21:52.640
to work in this way.

21:52.640 --> 21:54.640
This is a choice.

21:54.640 --> 21:56.640
I personally feel that

21:56.640 --> 21:57.640
as a researcher,

21:57.640 --> 22:00.640
is not compatible

22:00.640 --> 22:02.640
with my principal.

22:02.640 --> 22:04.640
So even though there is a price

22:04.640 --> 22:06.640
to pay,

22:06.640 --> 22:07.640
I think that I am just

22:07.640 --> 22:10.640
to all to forget my principal.

22:10.640 --> 22:12.640
Somehow carried myself

22:12.640 --> 22:14.640
for quite a good

22:14.640 --> 22:15.640
many years

22:15.640 --> 22:17.640
and I think that I just

22:17.640 --> 22:18.640
prefer to live

22:18.640 --> 22:19.640
with my principal and to live.

22:19.640 --> 22:20.640
But you are right.

22:20.640 --> 22:21.640
I mean,

22:21.640 --> 22:22.640
I think that peer review

22:22.640 --> 22:23.640
should be paid.

22:23.640 --> 22:24.640
I mean,

22:24.640 --> 22:25.640
if you get someone

22:25.640 --> 22:26.640
to review your paper,

22:26.640 --> 22:27.640
you cannot respect

22:27.640 --> 22:28.640
this person to spend

22:28.640 --> 22:30.640
hours on paid.

22:30.640 --> 22:32.640
But this is a long discussion

22:32.640 --> 22:33.640
that has to do with

22:33.640 --> 22:35.640
economics of academia.

22:35.640 --> 22:36.640
But I think that also

22:36.640 --> 22:38.640
if you renounce

22:38.640 --> 22:41.640
the

22:41.640 --> 22:42.640
these large language

22:42.640 --> 22:44.640
models are for instance

22:44.640 --> 22:45.640
infringing into

22:45.640 --> 22:47.640
the intellectual property rights.

22:47.640 --> 22:48.640
And instead,

22:48.640 --> 22:49.640
I think that researchers

22:49.640 --> 22:50.640
should ask themselves,

22:50.640 --> 22:51.640
okay,

22:51.640 --> 22:52.640
how can I

22:52.640 --> 22:53.640
monetize

22:53.640 --> 22:54.640
that?

22:54.640 --> 22:55.640
Not in a bad sense,

22:55.640 --> 22:56.640
but make a living

22:56.640 --> 22:57.640
like the

22:57.640 --> 22:58.640
philosophers,

22:58.640 --> 22:59.640
only philosopher,

22:59.640 --> 23:00.640
developer.

23:00.640 --> 23:01.640
How can I make a living

23:01.640 --> 23:02.640
about my idea?

23:02.640 --> 23:03.640
What I do?

23:03.640 --> 23:04.640
So that you can also

23:04.640 --> 23:05.640
get away from

23:05.640 --> 23:06.640
the system of short

23:06.640 --> 23:07.640
grants of it

23:07.640 --> 23:09.640
is the way in which academia

23:09.640 --> 23:10.640
live.

23:10.640 --> 23:11.640
And just quickly

23:11.640 --> 23:12.640
under a

23:12.640 --> 23:13.640
steel

23:13.640 --> 23:14.640
years after

23:14.640 --> 23:15.640
point colleagues

23:15.640 --> 23:16.640
to the slow science

23:16.640 --> 23:17.640
manifesto.

23:17.640 --> 23:18.640
2010

23:18.640 --> 23:20.640
and still a very

23:21.640 --> 23:22.640
science

23:22.640 --> 23:24.640
that's basically what you

23:24.640 --> 23:25.640
say.

23:25.640 --> 23:26.640
Thanks.