<div dir="ltr">Some <a href="https://thenewstack.io/whats-wrong-with-generative-ai-driven-development-right-now/">reading</a>.<br></div><br><div class="gmail_quote"><div dir="ltr" class="gmail_attr">On Wed, Oct 9, 2024 at 9:11 AM ElPaso via gdal-dev <<a href="mailto:gdal-dev@lists.osgeo.org">gdal-dev@lists.osgeo.org</a>> wrote:<br></div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">Il 09/10/24 00:55, Greg Troxel via gdal-dev ha scritto:<br>
> ElPaso via gdal-dev <<a href="mailto:gdal-dev@lists.osgeo.org" target="_blank">gdal-dev@lists.osgeo.org</a>> writes:<br>
><br>
>> I have read the discussion on lwn and I must say that I am more in<br>
>> line with the debian position.<br>
> My view is that code that comes out of generative AI should be viewed as<br>
> an improper derived work, and lacking adequate provenance/permission to<br>
> be added to an open source project, period. As Even says, we could<br>
> relax this in the future. It's very difficult to go back and remove<br>
> things.<br>
<br>
<br>
This may be true in some circumstances but that's not what I have seen <br>
so far using copilot wih GDAL: most of the times what it does for me <br>
much faster than me is cut-and-paste-replace or autocomplete taking code <br>
from other parts of GDAL or most frequently from other parts of the same <br>
file that I am editing, for example, when I was changing ogrlineref to <br>
use gdalargumentparser, after manually changing the first part of the file<br>
<br>
<a href="https://github.com/OSGeo/gdal/pull/10147/files#diff-b906434b9e6a52aef54e0894ba43a7202290b1af4964cc5d9f1ec8ae7a1c4e15R1271" rel="noreferrer" target="_blank">https://github.com/OSGeo/gdal/pull/10147/files#diff-b906434b9e6a52aef54e0894ba43a7202290b1af4964cc5d9f1ec8ae7a1c4e15R1271</a><br>
<br>
the AI was very useful to autocomplete the other command line switches <br>
(one by one, not all of them), I had to change/edit almost everything <br>
but the scaffolding was there, the source of the autocomplete was <br>
obviously GDAL itself.<br>
<br>
<br>
Where is the copyright issue in this use case?<br>
<br>
<br>
Another situation where I find it useful is when writing tests: most of <br>
the times tests are boring boilerplate code to construct the test data, <br>
for example: the next line here was generate from the comment<br>
<br>
<a href="https://github.com/qgis/QGIS/pull/58986/files#diff-8eebd3707bdc54c42ecd1a2abfca8f623a4c6a8c44c89f344a6762fe95a95059R4682" rel="noreferrer" target="_blank">https://github.com/qgis/QGIS/pull/58986/files#diff-8eebd3707bdc54c42ecd1a2abfca8f623a4c6a8c44c89f344a6762fe95a95059R4682</a> <br>
<br>
<br>
also the checks were automatically generated after I entered the first <br>
couple (they are essentially a copy-paste), the source is QGIS itself.<br>
<br>
<br>
That said I agree that when an AI will be smart enough to be able to be <br>
"creative" (I know, hard to define what it exactly is, but a one-line <br>
cut-paste-replace from the same code base certainly isn't) we will have <br>
a problem to define who/what is the author of the code.<br>
<br>
<br>
Perhaps we could find a way to allow the limited use of AI tools as <br>
autocompleters as long as the source of the generated material is <br>
obviously the code base itself (for instance when using the GDAL API).<br>
<br>
<br>
-- <br>
Alessandro Pasotti<br>
w3: <a href="http://www.itopen.it" rel="noreferrer" target="_blank">www.itopen.it</a><br>
<br>
_______________________________________________<br>
gdal-dev mailing list<br>
<a href="mailto:gdal-dev@lists.osgeo.org" target="_blank">gdal-dev@lists.osgeo.org</a><br>
<a href="https://lists.osgeo.org/mailman/listinfo/gdal-dev" rel="noreferrer" target="_blank">https://lists.osgeo.org/mailman/listinfo/gdal-dev</a><br>
</blockquote></div>