Pono,
Thank you for this presentation of an important current cultural notion and being proactive on addressing the complexity.
As an open-source code project that is driven by a standard reference (SVG) the implications can become suddenly manifold with respect to the Inkscape project that aims to produce software which enables the production of creativity using this standard.
No doubt idiots have already discovered clever ways to produce SVG files using the current landscape of generative AI systems, but inclusion of code is distinct from the making of imagery.
People are the ones who are prominent in the comprehension, creation, and correction of the code that drives this unique project.
Any and all usage of AI code should be disclosed. The vetting of its veracity takes careful human consideration. I am not utterly opposed to this form of cheating, in certain circumstances that may be technically challenging given our collective abilities, but I consider Inkscape a work of Art composed by people who devote their energy toward honestly understanding how it does all that it is capable of performing. We love having an activity that is genuine.
Disclosure. Validation. Acceptance. But do not destroy the natural creativity that erupts from human creativity when we want it to do something specific without appreciating the ability for past, present, or future contributors to glimpse and appreciate.
Coding is a skill. Like any language. It is easy to tell if one is capable of “speaking” it well or not - principally due to the ability to take time to interrogate one on why it was done in such a fashion. We want solid code, that works fast, but does not become so utterly esoteric that we are unable to follow the flow of thought that performs the desired activity.
Best,
Nathan (NPJ2000)
Sent from my iPhone
On Oct 17, 2024, at 2:10 PM, Daniel Pono Takamori pono@sfconservancy.org wrote:
Heya Inkscape Developers,
Following up on an issue that came up a few months ago about code and other uses of LLMs; we at Software Freedom Conservancy have been hearing from member projects that people are asking whether they can submit code to these projects that was generated by large language models (LLMs), sometimes called AI coding assistants. We'd like to better understand the nature of these requests, how much LLM-generated code has been offered to these projects, and generally how you're feeling about the situation.
To do this, we will be holding three different sessions, where you are welcome to join and share your thoughts and any details so far on these types of contributions. The sessions will all be run in the same way with the same topic - we are running three in order to hopefully allow as many people from around the world to attend as possible.
Our Director of Compliance, Denver Gingerich, will be hosting some sessions at the times below for all our member projects to discuss the issue.
date -d "2024-10-29 20:00 UTC" date -d "2024-11-01 14:00 UTC" date -d "2024-11-06 04:00 UTC"
We'll be meeting in this room for the sessions:
https://bbb.sfconservancy.org/b/oss-nnj-obi-jea
We are excited to see some of you there and are looking forward to drafting appropriate recommendations based on what we hear!
Thanks, -Pono _______________________________________________ Inkscape Devel mailing list -- inkscape-devel@lists.inkscape.org To unsubscribe send an email to inkscape-devel-leave@lists.inkscape.org