Hi, I don't know whether or not ChatGPT is trained on the source code of SageMath, but one of the biggest publicly available training sets of code is described here: https://arxiv.org/abs/2211.15533 In that training set, they explicitly remove any GPL'd code (e.g., SageMath): "Permissive license dataset: We develop a dataset of source code with only permissive licenses, i.e., with minimal restrictions on how the software can be copied, modified, and redistributed. We first provide the list of licenses which we classified as permissive in Appendix A. Note that we intentionally exclude copyleft licenses like GPL, as this community has strongly expressed the concern of machine learning models and inferred outputs violating the terms of their licenses Kuhn (2022)."
As things unfold in the years to come with people trying to use LLM's in the context of mathematical software, our academic community's choice of license could explain why LLM's output code that's much more like Sympy (say) than SageMath's, and perhaps why LLM's are not as good at using Sage. William P.S.In case anybody is curious, Sage was originally GPL'd because it is a derived work of Pari, and Pari is GPL'd. I made the choice to use Pari heavily in the implementation of Sage, rather than starting from scratch. I asked Henri Cohen why he GPL'd Pari and he told me that Richard Stallman personally "strongly encouraged" him to do so, because Pari depends on GNU Readline, and GNU Readline is GPL'd. Despite anything mentioned above, the GPL still seems like the right license for Sage, given that all the other competitors (Magma, Mathematica, etc.) are much more restrictive in their licenses. On Fri, Mar 24, 2023 at 9:13 PM Kwankyu Lee <ekwan...@gmail.com> wrote: > > > ... The syntactical correction Curve(x^3+y^3-x-y+1,P) still doesn't work > because we have an affine equation and are specifying an unrelated projective > space. > > > That is the most serious defect in the answer. I missed that. > > I think the conclusion should be that ChatGPT pretends to be a SageMath > expert with great confidence but .. gasp ... isn't! > > > and also I was not careful enough to claim that. > > > -- > You received this message because you are subscribed to the Google Groups > "sage-devel" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to sage-devel+unsubscr...@googlegroups.com. > To view this discussion on the web visit > https://groups.google.com/d/msgid/sage-devel/f49e1b3e-396b-4483-93d1-1af4d721476bn%40googlegroups.com. -- William (http://wstein.org) -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/sage-devel/CACLE5GDV9-m4usv2%3DfXckXdYFh%2BLogsoe2LGT8%2BEWkcaaZRYmA%40mail.gmail.com.
