Okay, I apologize for the tone of the conversation and for any offense caused to Jorge Ramos.
Please feel free to ask any questions. I’m happy to explain my opinion in more detail: Not all expectations of SymPy come from strictly applying algebraic algorithms to specific problems. Many people use SymPy as a flexible tool to solve mathematical problems, and in that regard, it overlaps with the goals of AI. Certainly, this overlap could reduce SymPy's "attractiveness" or "competitiveness" to some extent. I've occasionally participated in product design meetings in the industry. It’s increasingly difficult to engage people with ideas that rely solely on SymPy or similar computer algebra systems. People are already seeing the successes of AI products, and their expectations for these products are rising, especially when new developments come from companies like Google or OpenAI. Maybe Jorge Ramos or Maaz Muhammad are particularly passionate about SymPy and feel unhappy with my statements (or maybe not). I completely understand—it's similar to how someone might react if I were critical of their favorite musician or athlete. However, I want to note that I’ve spent a significant part of my life working with SymPy. I rarely make statements based on misunderstandings, and I have no intention of shaming a project that I’ve been deeply involved with. I’m always trying to learn new things and prepare for changes in the world. Our existing knowledge can become obsolete with new tools, and I’ve already noticed that there is a vast amount of knowledge outside SymPy, that could be used to understand or improve SymPy. And that's why I'm always watching over movements of AI, as well as Lean. To be more productive, SymPy itself can enhance its geometry capabilities by incorporating Wu's method or a deductive database approach, which are useful in addressing geometric challenges. I’ve also shared the implementation of the Area method with SymPy, which can be searched. On Friday, August 23, 2024 at 2:32:29 PM UTC+2 Sangyub Lee wrote: > > Do you mean that Sympy is losing competitiveness or attractiveness? > > I am not 100% definitely saying something like that, > I already know that coming up with strong opinions for that, in subjective > or religious tone, > decreases my professional credibility, so I definitely avoid expressing > such sentence. > > I already highlight the parts like 'heuristics', or 'solve', 'simplify', > and that narrows down the topic and some background, > and obviously some person already explains the general debate between > fuzzy computing and precise computing in software engineering. > and I am neither saying 100% support or anti statement for something. > > And it's problem of you not paying attention to some details like that, > but try to raise offense of my English skills. > it won't be productive or pleasant experience for for participating in > discussion or debate with you. > > On Friday, August 23, 2024 at 1:30:13 PM UTC+2 Jorge Ramos wrote: > >> I find your English pretty ambiguous. It's not clear to me if, for >> example, what you mean when you say: >> >> *"I think that the research reinforces that solve or simplify, or >> integral is losing competition*. >> >> *Because a lot of them are written with heuristics that won't win with >> AI,"* >> Do you mean that Sympy is losing competitiveness or attractiveness? >> >> >> >> On Sun, Aug 18, 2024 at 7:34 PM Sangyub Lee <syle...@gmail.com> wrote: >> >>> AI achieves silver-medal standard solving International Mathematical >>> Olympiad problems - Google DeepMind >>> <https://deepmind.google/discover/blog/ai-solves-imo-problems-at-silver-medal-level/> >>> >>> Recently, Google had announced the result that their AI >>> model, AlphaProof and AlphaGeometry can silver medal in IMO problems. Their >>> system is hybrid of symbolic models, and uses proof assistant Lean as >>> backend, which guarantees that the proof can be verified automatically. >>> ChatGPT had many problems that it can hallucinate the steps of proof, >>> and keep human verifying their result, as well as understaing the steps, so >>> expressing proof as formal proof statements is a gain. >>> >>> I think that the research reinforces that solve or simplify, or integral >>> is losing competition. Because a lot of them are written with heuristics >>> that won't win with AI, and we also have concerns about code around them >>> are getting messy. >>> >>> I think that if we want to avoid the losing competition, and make AI >>> systems work collaborative, symbolic computation should be focused to solve >>> only a few 'formal' problems in 100% precision and speed. >>> >>> I already notice that there is research to connect Wu's method to >>> AlphaGeometry >>> [2404.06405] Wu's Method can Boost Symbolic AI to Rival Silver Medalists >>> and AlphaGeometry to Outperform Gold Medalists at IMO Geometry (arxiv.org) >>> <https://arxiv.org/abs/2404.06405> >>> Although symbolic system would no longer competitive solution to general >>> math problems, the 'formal' symbolic systems can still be valued. (I also >>> hear that AlphaGeometry2 is using Wu's method, but I'm trying to verify the >>> sources) >>> >>> I also think that such advances in AI systems can raise concerns about >>> software engineering careers, or educational system, which may be >>> interesting for some readers in the forum. >>> >>> For example, math exams can be pointless in the future, even to identify >>> and train good science or engineers in the future, because trained AI >>> models can beat IMO. I think that in AI age, the education should change, >>> such that it is not bearing through boring and repetitive systems, which >>> does not even reflect the capability of future engineers or scientists. >>> >>> Also, I notice that software engineering is changing, because AI models >>> can complete a lot of code, and precision is improving, or people are >>> improving the skills of prompting. >>> It also seems to be deprecating code sharing efforts for open source >>> communities, because code can be generated rather than shared. >>> >>> -- >>> >> You received this message because you are subscribed to the Google Groups >>> "sympy" group. >>> To unsubscribe from this group and stop receiving emails from it, send >>> an email to sympy+un...@googlegroups.com. >>> To view this discussion on the web visit >>> https://groups.google.com/d/msgid/sympy/e7898bdb-d1e4-49fd-94c7-66ba8a840511n%40googlegroups.com >>> >>> <https://groups.google.com/d/msgid/sympy/e7898bdb-d1e4-49fd-94c7-66ba8a840511n%40googlegroups.com?utm_medium=email&utm_source=footer> >>> . >>> >> >> >> -- >> ---------------------------------------------------------------------- >> Then maybe we are gods after all,.... >> baby gods only just now waking up to our true power >> > -- You received this message because you are subscribed to the Google Groups "sympy" group. To unsubscribe from this group and stop receiving emails from it, send an email to sympy+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/sympy/1d4f6aa9-452f-42fe-98e7-f2497157d9c9n%40googlegroups.com.
