> Not to say that there mightn't be deeper holes lurking, only that we cannot > draw any conclusions about them from this sample. For instance, j is a lot > more expressive (in the formal sense) than most other programming > languages--meaning, it has little redundancy--for instance, >/@.(>&.) _is_ a > legal (if useless) j sentence--so it may be much harder to find patterns.
Not to mention that J has evolved recently complicating the learning process for a human or otherwise (e.g., that useless sentence parsing has changed in the last couple of years, if I am not mistaken). On Tue, Dec 6, 2022 at 6:43 PM Elijah Stone <elro...@elronnd.net> wrote: > > It also attempted an exposition of the infamous incunabulum - > https://medium.com/@solarbreeze69/chatgpt-explains-arthur-whitneys-j-incunabulum-5be2ea69a298. > Very good, seemingly, but also riddled with mistakes. > > The chatbot certainly has some shallow holes and some deep ones. I suspect > that its performance at j can be attributed to one of the former, viz., there > is not much material about j on the internet (in its training set), compared > with, say, a language like python (and its performance in python is quite > impressive, so I hear). > > Not to say that there mightn't be deeper holes lurking, only that we cannot > draw any conclusions about them from this sample. For instance, j is a lot > more expressive (in the formal sense) than most other programming > languages--meaning, it has little redundancy--for instance, >/@.(>&.) _is_ a > legal (if useless) j sentence--so it may be much harder to find patterns. > > On Tue, 6 Dec 2022, Raul Miller wrote: > > > Yeah, that chatbot seems to struggle with J syntax. > > > > Probably because it doesn't use a J session to test its "J expressions". > > > > (Intelligence which is not artificial would also struggle under such > > circumstances. Which, in turn, has significant economic implications: > > expertise which is never tested in practical contexts tends to sound > > good without being useful.) > > > > -- > > Raul > > > > On Tue, Dec 6, 2022 at 3:37 PM Jose Mario Quintana > > <jose.mario.quint...@gmail.com> wrote: > >> > >> Human: What do you know about the J programming language? > >> AI: I know that J is a concise high-level programming language > >> designed to express complex algorithms in a few lines of code. As a > >> functional programming language, J is well suited for data analysis, > >> statistical computing, and machine learning applications. It has been > >> used by companies like Google and Amazon, as well as many universities > >> and research laboratories. > >> > >> Human: Is recommended to write tacit programs in the J programming > >> language? > >> AI: Yes, tacit programming is a style of writing code that is favored > >> by many J programmers. This style emphasizes the use of small > >> functions and expressions, which are combined to create larger > >> programs. This allows for concise yet expressive code that is easy to > >> read and understand. > >> > >> Human: Can you give an example of tacit J code? > >> AI: Sure! Here is an example of a tacit J function to find the maximum > >> value in an array: > >> > >> max =: >/@.(>&.) > >> > >> This code finds the maximum value in the array by using two functions: > >> (1) >&. finds the maximum value in each sub-array, and (2) >/ finds > >> the maximum value in the entire array. > >> ---------------------------------------------------------------------- > >> For information about J forums see http://www.jsoftware.com/forums.htm > > ---------------------------------------------------------------------- > > For information about J forums see http://www.jsoftware.com/forums.htm > ---------------------------------------------------------------------- > For information about J forums see http://www.jsoftware.com/forums.htm ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
