5 ms·
ChatGPT is just the brand. I guess you're using GPT-4, but if you're using the default model (GPT-3.5) that would certainly explain below-expectation results.
by calderknight 2y ago
ChatGPT is just the brand. I guess you're using GPT-4, but if you're using the default model (GPT-3.5) that would certainly explain below-expectation results.
- chrisjj 2y ago> ChatGPT is just the brand. "ChatGPT" is a model, says its provider. But yes, the web page says model GPT-3.5.
- calderknight 2y agoMystery solved! The answer to your OP is that the reason anyone can trust ChatGPT for code is that they use a much better model than the one you're using! GPT-3.5 is ancient and way behind GPT-4. In fact, there are now tens of organisations who have developed model classes way ahead of GPT-3.5.
- chrisjj 2y ago> the reason anyone can trust ChatGPT for code is that they use a much better model than the one you're using! GPT-3.5 is ancient and way behind GPT-4. Testing that theory now using Bing Copilot GPT-4. Failed. unmatched close parenthesis Line: 11, Column: 11, File-Position: 408 Input is: Translate this to Common Lisp. def solve(graph): a = 0 while a < len(graph): for b in range(len(graph)): if b > a and graph[a].intersection(graph[b]): graph[b] = graph[a] | graph[b] graph.pop(a);a-=1 break a+=1 return graph a = [{1, 2}, {4, 5}, {1, 3}, {1, 4}] print((solve( a ))) Output is: (defun solve (graph) (let ((a 0)) (loop while (< a (length graph)) do (loop for b from 0 below (length graph) do (when (and (> b a) (intersection (nth a graph) (nth b graph))) (setf (nth b graph) (union (nth a graph) (nth b graph))) (setf graph (remove-if (lambda (x) (equal x (nth a graph))) graph)) (decf a) (return)))) (incf a)) graph)) (let ((a (list (list 1 2) (list 4 5) (list 1 3) (list 1 4)))) (format t "~a~%" (solve a)))
- calderknight 2y agoYeah, if going that route the answer to the OP is "most people aren't using it for Lisp".
- chrisjj 2y agoPerhaps Lips programmers have more sense than most :)