I then pointed out that the standard answer does not work, because of my changes to the scenario Copilot apologized and produced a nonsensical answer that confused who was on what side of the river at what point and who might eat what: 2/3
Conversation
Notices
-
Embed this notice
Anatol Stefanowitsch (astefanowitsch@mastodon.social)'s status on Thursday, 16-May-2024 18:45:33 JST Anatol Stefanowitsch - GreenSkyOverMe (Monika) repeated this.
-
Embed this notice
Anatol Stefanowitsch (astefanowitsch@mastodon.social)'s status on Thursday, 16-May-2024 18:45:33 JST Anatol Stefanowitsch And that is why I will continue to laugh at people who think that they can use LLMs for anything but creating statistically plagiarized answers to common questions. 3/3
-
Embed this notice
Anatol Stefanowitsch (astefanowitsch@mastodon.social)'s status on Thursday, 16-May-2024 18:45:34 JST Anatol Stefanowitsch Inspired by @devopscat and others, I created my own alternative version of the wolf-cabbage-goat problem, where the goat is carnivorous and the wolf is vegetarian, and presented it to MS Copilot. It ignored my changes to the scenario and produced the standard answer: 1/2
GreenSkyOverMe (Monika) repeated this.