@Halios@Hesamation Someone shall _own_ the proof if not knowledge, at least one person from OpenAI shall have understanding of the proof for a given problem. Else it's leading nowhere.
@boazbaraktcs I think that is because it was assumed that at least someone has ubderstood the subject matter. With AI based problem solving its not guaranteed.
@prathoshap offer two possible ways to test knowledge per course, one for those who want a bit more challenge, another for those who want certification lf completion. Or if someone is more inclined on implementation and labs, something similar can be done for those candidates as well.
@stellensatz I think for enterprise licenses AI companies guarantee that they won't use the user interaction for training externally available future models.
@stevenstrogatz its time, some frontier lab crawls Wiki for all the conjectures, arranges them in order of predicted difficulty and feed to some model or other
@lauriewired After certain limit it's monotonous. Idea behind a program is interesting. Debugging a complex bug is interesting, but just doing similar coding stuff over different geographic locations is not interesting.
@makash @_kashi_ks I too am learning Kannada script. Here is what I did to improve:
Requested ChatGpt to write my native language (Marathi) statements in Kannada. It helped a lot. One can anticipate a word as it's from a known language and then match the script to pronunciation.