zlacker

[return to "Chomsky on what ChatGPT is good for (2023)"]
1. submet+25[view] [source] 2025-05-25 17:47:54
>>mef+(OP)
Chomsky’s notion is: LLMs can only imitate, not understand language. But what exactly is understanding? What if our „understanding“ is just unlocking another level in a model? Unlocking a new form of generation?
◧◩
2. smokel+g5[view] [source] 2025-05-25 17:50:16
>>submet+25
Understanding is probably not much more than making abstractions into simpler terms until you are left with something one can relate to by intuition or social consensus.
◧◩◪
3. HPsqua+B6[view] [source] 2025-05-25 17:58:43
>>smokel+g5
Transforming, in other words.
◧◩◪◨
4. egbert+z7[view] [source] 2025-05-25 18:05:51
>>HPsqua+B6
Just because it can transform doesn't mean that the logic still remains correct.

I found this out when attempting to transform wiki pages into blog-specific-speak, repeatedly.

[go to top]