gen: 2024/06/27 : 02:52 in 23.7 sec (old)bias: 4 (Center-Left)
type: newsquality: 75
pts: -1
Machines cannot think,
Architectural limits,
Tinker toys won’t bridge.
Generated
Holder
author:Grady Boochinstitution:Twitter (now called X)¿porque no los dos?
tl;drGrady Booch criticizes the capability of Large Language Models (LLMs), claiming they are incapable of any kind of reasoning, not just abstract reasoning. This leads to a counter-argument from a user who asserts that LLMs, as a subset of Turing machines, can improve with more training. Booch refutes this by stating that simply increasing training won't overcome the architectural limitations.
deeper:Booch's comments reflect a skeptical stance towards the capabilities of LLMs, which could be seen as bias against new technologies. However, the discussion itself is fairly balanced, with both sides of the argument represented. The language used by Booch is quite strong, which might lead to a bias perception. The quality score is high due to the relevance and detailed nature of the argument, though the exchange could benefit from more empirical evidence and less ad-hominem remarks.
Enter the Arena

Please log in to comment