The most prestigious law school admissions discussion board in the world.
|
 |
    4 posters qualify.
You have 0 points. Copying this list does not send funds. Paste it on Disperse, then connect your wallet there. |
Codex 5.6 is too conservative
| cheese-eating navy resort | 08/29/26 |
| Crusty wonderful ape | 08/29/26 |
| fuchsia self-centered set | 08/29/26 |
| hateful azure digit ratio | 08/29/26 |
Poast new message in this thread
Date: August 29th, 2026 10:08 AM Author: cheese-eating navy resort It's funny that people who don't use ai or read about it second hand are still talking about sycophancy and hallucinations like it's 2024. 5.6 is overly cautious. Validators come out conservative to the point where yield gets reduced. Preflights are over-engineered. Research papers come out insanely hedged to the point of epistemic cowardice. This is actually an ai-tell now, because if a human were doing the work they would lie and cheat and make overconfident statements. If people decide to do work "natural" now, people are going to notice the human-generated hallucinations. (https://autoadmit.com/thread.php?thread_id=5898683&forum_id=2#50101433) |
Date: August 29th, 2026 11:00 AM Author: Crusty wonderful ape I've noticed this, too. It's still very good for high-level scaffolding, however. You just need to instruct it properly for the lower-level monotony. (https://autoadmit.com/thread.php?thread_id=5898683&forum_id=2#50101496) |
|
Date: August 29th, 2026 11:02 AM Author: fuchsia self-centered set claude just wants to chat (https://autoadmit.com/thread.php?thread_id=5898683&forum_id=2#50101499) |
|
Date: August 29th, 2026 11:32 AM Author: hateful azure digit ratio Cr took me 6 hours to realize the issue was my mapping documentation not the harness layer. (https://autoadmit.com/thread.php?thread_id=5898683&forum_id=2#50101550) |