bary12

bary12 on Hacker News

Re:30% Drop In o1-Preview Accuracy When Putnam Problems Are Slightly V... on Wednesday January 01, 2025 @06:31PM (#42570188)

I don't get why this matters at all? I looked at the o1-preview paper, Putnam is not mentioned. Meaning that: a) OpenAI never claimed this model achieves X% on this dataset. b) likely, OpenAI did not take measures to exclude this dataset from training. Meaning the only conclusion we can draw from this result is: when prompted with questions that were verbarim in the dataset, performance increases dramatically. We already know this, and it doesn't say anything about the performance of the model on unseen problems.

Austin's Butterfly: How to Learn from Feedback

Posted by bary12 on Wednesday June 05, 2024 @06:26PM
from the brainhacking.blog dept.

https://www.brainhacking.blog/p/austins-butterfly

( Read More... | 1 comment | 3 points | on HN )

Austin's Butterfly

Posted by bary12 on Wednesday May 29, 2024 @12:46PM
from the brainhacking.blog dept.

https://www.brainhacking.blog/p/austins-butterfly

( Read More... | no comments | 2 points | on HN )

User-Agent: AutoGPT

Posted by bary12 on Saturday April 15, 2023 @05:55PM
from the github.com dept.

https://github.com/Significant-Gravitas/Auto-GPT/issues/1724

( Read More... | no comments | 2 points | on HN )

Re:When Buddhism Goes Bad – How My Mindfulness Practice Led Me to Melt... on Tuesday February 21, 2023 @07:01AM (#34880079)

It's only a problem if they experience psychosis in their daily life when they don't want to. Any mental illness is only considered an illness when it causes impairment of function. other than that, probably equivalent to hallucinatory drugs.

Re:Show HN: Skip tests unaffected by code change. on Thursday October 13, 2022 @03:59AM (#33188448)

This looks interesting and definitely something missing from our CI pipeline. What languages do you intend to support?