HackerTrans
トップ新着トレンドコメント過去質問紹介求人

mucha

no profile record

投稿

AI and Travel Planning: The Case for Human Expertise

archaeology-travel.com
2 ポイント·投稿者 mucha·7 か月前·0 コメント

Netflix 'all in' on leveraging AI as the tech creeps into entertainment industry

cnbc.com
1 ポイント·投稿者 mucha·9 か月前·1 コメント

What It's Like to Be 'Mind Blind'

time.com
2 ポイント·投稿者 mucha·昨年·0 コメント

AI Hallucinations: Provably Unsolvable – What Do We Do?

mindprison.cc
4 ポイント·投稿者 mucha·昨年·3 コメント

コメント

mucha
·6 か月前·議論
Adam added a comment to that thread with the 75% number and more context.
mucha
·7 か月前·議論
LLMs are non-deterministic for everyone. Give it time.
mucha
·9 か月前·議論
https://archive.ph/f4s3A
mucha
·昨年·議論
That's actually an ad for X Premium+. No embarrassing ads!
mucha
·昨年·議論
What Satya says: “I’d say maybe 20%, 30% of the code that is inside of our repos today and some of our projects are probably all written by software,”

First line from the article: In April, Microsoft’s CEO said that artificial intelligence now wrote close to a third of the company’s code.

Software != AI

Source: https://www.cnbc.com/2025/04/29/satya-nadella-says-as-much-a...

CNBC misquotes Satya in the same article with his actual quote.
mucha
·昨年·議論
Interesting. How do existing systems catch Task Requirement hallucinations?
mucha
·昨年·議論
It looks like a lot of it was coming from the saveafoxsnark subreddit.

https://x.com/reddit_lies/status/1937527595236561090

https://pbs.twimg.com/media/GuKLR2IaYAAMziS?format=jpg&name=...
mucha
·昨年·議論
Aphantasia solved hallucinations for me.
mucha
·昨年·議論
What happens when the LLM responsible for checking decides to ignore your explicit conditions?
mucha
·昨年·議論
That's happening. Unmodified LLM outputs aren't copyrightable.
mucha
·昨年·議論
Free beer tomorrow.
mucha
·昨年·議論
Cool. What goods/services does your business provide to customers?
mucha
·昨年·議論
Capability != Reliability