506
Advanced OpenAI models hallucinate more than older versions, internal report finds
(www.ynetnews.com)
this post was submitted on 22 Apr 2025
506 points (100.0% liked)
Technology
69867 readers
2809 users here now
This is a most excellent place for technology news and articles.
Our Rules
- Follow the lemmy.world rules.
- Only tech related news or articles.
- Be excellent to each other!
- Mod approved content bots can post up to 10 articles per day.
- Threads asking for personal tech support may be deleted.
- Politics threads may be removed.
- No memes allowed as posts, OK to post as comments.
- Only approved bots from the list below, this includes using AI responses and summaries. To ask if your bot can be added please contact a mod.
- Check for duplicates before posting, duplicates may be removed
- Accounts 7 days and younger will have their posts automatically removed.
Approved Bots
founded 2 years ago
MODERATORS
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
just one more terawatt-hour of electricity and it'll be accurate and creative i swear!!
/S is mandatory
Because otherwise it would be totally believable
...
...
/s
This particular anti-AI stance always reminds me of religion gradually losing ground to science.
It’s been pointed out by some folks that if religion’s domain is only ‘what science can’t explain,’ then the domain of religion is continuously shrinking as science grows to explain more and more.
If your anti-AI stance is centered on ‘it wastes power and is wrong too often,’ then your criticism becomes more irrelevant as the accuracy improves and models become more efficient.
The assumption here is that the AI will improve. Under the current approach to AI, that might not be the case, since it could be hitting its limitations and this article may be pointing out a symptom of those limitations.
You’re obviously not interacting with AI much. It is improving, and at an alarming rate. I’m astounded at the difference between AI now vs 3 years ago. They’re moving to new generations in a matter of months.
My point is that the rate of improvement is slowing down. Also, its capabilities are often overblown. On the surface it does something amazing, but then flaws are pointed out by those who have a better understanding of the subject matter, then those flaws are excused with fluff words like "hallucinations".
All it needs to do is produce less flaws than the average human. It’s already passed that mark for many general use cases (which many people said would never happen). The criticism is now moving to more and more specialized work, but the AI continues to improve in those areas as well.