It's trivial to prove this is wrong: invert relationships it knows about and it fails to answer based on knowledge it previously demonstrated (even with loads of hints) https://chatgpt.com/share/680dc86c-f0dc-800d-9f04-57ba2f126a... https://chatgpt.com/share/680dc90b-de28-800d-92b6-f2ef824777... Note how applying increasing pressure to answer was what caused the hallucination: hallucinations aren't tied to if the mod…
This is a logical problem that requires reasoning. Bring the fact that Barry White’s mother is Sadie Marie Carter into the context and it should then be able to infer who Sadie Marie Carter’s son is.
You're saying it requires reasoning, then suggesting a non-reasoning based solution (steering the model in-context)
This is a logical problem that requires reasoning. Bring the fact that Barry White’s mother is Sadie Marie Carter into the context and it should then be able to infer who Sadie Marie Carter’s son is.
You're saying it requires reasoning, then suggesting a non-reasoning based solution (steering the model in-context) Reasoning doesn't solve it. O3 thinks and comes up empty, and O4-mini thinks and then hallucinates much worse than even 4o did: https://chatgpt.com/share/680e7a5c-ae3c-8004-bff9-5c8a0215a1...
The model only knows who the mother of Barry White is. It doesn’t know who the son of Sadie Marie Carter is. By reasoning I meant that it could figure out who the son of Sadie Marie Carter is if you brought into the context that Sadie Marie Carter is the mother of Barry White beforehand.
Call me optimistic but we only hear about the mistakes AI makes not the successes in stories like these. Failing to review the output is the real issue in my opinion. A machine writing a legal brief and only making 30 mistakes is still incredible.
Yep. > [learning a language is] only helpful for high professional tasks or close literary study or prestige. This is a person who doesn't understand actual face-to-face communication, like, at all. Even though translation apps are amazing, in a social interaction, there's no getting over the imposition of the halting, hesitant back-and-forth of device-assisted translation. Sure, you can almost always eventually get…
This means of communication is already outdated. If you live in an urban metropolis in the US and Europe especially, but increasingly everywhere, you cannot avoid using translation apps on an almost daily basis. I’ve had close intimacy with a partner whose language I barely spoke. The intricacies of language are suitable for technical tasks but language is something which is wielded at the bottom by force, the force of expression, of a gesture. The culture of the internet is building a language of its own with a non-linguistic grammar, as exemplified by memes which increasingly lack any immediately percievable linguistic content. And LLMs are the key to all this, they are the great mediator of the world, transforming any expression into any other. But the strength of the expression still lies in its use, far more than its content. Still, as I said, there is room for deep knowledge of the cultural intricacies intertwined into any particular language: but this is a regressive element in our expression, one which is being eliminated by industrial necessity.
I asked ChatGPT to give Wikipedia links in a table. Not one of the 50+ links was valid.
Share the link to the conversation.
Share the link or it didn't happen. And if it happened, you used an older model. And if you used the most up-to-date one, you didn't prompt it correctly.
Share the link or it didn't happen. And if it happened, you used an older model. And if you used the most up-to-date one, you didn't prompt it correctly.
So.. “trust me bro” is what you’re proposing? No thanks.
Make a claim, prove it, especially one so easily proven.