[greenflag]

Likely going to be a wave of research/innovation "regularizing" LLM output to conform to some semblance of reality or at least existing knowledge (e.g. knowledge graph). Interesting to see how this can be done quickly enough...

[mvcalder]

It will be interesting to see what insights such efforts spawn. For the most part LLMs specifically, and deep networks more generally, are still black boxes. If we don't understand (at a deep level) how they work, getting them to "conform to some semblance of reality" feels like a hard problem. Maybe just as hard as language understanding generally.

[kneebonian]

> Likely going to be a wave of research/innovation "regularizing" LLM output to conform to some semblance of reality or at least existing knowledge

This is a much more worrying possiblity, as there are many people who have at this point chosen to abandoned reality for "their truth" and push ideas that objective facts are inferior to "lived experiences". This is a much bigger concern around AI in my mind.

“The Party told you to reject the evidence of your eyes and ears. It was their final, most essential command.” ― George Orwell, 1984

[vore]

As fun as quoting 1984 is, there is a huge gap between that and just not making up the winner of the Super Bowl so confidently.

[visarga]

Probably the hottest research trend in 2023. LLMs are worthless unless verified.

[whimsicalism]

Really? I already get a huge amount of value out of LLMs even if they hallucinate.

Or is this just HN tendency towards hyperbole?

[visarga]

Interesting, care to give an example? Exclude fiction, imagination and role playing, where hallucination is actually a feature.

[visarga]

coming back with a link: https://mobile.twitter.com/ylecun/status/1625554772098002944

this tween from Yann LeCun came after my message was posted