[fiddlerwoaroof]

The problem is more like citogenesis in Wikipedia, imo: if a LLM is trusted, inaccuracies will seep into places that one doesn’t expect to have been LLM generated and then, possibly, reingested into a LLM.

[asdfaoeu]

That's already an issue without an LLM in the middle.

[fiddlerwoaroof]

Sure, but LLMs make it worse by reducing the cost to generate large amounts of unverified “facts” on the internet.

[kordlessagain]

And the problem lies with the Internet, not the hypothetical models.