[arrow7000]

Isn't this exactly how GANs work already?

[jedberg]

Yes. But from I've seen no one has applied it to the latest Generative AIs.

[dereg]

I’m pretty sure Anthropic’s Claude is doing that.

https://scale.com/blog/chatgpt-vs-claude

[arrow7000]

Maybe an adversarial approach was used in training these models in the first place?

[sharemywin]

It was they were' trained using reinforcement learning with human feedback to create the critic.

[jedberg]

I hadn't thought about human feedback being an adversarial system, but I guess that makes sense, since it's basically a classifier saying "you got this wrong".