I know @xai, @OpenAI and @AnthropicAI are going to try to raise more money than God in the coming months, touting more and more advancements over benchmarks. But are we really going to entrust all our automated factchecking capacity on this platform and others to Grok, which still acknowledges the need for precise corrections from ArAIstotle until this day? To err is to be an agent, but to arrive at epistemic consensus with different agents is to be more infallible.
(Putting aside the lack of other AI Search having multimedia fact-checking capabilities like granular factchecks of X articles, Spaces, Videos, etc.) See the link in the comments for the link to this thread ⤵️