That tweet shows that Musk has no idea how science, historiography, or LLM work, which isn't at all surprising. Thinking a search engine chat bot can evaluate and synthesize knowledge is idiotic.
How do you stop a search engine from exposing lies by simply finding information? What data do you train it on if mainstream media and scientific consensus are "wrong"? Twitter posts by "Autism Capital" and "Doge Designer"?
It's a long running joke that engineers are epistemologically retarded.
Reminder that "AI" and "reasoning" in this context are marketing terms that hook into the common usage meaning of these words, which has no relation to the actual processes and products being referred to as such.
Superhuman intelligence can't figure out that cats don't use "AI" apps.
View attachment 7537159
Recent events:
View attachment 7536980
link
View attachment 7536984
link
Both Musk and the people he amplifies lie incessantly. Grok is an extremely easy way to "fact-check" these antisocial degenerates: People just ask "Grok, is this true", and Grok says "No, it isn't", and then provides data it bases that answer on.
Funnily enough, people trust it and see it as an authoritative source, because people like Musk have so consistently misrepresented what it actually does.
They will not stop lying, since they would lose their audience (and blue checkmarks would also lose their livelihood) if they started to be truthful, so Musk is trying to bias the bot instead.