PSA: Trust me! Stay far away from Grok/bot/XAI!

A Reddit user warns against using Grok/bot/XAI, labeling it a "scam." They claim the xAI model is less capable than a "4B local model" and is "blind as a bat about it's own outputs," often approving "corrupt soup of overlapping triangles" as perfect. The user advises returning to "deepseek API's and local models," stating that "grok is a scam, we were missing nothing."

Unlike other user reviews, this post directly compares xAI's performance to a "4B local model," finding it less capable and prone to approving flawed outputs.