Elon Musk's new direction for AI chatbot Grok
Elon Musk is giving his AI chatbot, Grok, a major upgrade by retraining it with cleaner, more accurate data.
The next version—Grok 3.5 or 4—will aim for sharper reasoning and fewer mistakes by fixing errors and filling in missing info.
Musk says this is needed because there's "too much garbage" in current AI training sets.
Grok's focus on cutting down ideological bias
Grok launched in February 2025 as an alternative to mainstream chatbots like ChatGPT, with a focus on cutting down ideological bias.
Musk often calls other AIs victims of a "mind virus."
To keep things real, he even asks users to send in "divisive facts"—stuff that might be unpopular but is still true.
Grok uses synthetic data and xAI's Colossus supercomputer to boost accuracy and avoid those weird AI hallucinations.
Grok could become way more reliable
With this retraining, Grok could become way more reliable—and potentially influence how people trust chatbots.
Plus, thanks to a $300 million deal bringing Grok to Telegram, it'll soon be just a tap away for over a billion users worldwide.
If you're curious about smarter (and hopefully less biased) AI tools, this one's worth watching.