Benjamin De Kraker left xAI last night over a dispute about mentioning Grok 3 in a tweet. His post ranked several AI models for code ability, placing the unreleased Grok 3 below some competitors with a ‘TBD’ qualifier. xAI demanded he delete this clearly labeled personal opinion or face termination. The absurdity? Grok 3’s existence is public knowledge, repeatedly acknowledged by both xAI and Elon Musk. The company claimed even writing “Grok 3 – TBD” violated confidentiality, despite their own blog posts discussing it. De Kraker refused to delete his opinion and resigned instead. He maintains hope for xAI’s success and speaks positively about the specific Grok feature he helped develop, though he can’t disclose details yet. I see three possibilities here: 1. xAI wanted him gone for other reasons and used this as an excuse 2. An HR department power trip 3. They disliked him ranking Grok 3 below competitors pre-release Most likely it’s a mix of factors 2 and 3. The irony is that forcing his resignation generated far more attention than his original ranking ever would have. For a company supposedly championing free speech, attempting to silence an employee’s mild, qualified opinion seems completely counter to their stated values. This situation highlights a growing tension in AI companies between transparency and control. While leaders publicly discuss unreleased products, employees face strict limitations on sharing even basic opinions about widely known developments. The incident also raises questions about how AI companies handle internal dissent and different viewpoints. If expressing a careful, qualified opinion about a publicly acknowledged product leads to threats of termination, what does that suggest about the company culture? De Kraker chose principles over position – keeping his “speech and dignity” rather than cave to unreasonable demands. His experience serves as a reminder that even in cutting-edge tech companies, basic rights like expressing mild personal opinions still need defending.
Menu
