Elon, stop trying to make Grok happen
Original reporting by The Verge
Elon Musk’s ambition to build a “truth-seeking” AI chatbot appears to be running into a harsh reality: Grok isn't very good, and few people are using it. A recent Reuters investigation reveals a striking disparity in adoption, particularly within the U.S. government. Reviewing hundreds of federal AI uses, Reuters found Grok or xAI mentioned in only three instances, typically for basic tasks like document drafting, and always alongside market leaders like OpenAI and Google, whose models appeared in hundreds and dozens of examples, respectively. This sparse usage extends to more ambitious government projects, where Grok again registered a mere three entries compared to its rivals’ substantial presence.
The performance gap
Sources close to the Pentagon suggest a simple explanation: Grok "is just not the best model out there," with staffers preferring superior alternatives. Public leaderboards confirm this, consistently placing Grok outside the top tier. This underperformance is particularly awkward given that Grok and the broader xAI initiative are central to SpaceX's upcoming IPO, with the company touting a colossal $28.5 trillion AI market opportunity. Furthermore, Musk’s team has reportedly resorted to contentious training methods, including using OpenAI’s models, and the chatbot’s public "spicy" persona has led to a track record of offensive outputs, raising significant reputational and legal risks. The core truth about Grok, it seems, is far from profound: it’s an underperforming product that carries immense pressure for SpaceX’s future.
The data paints a clear picture: Grok, despite its lofty claims and central role in SpaceX’s projected valuation, is currently an underperforming player in a fiercely competitive AI landscape. Its minimal government adoption, coupled with expert consensus on its technical inferiority and a public persona marked by controversial outputs, casts serious doubt on its viability as a frontier model. The reliance on distillation from rival models further underscores its developmental shortcomings, while Musk’s reported strong-arming of potential investors suggests a desperate attempt to create artificial demand rather than genuine market traction.
Future Challenges Ahead
The implications extend far beyond Grok itself. This situation represents a significant test for xAI’s credibility and, by extension, SpaceX's audacious IPO strategy, which hinges heavily on the chatbot's success. Should Grok fail to mature into a truly competitive enterprise solution, it could severely undermine investor confidence and the astronomical valuations projected. Moreover, it raises broader questions about the feasibility of developing "less censored" or "truth-seeking" AI for professional environments, where accuracy, safety, and reliability are paramount. For xAI, the path forward demands not just technological leaps, but a fundamental reassessment of its product strategy and market positioning to avoid becoming a cautionary tale in the rapidly evolving world of artificial intelligence.