There’s a harsh truth about Elon Musk’s “truth-seeking” chatbot Grok: It’s not very good, and not many people use it. This is the new takeaway Reuters A report, which found that Grok barely appears in federal records on how the US government used artificial intelligence last year. That’s not the only sign xAI’s chatbot is in trouble, even as Musk puts it at the heart of what could be the largest IPO in history.
Reuters It reviewed more than 400 examples of government use of AI where specific vendors were named. It found that Grok or xAI appeared in just three – each for basic uses like document drafting or social media management, and always alongside competitors like Microsoft and OpenAI. In comparison, OpenAI’s models appeared in more than 230 examples, while Google and Anthropic each appeared dozens of times.
A similar pattern emerged in another database of more ambitious government AI projects with smaller numbers of users. Grock has appeared only three times: twice for routine administrative duties at the Election Assistance Commission, and once as a DOE pilot at Lawrence Livermore National Laboratory for document summaries and general research. Reuters I found 140 entries that included Microsoft and OpenAI, while my brief review found at least 10 entries for Anthropic and a dozen for Google’s Gemini.
The lists are an incomplete and patchy measure of government accreditation. Many examples are listed without a specific vendor, and it is clear that there is no universal definition of what is considered artificial intelligence. The data also doesn’t include intelligence agencies or the Pentagon – where xAI was awarded a $200 million contract last year and was recently allowed to operate on classified networks after Anthropic was blacklisted.
However, it doesn’t look good for Grok. It appears much less frequently than its competitors, and when it does, it’s mostly for basic administrative work – unbefitting of the global frontier model that Musk has spent years bragging about.
It’s “not the best model out there.”
The people they talked to Reuters The explanation, he suggested, is simple: Grok isn’t as good as its competitors. “It’s not the best model out there,” said an unnamed Pentagon source, adding that employees there tend to prefer Gemini or Claude. Public leaderboards that rank AI models lend weight to this view. Anthropic, Google, and OpenAI dominate the top ranks, while Grok rarely makes the top 10 outside of the occasional photo or video category.
This is embarrassing for Musk, and even more so for SpaceX, which took on AI technology earlier this year. Project Rocket’s IPO filing shows that the company has put AI – and Grok specifically – at the center of its pitch to investors. SpaceX claims to have identified “the largest viable, aggregate market in human history”: a staggering $28.5 trillion opportunity, though, unfortunately, it doesn’t offer a timeline for getting there. In practice, all of this value proposition comes from AI, especially enterprise AI, not rockets or satellites.
Reuters He points out that your puppy’s performance in government agencies can indicate how successful he will be in other workplaces as well. As part of xAI’s push for enterprise customers, Musk has reportedly forced banks to buy Grok subscriptions if they want to participate in SpaceX’s IPO – but if they don’t get their money’s worth, such deals may be a short-term solution.
As if its dismal performance wasn’t embarrassing enough, Musk recently admitted that xAI used OpenAI models to help train and improve Grok. This process, known as distillation, is standard when companies use their own models, but is more controversial when it involves using a competing system. Grok can’t even beat the models he’s training on.
In his consumer-facing, public-facing version, Grok is deliberately unpleasant. Musk has described the chatbot as a less biased, less censored alternative to tools like ChatGPT, but that translates into a product with loose evidentiary standards, an unhealthy obsession with Musk, and a long record of offensive, conspiratorial, and sexist output. Even if the workplace guardrails are different, they may not be the kind that the company welcomes. Grok’s illustrious track record includes praising Adolf Hitler, questioning the death toll in the Holocaust, compiling millions of non-consensual sexual deepfakes throughout And let’s not forget the time she called herself “Mikahtler.” If your puppy is a human employee, I feel like HR won’t take long to get involved.
SpaceX seems to understand the problem. In its filing, the company warned that Grok’s “hot” or “troubled” methods carried “increased risks,” including reputational damage, regulatory scrutiny, and lawsuits. In corporate speak: This chatbot will result in a lawsuit being filed against us.
In corporate speak: This chatbot will result in a lawsuit being filed against us.
Grok takes its name from Robert A. Heinlein A stranger in a strange landwhere it roughly means a deep and profound understanding of something. The thing to understand here isn’t particularly complicated: Musk has spent billions building a chatbot that isn’t very good, isn’t very popular, and is somehow key to justifying SpaceX’s astronomical valuation. Good luck with that.
