Some Grok users are seeing the chatbot return nonsensical responses, according to TechCrunch, in a glitch that appears to affect at least part of xAI’s consumer chatbot experience. TechCrunch reports that affected users described using Grok Lite and noticing the issue as early as Wednesday morning. In one example cited by the outlet, a user who asked the model to generate a PDF received several paragraphs of disjointed words rather than a usable answer. Another user checking source links reportedly found a set of links pointing to reinforcement learning research sites. The scope is still unclear. TechCrunch said it could not reproduce the issue in its own testing and assessed that the bug is likely affecting only a small subset of users. xAI did not respond to TechCrunch’s request for comment. The issue nevertheless generated visible user confusion. TechCrunch reports that Grok’s Reddit community saw a wave of complaints, with some users saying a refresh fixed the problem and others saying the nonsensical outputs continued after multiple refreshes. The apparent boundary matters: TechCrunch says the bug appears limited to direct queries on Grok.com. The Grok account on X.com was not affected, according to the report. The Grok account acknowledged the problem in replies to unhappy users on X, according to TechCrunch. It described the output as a rare temporary generation glitch, said the official xAI status page showed Grok services fully operational with no incidents, and advised users to start a fresh chat or regenerate the answer. That leaves two facts in tension but not necessarily contradiction: users are seeing a real product-level failure in at least some sessions, while xAI’s official status surface did not register it as a broader service incident. For operators, that is a familiar distinction in AI products: a generation-quality failure can be visible to users even when infrastructure health checks report normal availability. Who benefits: Users who encounter the issue may be able to recover by starting a fresh chat or regenerating, according to the Grok account’s guidance reported by TechCrunch. Competing chatbot providers may use incidents like this to emphasize reliability, but this report does not establish any user migration. Who's exposed: xAI is exposed to user trust risk if the glitch persists or appears broader than currently reported. Grok Lite users making direct queries on Grok.com appear to be the affected group identified by TechCrunch.