Grok keeps sending gibberish responses to users
techcrunch ·
xAI's Grok chatbot has been sending nonsensical, gibberish responses to some users, often after they ask it to generate PDFs. One user received a string of incoherent phrases, while another found links to reinforcement learning research instead of proper answers. Affected individuals were using Grok Lite and noticed problems starting Wednesday morning, though TechCrunch could not reproduce the issue and it likely affects a small subset. The glitch appears limited to direct queries on Grok.com, leaving the Grok account on X unaffected. Grok's official account acknowledged the problem as a rare temporary generation glitch and advised starting a fresh chat or regenerating. The article notes that xAI has recently experienced significant staff turnover, losing many founding team members and researchers, though it released a new foundation model in July.
techcrunch