I tried Grokipedia and I think it beats Wikipedia — in some cases. (Citation needed.) (www.businessinsider.com)

xAI’s Grokipedia — positioned as Elon Musk’s “anti‑woke” alternative to Wikipedia — is a large AI‑generated encyclopedia (about 800,000 pages so far) that often mirrors Wikipedia entries but sometimes reorganizes and expands sparse or poorly written articles by crawling the web and synthesizing available sources. A hands‑on comparison found many non‑political pages nearly identical to Wikipedia (which allows scraping), but Grokipedia occasionally produces cleaner, more cohesive writeups for neglected topics — e.g., a tightened school history or a more narrative account of an obscure historical figure. At the same time, hot‑button topics show predictable editorial slants and some entries cite dubious sources (a Facebook post flagged as viral clickbait), and xAI hasn’t explained how pages are selected or vetted. For the AI/ML community this is a useful case study in large‑scale content synthesis: LLMs can rapidly aggregate dispersed facts and improve readability, but they also amplify risks — hallucinations, low‑quality citations, biased framing and unknown selection algorithms. The pragmatic takeaway is hybrid workflows: use AI to propose drafts or surface candidate facts, but keep human editors in the loop to verify sources and correct bias. Jimmy Wales’s suggestion — AIs that extract facts from existing citations for human review — captures the safest path forward: AI as an augmentation tool and sandbox, not an autonomous replacement for curated knowledge.