In yesterday's thread about Grok 4 [1], people were praising it for its fact-checking and research capabilities.
The day before this, Grok was still in full-on Hitler-praising mode [2]. Not long before that, Grok had very outspoken opinions on South Africa's "genocide" of white people [3]. That Grok parrots Musk's opinion on controversial topics is hardly a surprise anymore.
It is scary that people genuinely use LLMs for research. Grok consistently spreads misinformation, yet it seems that a majority does not care. On HN, any negative post about Grok gets flagged (this post was flagged not long ago). I wonder why.
[1] https://news.ycombinator.com/item?id=44517055
[2] https://www.ft.com/content/ea64824b-0272-4520-9bed-cd62d7623...
[3] https://apnews.com/article/elon-musk-grok-ai-south-africa-54...