Kirill Kliavin 🇺🇦
@kakauandme
Building impactful products 🤓
ID: 83937858
https://github.com/kakauandme 20-10-2009 22:12:00
4,4K Tweet
277 Followers
574 Following
The race to minimise LLM hallucinations is still on. 37% of Perplexity responses contain incorrect information, according to a study by Columbia Journalism Review . ChatGPT is 67% misleading, and Grok is wrong 94% of the time. cjr.org/tow_center/we-…