from Hacker News

Claude Sonnet 3.5 – The New Benchmark in Conversational AI

by naveensky on 6/27/24, 6:12 AM with 10 comments

• by opdahl on 6/27/24, 8:59 AM

  This reeks of AI generated blog spam created from a few other articles as input.

• by lobochrome on 6/27/24, 8:40 AM

  Getting it to enforce JSON is still a pain and not guaranteed. GPT-4o does this better.

  Also it’s vision capabilities are much less evolved.

• by tmaly on 6/27/24, 12:01 PM

  Honest question, was the article written by Claude?

• by atul_lakhotia on 6/27/24, 8:58 AM

  The results from Claude AI was lot better than ChatGPT. I had tried writing an ad copy for a client- Chatgpt gave very typical suggestions while Claude AI gave interesting different suggestions.