News for hackers. Stuff that matters.
Re:Claude 3.7 Sonnet and Claude Code by nomel on Monday February 24, 2025 @06:48PM (#43166338)
That 0.77 can save hours of work though, fighting with or being misdirected by other LLM. And, relative to hourly rate, or a cup of coffee, it's incredibly insignificant, if just used for the heavy questions.

My LLM client can switch between whatever models, mid conversation. So I'll have a question or two in the more expensive, then drop down to the cheaper for explanations/questions that help me understand. Rewind time, then hit the more expensive models with relevant prompts.

At the edges, it really ends up being "this is the only model that can do this".

Re:Claude 3.7 Sonnet and Claude Code by elliot07 on Thursday February 27, 2025 @07:26PM (#43200237)
After playing with it for a few days, I agree the cost is easily worth the $$. It's almost in a different playing field of "normal" LLM usage.. and more so in the "good productivity tools" realm.