Hacker News Reader

I see your point. On the other side, I can think of one reason for wanting to remove superfluous words: the user pays per token.

View on HN

If you're paying per token for ChatGPT, I am surprised. You pay nothing to get access to ChatGPT. Plus subscribers get access to GPT4, but they pay per month (with ratelimits are per N reqeusts / X hours), not per token.

If you're paying for API, you have text-davinci, it is not behaving the way like free ChatGPT behaves.

> If you're paying for API, you have text-davinci, it is not behaving the way like free ChatGPT behaves.

No, you can get both gpt-3.5-turbo (GA) and gpt4 (behind a waitlist) via API, not just text-davinci and other non-chat models.

Try asking same complex question from OP to gpt-3.5-turbo and text-davinci. 80% chance they'll be very different no matter the temperature.

More like 99% chance, as GPT-3.5-turbo is just as large as GPT-2-XL.

Reply on news.ycombinator.com