melvinmt

Joined 15 years ago

Karma

3,078

Submissions

950

Member since

Feb 1

https://www.linkedin.com/in/melvintercan

View on Hacker News

Recent submissions

Comment · 5 years ago

Let's take this example: https://beta.openai.com/examples/default-factual-answering

95% of the token usage in this example would consist of the prompt and I would only get 1 sentence in return. So if I wanted to generate another sentence, I would have to pay 95% of the cost towards the prompt again and again... Isn't there a way to create a template for the prompt so you only pay for the generated sentences?

Comment · 5 years ago

> In the simplest case, if your prompt contains 10 tokens and you request a single 90 token completion from the davinci engine, your request will use 100 tokens and will cost $0.006.

It's weird that the prompt characters are also being counted towards your token usage. So you're penalized if you have an elaborate prompt with lots of examples (as shown in the docs)?

Comment · 5 years ago

> lacked transparency (Substack)

What do you mean exactly by this?

Comment · 5 years ago

Yes, part-time PhD's are pretty common in Europe. They also commonly take 3-4 years instead of 6...