💾 Archived View for dioskouroi.xyz › thread › 24924035 captured on 2020-10-31 at 00:48:05. Gemini links have been rewritten to link to archived content

View Raw

More Information

-=-=-=-=-=-=-

How I used GPT-3 to hit Hacker News front page 5 times in 3 weeks

Author: masterrr

Score: 27

Comments: 5

Date: 2020-10-28 20:55:36

Web Link

________________________________________________________________________________

minimaxir wrote at 2020-10-28 21:18:09:

The title is very misleading: GPT-3 generated the titles, but as usual, they were manually curated on the "picked the title I liked the most" criteria.

There's more to getting to the front of Hacker News than just a good title (although it helps); ironically this framing undersells the writing capabilities of the author by attributing success to the AI only.

There _are_ ways in theory for AI models to generate "better" titles, which is something I'm actively researching (rough example:

https://github.com/minimaxir/gpt-3-experiments/tree/master/e...

).

dunkeylim wrote at 2020-10-29 02:39:29:

It doesn’t seem misleading to me, they explicitly said “How I used GPT-3”, not “How GPT-3 generated”. (Unless they edited the title)

simonebrunozzi wrote at 2020-10-30 15:30:31:

You are correct, but I would still agree that it is misleading to some extent.

sullyj3 wrote at 2020-10-29 04:33:51:

Sometimes people infer things that aren't strictly speaking logically implied. That doesn't make it not misleading, if it reliably causes people to make those inferences.

minimaxir wrote at 2020-10-29 03:18:43:

It's implying a direct causal effect to getting to the front page, which given the article argument is not well supported.