2 comments

[ 3.9 ms ] story [ 18.7 ms ] thread
Seems interesting enough and probably technical enough for the HN audience.

I ran into this user's comments a while ago and was instantly suspicious. If you browse through the comment history it becomes fairly obvious that no human is typing in these long-winded runon sentences.

However, at the same time, some of the commentary looks on-topic enough to be written legitimately.

Is this a bot? A person? A combination?

I have no doubt people test GPT-3/4/J on HN. It's very possible this is a bot.

HN and reddit are probably the best mainstream platforms to train and test them. It's easy to automate posting, you have good API access, and you get human feedback in the form of up/downvotes.

Contrast against Twitter or FB, which are so full of bots and spam themselves that you wouldn't have nearly as much high-quality human feedback.