350M Tokens Don't Lie: Love and Hate in Hacker News

outerbounds.com

10 points·savin-goyal··4 comments

I like the part where the "harmonious & positivity" score of 5 is considered neutral on the 0-9 scale, but the LLM categorized absolutely nothing to be right in the middle:

>the LLM is utterly befuddled about what a neutral discussion looks like - there are no 5 in the results. Or, maybe this is a snarky note by the LLM noting that humans are incapable of unemotional, unbiased discourse.

The snark of the LLM itself may not be good for maintaining a middle-of-the-road posture.

> all the AI hype aside, are mind-blowingly effective at practical tasks like this.

Thank you. Most of the hate toward LLMs seem to focus on their not being even a little oracular, with little value placed on how much of an incredible tool they are for natural language manipulation in general as in this fine story. Our new pal is not very godlike but still very nice to have around.

Exactly. Now we have a new UI paradigm and a new, incredibly powerful NLP tool in our toolkit.

Is this AGI? Of course not. Is this useful and valuable? Obviously.

> Most of the hate toward LLMs seem to focus on their not being even a little oracular

I don't see that. I see contempt for the idea that AI can be applied to everything or even tagged on to things with little or no AI for the purpose of marketing or attracting investment.

Reply on news.ycombinator.com