Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

When did AI take over Hacker News?

zachperk.com

Post reply on HN
Print view View original

211–220 of 245 posts

Re: When did AI take over Hacker News?

#211

Post by Eddy_Viscosity2 »

ryandrake wrote: post #17
rising-sky wrote: post #9

What I found insightful about this article was the framing of another article cited. > " This pretty negative post topping Hacker News last month sparked these questions, and I decided to find some answers, of course, using AI" The pretty negative post cited is https://tomrenner.com/posts/llm-inevitabilism/ . I went ahead to read it, and found it, imo, fair. It's not making any direct pretty negative claims about AI,…

Weird what counts as "negative" on HN. Question something politely? You're being negative. Criticize something? Negative. Describe it in a way someone might interpret badly? Negative. Sometimes it seems like anything that's not breathless, unconditional praise is considered being negative and curmudgeonly. It's turning into a "positive thoughts only" zone.

It's ok to be negative sometimes. Not just ok, but a necessary mechanism for course-correction. So even if sometimes comments might be negative, that is fine.

Now of course I'm not including aggressive or rude posts, because they are a different category.

Re: When did AI take over Hacker News?

#212

Post by tempodox »

insin wrote: post #85
jacquesm wrote: post #29

I'm more annoyed at the - clearly - AI based comments than the articles themselves. The articles are easy to ignore, the comments are a lot harder. In light of that I'd still love it if HN created an ignore feature, I think the community is large enough now that that makes complete sense. It would certainly improve my HN experience.

I added muting and annotating users to my Hacker News extension: https://soitis.dev/comments-owl-for-hacker-news

It even works for Safari, which I didn't expect, and it's free.

Thank you so much!

Re: When did AI take over Hacker News?

#213

Post by camillomiller »

zahlman wrote: post #206
camillomiller wrote: post #204

Earlier quoted context omitted.

Did you read my comment, instead of framing it as ideological? Such a framing is a quite interesting way to dismiss the issue at hand isn't it? Do you share the aknowledged reality that Meta has fostered genocide in Myanmar, as per their own admission in front of Congress, and that Instagram has led teenage girls to depression, as per admitted by internal documents seen in discovery that prove how they weaponized tho…

> Did you read my comment Yes. From Wikipedia: > An ideology is a set of beliefs or values attributed to a person or group of persons, especially those held for reasons that are not purely about belief in certain knowledge, You are passing judgment and using emotionally charged words (such as "weaponizing", which also implies intent and motivations not in evidence) to make a point about what you consider moral. And y…

I am pointing at quite a big factual moon, but you prefer to dissect my rhethorical finger. If there is a rhetoric problem here, is how obfuscating this very clear ethical discourse in rhetoric disquisitions is just an elegant way to toss the ball off the field. A diversion, that only serves the perpetuation of the status quo.

Re: When did AI take over Hacker News?

#214

Post by tempodox »

ryandrake wrote: post #20

I noticed at one point a few days ago that all 10 out of the top 10 articles on the front page were about AI or LLMs. Granted, that doesn't happen often, but wow. This craze is just unrelenting.

My subjective impression is that it has become even more prominent in the past few months. I suspect the providers are feeling increased pressure to monetize and are boosting their astroturfing and creative marketing efforts accordingly.

Re: When did AI take over Hacker News?

#216

Post by svara »

dangus wrote: post #210
svara wrote: post #197

Earlier quoted context omitted.

Not at all. I don't want to address directly your claim about lack of generalization, because there's a more basic issue with the GP statement. Even though I will say, today's models do seem to generalize quite a bit better than you make it sound. But more importantly, you and GP don't mention any evidence for why that is due to specifically using next token prediction as a mechanism. Why would it not be possible for…

> you and GP don't mention any evidence for why that is due to specifically using next token prediction as a mechanism. Again, inverted burden of proof. We don’t have to prove that next token prediction is unable to do things that it currently cannot do and has no compelling roadmap that would lead us to believe it will do those things. It’s perhaps a lot like Tesla’s “we can do robocars with just cameras” manifesto.…

I don't know what you're talking about or how anything I'm saying inverts a burden of proof (of what exactly?).

If you're saying "X can't do Y because Z" you do need to say what the connection between Y and Z is. You do need to define what Y is. That's got nothing to do with a burden of proof, just speaking in a understandable manner.

The Tesla tangent is totally unhelpful because I know exactly how to make those connections in that example.

Re: When did AI take over Hacker News?

#217

Post by zahlman »

camillomiller wrote: post #213
zahlman wrote: post #206

Earlier quoted context omitted.

> Did you read my comment Yes. From Wikipedia: > An ideology is a set of beliefs or values attributed to a person or group of persons, especially those held for reasons that are not purely about belief in certain knowledge, You are passing judgment and using emotionally charged words (such as "weaponizing", which also implies intent and motivations not in evidence) to make a point about what you consider moral. And y…

I am pointing at quite a big factual moon, but you prefer to dissect my rhethorical finger. If there is a rhetoric problem here, is how obfuscating this very clear ethical discourse in rhetoric disquisitions is just an elegant way to toss the ball off the field. A diversion, that only serves the perpetuation of the status quo.

You are not "pointing at" the moon; you are putting it somewhere it does not belong, and then getting defensive when people complain about the effect on the local change in gravity as if you can't understand why people would object to that.

I am tossing the ball off the field because this field is intended for a different sport entirely, and there is already a game in progress.

Re: When did AI take over Hacker News?

#218

Post by rising-sky »

epolanski wrote: post #91
rising-sky wrote: post #9

What I found insightful about this article was the framing of another article cited. > " This pretty negative post topping Hacker News last month sparked these questions, and I decided to find some answers, of course, using AI" The pretty negative post cited is https://tomrenner.com/posts/llm-inevitabilism/ . I went ahead to read it, and found it, imo, fair. It's not making any direct pretty negative claims about AI,…

I've always found HN's take on AI healthily skeptical. The only subset where HN gets overly negative is coding, way more than they should.

I tend to agree with this. I just the "pretty negative" adjective jarring in this case and wanted to get a sense of what some in the community here think. Seems mostly in line with your sentiment

Re: When did AI take over Hacker News?

#219

Post by svara »

Jensson wrote: post #208
svara wrote: post #197

Earlier quoted context omitted.

Not at all. I don't want to address directly your claim about lack of generalization, because there's a more basic issue with the GP statement. Even though I will say, today's models do seem to generalize quite a bit better than you make it sound. But more importantly, you and GP don't mention any evidence for why that is due to specifically using next token prediction as a mechanism. Why would it not be possible for…

> Why would it not be possible for a highly generalizing model to use next token prediction for its output? The issue is that it uses next token prediction for its training, it doesn't matter how it outputs things but it matters how its trained. As long as these models are trained to be next token predictors you will always be able to find flaws with it that are related to it being a next token predictor, so understa…

This is a perfectly fine line of argument imo but the GP didn't say that.

LLM research is trying out a lot of different things that move away from just training on next token prediction, and I buy the argument that not doing anything else would be limiting.

The model is still fundamentally a next token predictor.

Re: When did AI take over Hacker News?

#220

Post by duxup »

bfg_9k wrote: post #179
duxup wrote: post #140

Earlier quoted context omitted.

I find asking questions on the internet are increasingly seen as a negative, right out of the gate, no other questions asked. I get it to some extent, a lot of people looking to inject doubt and their own ideas show up with some sort of Socratic method that really is meant to drive the conversation to a specific point, not honest. But it also means actually honest questions are often voted or shouted down. It seems l…

I don't get it. Asking questions is never a hostile thing, regardless of the context. Honest or not, questions are simply.. that. Questions. If someone is able to find a way to take offence from a question being asked, that's pathetic.

>Asking questions is never a hostile thing

I think many people are looking for context before diving into a conversation, I think that's a human thing. It can be a waste of time / disappointing to engage in a conversation and find the other person is really not participating and is there to drive the conversation to their point.

Post reply on HN