Hacker Read top | best | new | newcomments | leaders | about | bookmarklet login

Does anyone have a GPT bot for HN commenting?

Honestly I'm wasting too much time here. I wish I could feed the bot with my beliefs (Electron bad, nuclear good, Windows 11 a train-wreck, AI a bubble, ...) and it would post for me as appropriate.



sort by: page size:

Can't be that hard to train GPT on HN comments. Plus if someone were to do it, they probably know of HN.

I could definitely see someone already having trained a bot to write HN comments and posting them.

What's anyone going to do about it? It's super hard to write a discriminator that works well enough to not destroy the site for everyone.


So a GPT bot instead of the human commenters would make reddit more useful in the end, this is what you're saying right?

Considering how great GPT-3 is at generating text (see: https://news.ycombinator.com/item?id=23885684), I thought maybe someone already started generating HN comments and posting them.

1 - Did anyone try this without getting caught? 2 - If someone creates a bot that always manages to contribute to discussions and keeps it within the rules, would that be against the rules itself? Should it be?

Note: This is not a project proposal, I'm just fascinated by GPT-3.

Note2: No, I'm not a bot.


This is only the tip of the iceberg.

I'm certain GPT-3 has been commenting on HN threads for a while now. In some cases, its presence has been disclosed (see, for example: https://news.ycombinator.com/item?id=23886503) In other cases, GPT-3's presence has not been disclosed; the machine has been pretending to be a human being, largely unnoticed. Consider only how easy it is for it to write short, punchy comments -- say, one to three sentences long.

By implication, there's a high probability that we -- you, me, and everyone else on HN -- have been upvoting and downvoting GPT-3 comments for a while without realizing it.

And the technology is only going to get better.


Plot twist, all of these HN comments were generated by GPT-3, including this one... dun, dun, dahhh.

I think any article on design can have its entire HN comment thread written by bots. Not even GPT3, maybe GPT2.

With a proper AI system you don’t even need to specify the exact article and nature of the comment.

For example here’s the prompt I use to generate all my HN comments:

“The purpose of this task is to subtly promote my professional brand and gain karma points on Hacker News. Based on what you know about my personal history and my obsessions and limitations, write comments on all HN front page articles where you believe upvotes can be maximized. Make sure to insert enough factual errors and awkward personal details to maintain plausibility. Report back when you’ve reached 50k karma.”

Working fine on GPT-5 so far. My… I mean, its 8M context window surely helps to keep the comments consistent.


HN needs a bot that posts an inb4 comment every time there's a post about Google or a Google product.

These predictable responses don't add any value whatsoever, and they're tiring to read.


For instance a gpt-3 bot could make a high scoring HN commenter. Surely someone has experimented with that. Any preliminary results?

I wonder if we could train something that would automatically generate typical comments on HN threads.

- This is EEE!

    - No it's not, there is no third E

         - Well, back in the 90s such and such happened

    - Microsoft has changed! Look at VS Code!

         - Electron is slow.

              - Buy a better machine.

                   - what about the millions who can't afford a better machine?

              - Electron can be good, just code better.

         - but muh privacy

              - exactly, VS Code is going to take over the world with it's knowledge about your coding habits

    - It's all about Azure nowadays, anyway.

         - Azure has terrible customer support, we should use bare metal.

I've trained a Transformer encoder-decoder model (this was slightly before GPT2 came out) to generate HN comments from titles. There is a demo running at https://hncynic.leod.org

It's also used for the comments at robots.thoughtbot.com

I want a GPT-3 trained on HN - if it can look at your title and generate basically your content/comment, don’t post it. Similarly, if it can look at the comment you’re replying to and generate essentially your reply (let’s say in top-5 most probable), don’t post it.

Edit, love the irony that both responses are about GPT. Point proven?


Nice. I just sort of assumed early on my comments were training some future AI, and I hope that in some small way I have been able to moderate some of its stupider urges.

A version where you can turn knobs of flavored contributors would be pretty funny. I know my comment style is easily identifiable and reproducable, and it encodes a certain type of logical conjugation, albeit biased with some principles and trigger topics, and I think there is enough material on HN that there may be such a thing as a distinct, motohagiographic lens. :)


> Can't be that hard to train GPT on HN comments.

Yes, which will produce things that are stylistically similar to HN comments, but without any connection to external reality beyond the training data and prompt.

That might provide believable comments, but not things likely to be treated as high-quality ones, and not virtual posters that respond well to things like moderation warnings from dang.


Now every AI article has a comment from someone being suspicious that a comment was generated.

[This comment, like every other comment on HN, was generated by GPT-3. You're the only human here.]


I read this, thought "This looks like it was written by GPT." Checked the profile:

> Real Edwin exiled by HN moderaters, PC edwin has the same opinions but translated to silicon valley sensibilities using ChatGPT.

Which makes me feel like I can recognize AI comments, but no way to know how many I fail to spot.


The other way to look at it is that HN comments are indistinguishable from GPT-3 generated sentences. Hell, even a standard Markov chain would suffice.

Markov-chain generators are extremely lacking in long-term coherency. They rarely even make complete sentences, much less stay on topic! They were not convincing at all-- and many of the GPT-2 samples are as "human-like" as average internet comments.

Conjecture: GPT-2 trained on reddit comments could pass a "comment turing test", where the average person couldn't distinguish whether a comment is bot or human with better than, say, 60% accuracy.

next

Legal | privacy