Are you saying that you are using LLMs to give yourself a filtered UX on the people and topics you want to respond to, but then still doing your own comments and voting independently from LLMs?

If so, that seems fine. There are plenty of alternative UX tools to get to HN info in different ways. You made another one for yourself, which is A-OK.

But if you are using LLMs to draft comments for you or tell you what you should be up/down voting, that would cross a bad line.

Instead of worrying about the implications on HN, I'd rather worry about the implications on you as a consumer of HN.

I wouldn't be comfortable creating a bubble inside of a bubble, but to each their own.

As long as it's just consuming information from the site and you're essentially forming your own custom filters using it, I don't see an issue.

> Is this fine, or would you rather I didn't?

If you were the owner and operator of the site, what would you want?

As a result of this filtration my impact as a voter on the site is amplified.

Curious why that feels desirable.

We are getting close to this: https://xkcd.com/810/

> Is there an ethical obligation to interact with online services through their available UI, or are we free to make our own?

> As a result of this filtration my impact as a voter on the site is amplified.

Your obligation lies not towards UI, but towards mitigating the normal human instinct to wrap oneself up in a filter-bubble blanket — and in a social context, to be sure you aren’t proselytizing your filter-bubble upon others.

I regularly filter out AI content from HN whether it likes it or not, which is perfectly fine in a void — but I also filter out my votes on that content, rather than participating with it. Not because I’m biased — we’re all biased, or else we wouldn’t vote/flag/report at all! But because it’s not ethical to amplify/squelch things broadly when one has focused tightly.

So, if you’re filtering to exclude a bunch of posts and then disregard them, yeah, that’s what I do without tool assistance. Normal and fine, standard human.

If you’re filtering posts for keywords and then upvoting posts about those keywords without having read them, that’ll earn you a permaban once it’s noticed, because that’s not participation, that’s just SEO boosting votemanip.

If you’re filtering posts for keywords and then reading them and upvoting only the posts you think were meaningfully worth your time to read, then sure, have fun voting the castle, nothing wrong with that.

If you’re filtering comments for keywords and upvoting only those comments that you think are noteworthy amidst all such comments on those keywords, that’s mostly fine, so long as you don’t slip into “yeah that” boosterism of your own viewpoint.

If you’re filtering comments for keywords and replying or upvoting or downvoting or flagging more often than not and more than a couple of times a day, it would be reasonable to label you a proselytizing missionary rather than a genuine site participant. That’s not really a healthy form of participation, whether it’s tool-assisted or not, and I tend to view it as unethical even though — especially because — it’s a provably efficient way to sway the beliefs of the masses without their awareness.

So, that’s the dilemma you face: the more tightly you filter, the more cautious you need to be not to cross into unethical memetic propagation. If you just skip all AI posts then you only have to avoid voting on most AI comments. If you have a fifty keyword reading list that is polymath-diverse, then you don’t need to worry so much; but if it’s fifty words about blockchains, you probably shouldn’t be voting at all. Along those lines, here’s a litmus test; it won’t absolve your concerns, but it will help you get a coarse read. Take your keyword list, shuffle-chunking it into 7-word arrays, and then put them into https://www.datcreativity.com/ (either at the website or using the about-linked GitHub repo to DIY locally). If it grades your wordlist as high creativity, you don’t need to worry much about this. If it grades your list as low creativity, you do need to worry.

Also. Open up a copy of Brin’s “Earth” and read the chapter titled “Holosphere”, pages 253-262. I think it’s directly relevant to your concerns; not least of which because it describes the exact thing you’re making, but also because it digs into the social impacts of that thing in the context. I don’t assume it will conclude your concerns, but that’s rarely an outcomes with ethical dilemmas of definite personal convenience versus possible societal harm.

The final concern to keep in mind — especially when building a filter bubble of any opacity, but even when not, too — is that your participation should be unique to the post you’re participating with. Like: If you have a beef to pick with Canonical, don’t just concern-shop it to every Ubuntu post. If your comment about Canonical could be copy-pasted to any post about Ubuntu, then it’s not a good comment and you probably shouldn’t post it. You’ll end up with your generic comment downweighted in the backend and may earn a ‘stop it or else’ email from the mods if you do so often.

(And whatever you end up deciding, thank you for caring at all about the ethics of this!)

try hermes