As a result of this filtration my impact as a voter on the site is amplified. Previously I may have seen one mention of topic X per day where now I see twenty. My voting criteria haven't changed but the volume has. I could always have manually searched, but that would be work, and now it isn't.
Is this fine, or would you rather I didn't? Is there an ethical obligation to interact with online services through their available UI, or are we free to make our own?
Maybe the ease of using and abusing such tools will make HN style voting, and the resulting page ordering, uninformative.
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.
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.
Even early on, there were graphical browsers, text mode browsers, tools like wget and curl (not everyone had access to a graphical desktop!). Services like AltaVista allowed one to approach the web in yet another way, spidering the web and surfacing information through full-text-search that couldn't be found through just surfing.
And this is before we get to people with visual impairments using screen-readers or braille terminals. Or folks connecting to the internet over expensive satellite or infrequent wifi who might want to batch their web-fetches and read in their own time.
As far as I'm concerned you're just continuing that tradition with more modern tools. Semantic tools. Carry on. Put your tools on a git site someplace if you'd like to share!
ps. see also the exploits of https://en.wikipedia.org/wiki/James_%22Kibo%22_Parry (he grepped the entire usenet feed for mentions of his name, and allegedly always answered) . Use your powers for good! (see eg altairprime's comment for a definition of 'good' https://news.ycombinator.com/item?id=48260183 )
pps. if I want to telnet info.cern.ch 80 I should still darn well be able to! Uphill, both ways, in driving snow!
ppps. https://line-mode.cern.ch/www/hypertext/WWW/TheProject.html <- eg using a line mode browser.
If you were the owner and operator of the site, what would you want?
Curious why that feels desirable.
I don't particularly desire to do so (anymore), but even if I rolled a d20 to a nat 1 and that critical miss of a color was unconditionally applied to my new car, wholly decoupling any involvement of my desires with the obvious and catastrophic outcomes of a polished mirror finish on other drivers, I still owe it to my society to evaluate whether proceeding to drive that vehicle is ethical or not.
You can't conditional 'is this ethical' on whether or not one desires one of the outcomes ('amplified impact', in this case). Whether or not I desire the specific outcome of the one hundred vehicle crashes my mirror-polished car would cause, the harm of that outcome itself imposes a rather severe ethical burden on the choice to drive that car at all. As is perhaps obvious (or perhaps not!), it would be highly unethical to drive such a vehicle on open-to-the-public surfaces regardless of whatever your desires and motivations may be.
So, similarly, whether or not they desire to disrupt the balance of voting at HN, they are correct to recognize that their plans will, regardless of their interest or lack thereof, impact the site's algorithmic and voting outcomes. That disruption could present an ethical concern, and so they brought that concern to the community that would most be impacted — us, here! — and asked for our input. That is, in general, a sensible approach to working through the ethical concerns of complex outcomes impacting societies. The example provided is exaggerated to prioritize efficacy over plausibility. And it's to the credit of OP that they asked even though this impact would bring the site into closer alignment with their own interests and desires. In a contrasting example of outcomes to such concerns, HN users still post archive.is links to this day even though that site's owner/operator intentionally 'infected' all of our browsers with DDoS malware when visiting their site in order to abuse a single person. 'HN user voluntarily asks for serious ethical guidance from a community impacted by their software' was not something I ever expected to see here. Made my day.
ps. Uh, I'd forgotten something. I did as a kid want a car just like the spaceship from Flight of the Navigator for years, even though society would punish me for it, and had forgotten about that for decades. It took discovering Asimov's Zeroth (Robots and Society) in my early teens for me to understand why I needed to account for a society's needs when evaluating the ethics of my own decisions, rather than just taking for granted the ethical patterns I was raised with. Thanks for the surprise nostalgic memories :D
> 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!)
