Hacker News Reader

Comment by xnx | Hacker News Reader

If X wanted to be a real pot-stirrer they'd enable commenting on ads all over the web: Google, Instagram, Facebook, news sites, etc. There should be enough unique identifiers in an ad to make this possible. This is a service I might pay for. A bonus side-effect for X is the browser extension would also capture a bunch of real user ad impression and click data.

View on HN

This is something I've been meaning to do for like 20 years. Install the plugin and you can comment on any DOM element on the Internet. YouTube comments disabled? Comment anyway. Dislike an ad? Comment on it. Banned from Reddit? Reply anyway!

Collaborative web annotation extensions have been done, they've never taken off. Now that mobile has eaten the web it seems even less likely that the idea could work.

Facilitating discussion of arbitrary webpages, though, has been stupendously successful, in the form of social media. I think that’s the problem with the extension idea, it’s social media that doesn’t realize it’s social media, so doesn’t pay enough attention to the make-or-breaks of social media like content quality, drawing people good for content quality, spam prevention, moderation, and circles of shared interest. Getting HN-like commentary on every webpage is interesting, what you get if you just turn on WordPress comments then do nothing, isn’t.

Social media is also good for discovering the original content. Like, I would never find half the blogs I read on HN, but they're great.

I think I came up with this idea in the world where you literally browsed the web. Like go to Yahoo, see what new sites there were, and read 'em ;) That is ... not what we do anymore.

The only way I can imagine it working would be an extension that just hunts down pre-existing Reddit (or similar) threads for a particular URL.

Oddly, this idea seems to share some of the same problems of the "Metaverse" idea. Facebook's Metaverse is a videogame that doesn't realize it's a videogame, it's a centralized service that you're trying to layer on top of the internet, and it ends up being totally empty and with no community, because it's for "everyone", which means it's for nobody. Discord and Reddit both understood the problem and partition themselves into user-moderated interest-oriented spaces; Facebook's idea of a metaverse doesn't do this, and a pan-web comment system sort of can't by design.

Why the Lucky Stiff had one for a while. It was fun, but it worked by putting in his server as your DNS, and it (from what I recall) man-in-the-middled the entire web to insert comment blocks on every page. I don’t think that would work anymore, and I imagine it was a nightmare to keep running.

I refuse to believe that Google's pathetic 10+ year old excuse for blocking extensions on mobile (performance) can hold forever.

Mobile users are the ones by far the most in need of agency. Google saying no here is a top existential threat to their dominance. Hopefully now that Firefox is starting to re-allow webextensions on mobile broadly, pressure can mount. Keep trying folks!

Beyond "can mobile run extensions", I think the main thing missing so far in collaborative web extensions is that few are actually interoperable in any significant way. They're all isolated services. That's helpful for content discovery, but I don't think a centralized service is going to cut it for letting a new depth to the web really form.

I also would say that Hypothesis has been a fairly successful fairly great collaborative web extension. I don't often run into other people commenting on sites or articles or comments, but it happens, and the effort in general is well loved. It somewhat uses standard protocols, but is designed for their service. https://web.hypothes.is/

Mobile absolutely can use extensions. I run them on kiwi. A subset of what's on my desktop, but extensions nonetheless. Every time I wind up accidentally opening chrome, it's eye opening just how awful mobile web is, particularly how many ads you see

I'd love for one to take off.

You might be right that they wouldn't take off even if they would work really well.

But I'd like to think they could :)

With a good UI, a frictionless login, clear personal data separation and a sufficiently widespread browser extension platform available on mobile, this would be awesome...

Maybe Safari would be a good first target with their WebExtension support.

There is of course also the technical problem of dynamic rendering and unpredictable DOM structure.

And the whole problem of reaching a social media user base and providing valuable interaction tools to them.

Especially when considering ads, the main question is how to locate these DOM elements — for example, the filter lists shipped with uBlock origin do a great job here.

I'm wondering if maybe having more flexible locators available could help identify DOM subtrees by their content instead of a CSS selector.

Similar to the ones available in e2e testing frameworks, but maybe even more powerful things could be conceived, or matching snapshots etc...

Could certainly imagine this taking off, but imagination is not reality :)

I will posit that the main problem with the idea is that when I'm browsing some page I probably don't want to see comments from all the people that have been banned from that page, or who couldn't be bothered to make an account. If I like a page I probably agree with their moderation decisions more than not, and being able to browse a big pile of rejected comments doesn't interest me at all. And even if there are people there commenting just because it's easier than making an account, that's going to facilitate very low-effort drive-by comments from people who are coming from somewhere else on the web.

And if people want to talk about some page that doesn't have comments, that's what Twitter and Reddit, historically, are for. The hard parts are not technical, they're social.

I remember gab did this for web pages but google/mozilla banned the browser extension for some reason.

I always thought it was a cool idea to be able to see comments left by other people though.

Hypothesis might come closest: https://web.hypothes.is/

To make it work, for each page I visit the plug-in will make a request to your server to load comments. That means you're getting a full record of where I go on the web. Huge privacy issue. Chances of me installing such a plug-in: 0.000.

Maybe it's possible to come up with some architecture that anonymizes the requests using a hash of the URL+XPATH and some proxies; not obvious to me offhand.

Definitely not for everyone, but you might be surprised what extensions people install just for the chance of getting a coupon (often expired or invalid) for random ecommerce sites.

Don't send every url, click a button to check for comments in urls you are interested in and aren't worried about anonymity on. If you are leaving "likes" or a comment anyway you are revealing you are at least aware of that url in the first place.

Could do it similar to how sponsorblock works, where it sends a hash that returns a pool of videos, and the extension then filters down to the one that you're actually looking at, locally

Maybe we could store the comments in IPFS or something. You don't want to be tracked, and I don't want to host the ASCII art CSAM that some jackass will probably upload.

Unless millions of people install the plugin, that sort of thing would more or less have the same impact as making those comments out loud in an empty room.

I think you'd be surprised how much reach a small niche community can have. Homestucks used to basically own an extension that let you doodle on pages. Even other webcomics ended up with homestuck doodles instead of their own. could not have been more than a couple hundred users but I ran across those things everywhere I went.

This idea has been done a few times I think and it always has the "empty room" problem. I'd also submit that it doesn't solve the moderation problem either, just pushes it to another entity, which can either 1) choose not to moderate, which will drive anyone off who doesn't enjoy having the worst comments being littered everywhere they browse, or 2) choose to moderate, except now they're trying to moderate something the size of the entire web.

Plus, ranting to yourself doesn't increase the attack surface of your browser.

Now that's a very good point, at least unless you have some fancy AI system that turns ranting into terrible cod- oh that's ChatGPT.

Could bootstrap the process by aggregating/summarizing existing comments/sentiment from around the web. This is much more feasible in the era of LLMs.

You need a way to track the sequence of comments. Not the comments, but the sequence of them, in a sha form. Some kind of ledger is the first thing that comes to mind.

Just don't say the bad word

Some kind of square shaped rope data structure

Without moderation, it would quickly become a shit show once some user threshold is reached.

(Prerequisite: company offering moderation-as-a-service?)

I can't imagine there'd be many people aware enough of advertising to use this who would not rather just use an adblocker.

To some, it might be worth turning off adblocking for the humor appeal. It might even be a way of making more interesting advertising. For example, the ad could be "butter, butter, butter" and the ad comments are "Parkay."

Starting in 1973, a long-running advertising campaign was introduced for Parkay featuring a mechanically animated "talking tub" of the product. A typical ad depicted a sort of humorous verbal sparring match between a container of Parkay repeatedly saying "Butter" and a human actor insisting that it is actually Parkay. When the spokesperson tastes the product, they conclude that its creamy texture and buttery taste mean it really is butter, to which the container replies, "Parkay."

https://en.wikipedia.org/wiki/Parkay

It seems like they are going the exact opposite direction: https://mashable.com/article/twitter-x-new-clickbait-ad-form...

With all due respect, this idea seems like the preamble for "How to stop all companies from advertising on your platform". Why would any company pay Twitter for this privilege?

It makes the advertising experience more trustworthy, as weird as it sounds. You basically cannot trust any ad (yes, I know HN readers don't trust ads anyways but not so for everyone else), and even tons of normal users have grown skeptical of Instagram feed ads for example.

Honest advertisers won't be affected, because it's honestly not that easy to get the notes up on your post. It needs a very heavy majority of votes.

> Honest advertisers won't be affected, because it's honestly not that easy to get the notes up on your post. It needs a very heavy majority of votes.

Oof. With Amazon reviews as our working model, imagine how long it takes a large restaurant chain to lie about a competitor.

... you are right. it will inevitably turn to shit, but I'll enjoy it while it lasts.

The only thing that makes me hopeful is that not all accounts can vote, and that the notes only show up with a vast majority of votes. Amazon on the other hand is easier to game just because you usually need to have purchased or received the product to review it (Meaning the seller can easily bribe buyers and even fake them).

No company would pay Twitter for a comment section to be shown on their ads just as no company pays for an adblocker to hide their ads. A hypothetical service would be paid for by the commenters.

This idea goes back to Xanadu, with the first implementation I know of being Tribe, from the 90s. IIRC the web site owners complained it was somehow unfair.

"I hate sunshine, it shows what is really going on!"

- Some advertiser, somewhere, without a doubt

I remember Third Voice, even installed it iirc: https://en.m.wikipedia.org/wiki/Third_Voice

That would inevitably spiral into spam in ad comments.

This is a great idea. I think comments on ads would be tops. It would look like the comment section of Yelp meets HckrNews, maybe even legal action, but context of ads would be so helpful!

Gab did that, didn't they?

Reply on news.ycombinator.com