3rd Party Developers group

PitaJ

I don't think anything needs to be changed besides the URL in config.json

oplik0

@julian said in Pushing forums in the AI SEO domination era:

Edit: We currently use microdata to present our metadata, so that may need adjusting. Might be when we moved to Harmony that wasn't ported over.

AFAICT it's Google who isn't following the spec here and the original structured data was valid (e.g. schema.org validator doesn't report any errors with it), but I guess Google has some more strict preferences, and in the end what matters is what Google believes is correct (gotta love monopolies).

Hah, didn't I call it when I said that the ride of LLMs and ChatGPT, et al. would mean that actual user-generated content would become all the more valuable? It seems Google is coming to that conclusion too.

Paid API access support when?

oplik0

It seems that Google is pushing for more structured data, with focus on content actually made by humans. They just announced support for DiscussionForumPosting structured data, and along with a "perspectives" filter and Discussions and Forums "category" in results that are trying to highlight social media/forum discussions, it seems their response to AI generated content breaking SEO (see this viral thread from founder of a company actively making the web worse) is promoting user-generated content and push the problem back to social media and forum moderation (see also their updates to quality rater guidelines adding an "experience" metric).

But why does structured data help here? Why only now? Well, I imagine it's also their response to LLMs. They didn't need that much structured data when one could be reasonably sure the free-flowing text was done by humans. So Google focused on analyzing the text themselves and didn't really need the websites to tell them what that is, since what mattered was the content more than who made the content. Now creating new text that looks human to their algorithms is becoming cheap, and trying to automatically moderate the entire internet isn't feasible. So it seems Google realized that they can't trust content providers to provide decent content, and while obviously bots are rampant in user-generated content too, it's not them who have to deal with moderation. Wikipedia and even Reddit seem to be more trustworthy than news articles or blogs for many issues (e.g. product reviews), so they need to adjust to not lose users (even if they are very much part of the issue themselves).

Anyway, NodeBB already implements this - so what's changing is that Google should now be able to pick up this metadata.
I'm saying should because it seems to be slightly broken currently - it's throwing errors because of invalid name data type for NodeBB (which from my understanding is done correctly here and should be serialized into text from the href value), and it seems to not even pick up on this on Discourse even though it implements it too. But I'm assuming these are implementation issues and will either be fixed by Google or worked around on forum software soon.

EDIT: it was fixed on NodeBB side: https://search.google.com/test/rich-results/result?id=wCSb0uOG9XPulXXSxHFZnA

The point is, we're probably going to see resurgence of forums in search results. Probably along with more forum spam, but I hope the increased visibility will be worth the moderation burden. And maybe people won't need to add site:reddit.com to google searches soon

Side note: does this mean web 3.0 might actually somewhat succeed due to SEO benefits? I mean the "semantic web" web 3.0, not the cryptocurrency-based "web3" one.

PitaJ

NodeBB doesn't really support multiple domains, and we recommend you redirect one to the other instead.

PitaJ

What browser are you using? Mobile or desktop? What operating system?

oplik0

I mean, it's important to note that I'm talking about AI as moderation help, not sole moderation. Auto-flagging, or putting in moderation queue instead of making the post public.

And almost regardless of ideology, I hope most people see the need for moderation - at least on the basic spam level. And for most people this quickly escalates to illegal content, and some general content policies (things like obvious harassment, often nudity, etc.). And as the issue of the content that needs moderation gets worse (gotta love AI companies selling the cure for the disease they are themselves creating ) there is a need for new tools in that area.

Ultimately, everyone wants to make a "free speech platform" but the way internet works is utterly incompatible with fully embracing this idea. And sadly, making it work differently may be much worse (and more dystopian) than just moderating.

oplik0

@omega said in OpenAl Plugin:

None of this ChatGPT/AI stuff is at it seems, and it can not be judged or assessed in totality as some technical challenge - The vampire can only come into your house invited.

I kind of agree - I hate that so many people seem to be treating "hallucinations" as some bug that can just be fixed and not part of the fundamental way these models work, and just cherry-pick the good outputs ignoring all the work that goes into getting them (seriously - how the hell do people program using ChatGPT? Like, Copilot is useful for boilerplate and just as a good autocomplete, but for actually writing anything more complex than algorithmic class assignments it takes me more work to get a decent code from a chatbot than just writing it myself).

But a good autocomplete is still worth something, and even the current "AI" can do a lot more. I consider them incredibly useful, but even more overrated, tools. This is why I'm suggesting using it not as a response generator (which will produce a lot of garbage, even if some results will be good and it's an easy way to say you followed a trend).

Also, another thing that both Discourse and NodeBB now have (see chatGPTmoderator plugin) is AI-based moderation, which is actually great - a lot of moderation is already somewhat automated, and this is just automation that works far better with text semantics and not just literal words or obvious usage patterns. It's not perfect by any stretch of imagination, but it can allow for more efficient moderation and that may be all that's needed.

oplik0

@Henryk said in NodeBB + ReactJS:

React, for example, has eaten the world and birthed many alternatives with similar ergonomics, smaller footprints and faster renders. Meanwhile nodebb is using... jQuery and BenchpressJS? Oof.

I mean, NodeBB is slowly moving away from jQuery, but the issue is the plugin ecosystem - removing jQuery would be a very breaking change, much, much bigger than the move from less to scss (since most plugins would still at least work even with their css broken, and most of the time the transition didn't actually require rewriting anything unless one was using more advanced features). I suspect it'll end up moving to something like cash at some point.

And I suspect you're confusing Benchpress with some other templating engine, since (which has both positive and negative aspects) it's basically used only by NodeBB and tailored towards it. It's a tiny and from what I've seen fairly well optimized templating engine - I suspect it'd be faster than react for templates just because it's significantly less complex.

I honestly imagine that if NodeBB was written, or fully rewritten, now it'd be using some framework, likely React/Preact. But the issue is that when you have a long-lived codebase with existing ecosystem the benefits need to be much larger than for the initial choice. Because even with diminishing returns on smaller improvements, the required amount of work makes it hard for a rewrite to beat them, much more so if the changes will require more changes by third party developers too. And you still have to ask yourself if working on DX and performance is worth the opportunity cost, or maybe it's "good enough" and you should focus your efforts elsewhere.

It's annoying, and if I had the time and will to develop forum software from scratch my tech stack would look different than NodeBB, but as with many things it's only easy on the surface and I still prefer the way NodeBB is built to php of Flarum (also - they're an example of an early framework bet that didn't work out ecosystem-wise, since mithril ended with a far smaller community than React and has gone through a few periods when it had years between new releases) or the weirdly dockerized Ruby on Rails w/ ember.js that Discourse uses.
I may not be a big fan of jQuery, but at least even when writing for core I have the choice of just not using it and going with modern JS instead.

oplik0

@crazycells said in Future of Post Sharing Options?:

hi @oplik0 , do you mean something like this "share" button? Actually, this would solve all the problems... Because it opens up all the options including text sharing and it is better than what I proposed...

Yup, it's a pretty nice API, though as I mentioned it still has issues with support on desktops despite being in the works for 7 years now (which is largely because Windows doesn't have good story around sharing, and even after a few minutes of googling and even trying to get bing chat to spit something out I have no idea how the hell you integrate with it when writing a native app. EDIT: after writing this I remembered that it was added when MS was pushing UWP, and searching specifically for UWP share integration actually yielded results. So you can guess why no one is using it [tbf you can now make win32 apps use this API, but only because MS allowed adding UWP components to standard apps, and it takes that to interact with the new API])

oplik0

@baris said in OpenAl Plugin:

Same post-queue rules would apply so not sure how it's different than some user creating an account and posting spam generated by chat-gpt

It stops being their API usage

But yeah, with the UX mentioned here it wouldn't really be abusable, but I feel it'd end up being a novelty rather than something useful.
My main issue is the question of when it'd be more useful to tag the chatbot than use its own interface.

One idea is Q&A forums, where providing an "attempted answer" (especially if it's using data from the forum) and still having the opportunity to see the community clarify or correct that may be useful - but I'd say automatically feeding question topics for a response may be better.

Probably the most useful, but also more abusable, would be to integrate a chatbot with the composer, so that it's useful for making posts, rather than after one makes a post (in part also because my experience with even newer models is that they still need to be very highly guided to do anything even slightly complex, so the feedback loop needs to be faster than posts).

3rd Party Developers

Posts

Member List