I added a real-time chat to my blog, people used it to attack me (en.andros.dev)

I'll have to be honest here man, yeah no shit. I can't think of any good reason of why you would add something like this. A standard comment section is typically bad enough, and at least you're able to moderate that. An anonymous real-time chat embedded within your own blog? That's just asking for issues.

The chat is still there in the corner of this post, with people typing nasty shit into it.

Well have fun delving into programmatic censorship I guess. Or just take it out. If you take the former path and post about trying to fix it here then I'm sure you'll get a lot more free testing of your attempts to enforce civility through regex or whatever. Enjoy learning about the Scunthorpe problem.

Well, you've just painted a big target on your blog. This is the internet, after all.

"Every input is hostile until proven otherwise": you said it yourself. One of those sad things that somebody new learns on the Internet every day.

lol hate speech (fun speech lbh) drives traffic

TL;DR: this is a toy example, so it doesn't require BEAM VM

Hard to take this seriously without the loadtest with the the large number of concurrent users.

For a number of users concurrently reading a long-tail blog post - you will get good results even using DotCom era HTTP polling.

---

See:

Django LiveView vs Phoenix LiveView: a real benchmark

https://en.andros.dev/blog/80134668/django-liveview-vs-phoen...

I was watching porn recently and the particular site I was on had recently added a real time chat widget too! Aside from scam bots trying to lure users off onto telegram, the last few messages were guys talking about Monty Python's The Life of Brian.

So that’s what’s beneath the usual HN tech bro libertarian veneer.

Oh my bad are yall under the impression that HN is not the majority of the traffic right now lmao

The homepage design is nice, and I think anonymous chat is fun too. Even though the programming topic probably has a pretty clear target audience, I'm surprised that kind of hateful chat still shows up.

The obvious solution is shutting it down it does offer nothing.

Short of that shadow ban everyone everyone sees their own messages, and add a few fake ones every now and then.

This should come as a surprise to absolutely nobody.

But hey traffic is traffic I guess.

Reminds me of Maria Abramovic Rythym 0 performance art.

1: https://www.thecrimson.com/article/2023/3/30/maria-abramovic...

What if he filled the chat with bots and is using all of this as stealth marketing?

[dead]

This sounds like the kind of feature that ends up being quick to initially implement but takes up way more engineering time to maintain than everything else combined. The author concedes at the end that most genuine uses of the feature are people just saying hello. So maybe the solution here is to disallow unrestricted input and just have a few buttons with fixed things to say.

I’d suggest one additional heuristic: once your bad word detector has triggered, shut down the chat for 10 minutes. That way waves of assholes don’t monopolize the chat, and when they move on to the next target the functionality returns to your site automatically.

You just got yourself a product right there if it works and can market it as Roast me Blog with just trigger filters in your chat to make it healthy roast in place of abuse and highlight top roasters where people could upvote right there in your chat. Negativity turned into positive healthy banter ..

The social web has become the adversarial web, unfortunately. You have to have a huge squelch knob if you want any meaningful signal now.

According to dead internet theory, these might just as well be bots.

Content moderation and social anonymous behaviour aside, your real-time chat is really poorly implemented UX wise.

I saw the same message appearing more than once, I never saw my own message appear, it has a weird lag/delay feel to it. I didn't really enjoy that experience at all.

Well, Django sucks for real-time chat

Humans will really take any chance they can to troll, I would definitely get more moderation

From the earliest days of websites allowing text input from users, this has been a thing. The same with any website that allows you to draw things, it will quickly show people drawing crude (in content not skill) objects. This is just human nature. Once people realized you could use these forms as attack vectors, we were off to the races. Now that bots can do it for you, I'd only imagine the time before first bot using the form is in minutes. I have seen all sorts of things suggested as workarounds to mitigate bot form submissions, but eventually, you will be spammed at the least with the feature. Adding something like a bot chat, of course people are going to be abusive to it much more than a simple form field.

For someone to be surprised by this today suggests to me that the person is really really new to managing a website.

Next we're going to hear a complaint from Mark Zuckerberg about how people are using Facebook and LLaMa to attack Meta.

Obviously insults and attacks are unreasonable, but I think the tag injection etc should be an expected one if you’re posting on HN right? Like somebody is going to try it for kicks and I wouldn’t say it’s even necessarily malicious