Hacker News Reader

Comment by quiet35 | Hacker News Reader

I like the idea and even considered contributing to the list, but this stopped me:

> NAQ (Never Asked Questions)

> My website is on your list!

> Cry about it.

That's quite a suspicious attitude. Clearly the maintainer believes he is infallible. I understand the emotions behind this, but this is not how a public blacklist should be maintained.

View on HN

Yuuup. My personal website has been inaccessible to a few friends, they thought my server was down. It turned out they had some blocklist (not related to AI) installed on their PiHole, and for whatever reason my website was on that list. It is, in fact, to this day, because my request to unblock it went completely unanswered. I still don't know why the website is on the list.

Perhaps it got hacked and was hosting malware without you being aware? They are pretty good at hiding it from the site owner (showing the original website to you, but not to others).

The server is and has been clean the whole time. I don't even run WordPress or anything similar on that server that would be a common hacking target. If it was hacked, I'm pretty sure Google Safe Browsing would be the first to flag the site, not some random PiHole list.

Go to the Adguard GitHub (or use the extension) and report it. And get all your friends to switch to Adguard extension and Adguard Home (Pi Hole alternative) as blockers.

Easylist and its sublist are notorious for being poorly maintained and ignoring issues opened against it. Adguard is much more active in maintaining its lists. Especially Adguard its language blocklists have much, much less breakage and missed ads than Easylist.

>> And get all your friends to switch to Adguard extension and Adguard Home (Pi Hole alternative) as blockers.

Nice of you to slip this "easy" step into your advice. Give me a break!

..?

If you know how to run a Pi Hole, you know how to run Adguard Home. And installing Chromium / Firefox / Safari extensions isn't exactly rocket science.

The crux is in the sentence of yours:

>...all your friends to switch to ...<

:-))

PiHole should err on the side of false negatives, uBO on false positives. Difference being uBO only takes a click to disarm for a site.

Personally I don't want to introduce any chance of my DNS being a problem.

Not any more it seems: https://github.com/alvi-se/ai-ublock-blacklist/commit/7ebaa7...

Probably because there's about the same chance of them being innocent as the "Help I was wrongfully banned by VAC :(((" posts in the Counterstrike community.

Reminder that false positives are not only possible but likely. I remember one instance where you could get people banned by sending them a specific string of characters over chat. Anticheat was scanning the entire contents of RAM looking for it.

These days anticheat software is likely to snap at anything. Who knows what they think of the development tools Hacker News users are likely to have on their computers? They really hate virtual machines for example. There's no telling how they'd react to a debugger or profiler.

Yeah that's what the people love to say on the Steam forums when they've gotten busted in one of its many ban waves.

Both of you can be right.

Never claimed otherwise. Just saying it's a fact of life that every test has false positives and false negatives. The "is this player cheating" test is no exception.

Check out this amazing episode:

https://news.ycombinator.com/item?id=26296339

Dude got so fed up with long loading times he debugged the game and not only discovered the cause but actually fixed it. Billion dollar corporation couldn't be assed to do it.

Gotta wonder if this guy wouldn't have gotten banned by the anticheat for having the audacity to hook into the game with a debugger or something. Only cheaters do that sort of thing right?

Yeah that would be a reasonable thing to ban for. Companies can't afford to audit every single unauthorized tampering with their software to ensure that it's benign. If it results in better cheat detection, far better to have a policy that's unobstrusive and non-applicable to 99.99% of users and something the marginal outliers will understand is a risk.

It's a false positive. He would get banned even though he was not cheating.

Whether it's "reasonable" or not comes down to politics. Optimizing for either false positives or false negatives is a policy decision. Do you punish innocents to ensure you catch every single cheater? Do you let cheaters go to ensure you don't punish innocents?

I don't really intend to discuss the above questions. I'm just pointing out the fact one of those so called complainers in the forums could very well turn out to be one of these false positives. That's what the system is optimized for.

I would add that with this attitude and how new this initiative is, there's very little chance it will still be updated 5 years from now. Really this sort of thing needs to come from Easylist or similar, who have a track record of maintaining these for years.

I don't understand the need for the author to commit the rest of his life to this or start a foundation. It is a good list for now and if its never updated again, that seems fine.

in that case they should just contribute to one of the existing, more established lists. We don't need n+1 standards...

Which lists is open to this kind of contribution?

If a blocklist doesn't get updated it is outdated in a week.

Some tools are useful without updates. A blocklist for AI content farms that are sprouting like crazy is not helpful if it isn't updated.

You forgot:

> A personal list for uBlock Origin

Fork it then!

If the website is not AI slop, presumably they would remove it from the list.

Also seems a bit hypocritical given the screed about how such a list is necessary because the AI content might output hallucinations or damaging content without review.

But if it’s the author’s blocklist that is wrong, unverified, and causing harm to others? Cry about it.

Reply on news.ycombinator.com