Wiki access blocked

[Aistis]

I sometimes find it too exhausting to find any information about certain things Pascal or Lazarus related on the internets. Parts of the wiki is outdated, the information elsewhere is either very old or just plain wrong, and I'm too dumb to understand some parts of the manual that might be a little too sparse at places.

I started using ChatGipity for it and I was surprised at how good it performed. It gives you surprisingly good answers + code examples. If you're at least a bit familiar with programing and Pascal/Lazarus, you should be able to catch the times it might provide you with bs. But I much prefer that to going down the rabbit hole trying to decipher the information I have to manually search and sift through outside of the forums/wiki/manual.

I do like the local manual in Lazarus much more though. So long as it has actual useful descriptions and sometimes even examples in it. Too bad it's still quite sparse in places, but ChatJayPetey seems to fill in the gaps for me in that regard.

In a sense, I guess I'm  part of why this problems exists in the first place.

[Joanna]

We should all get brain chips to code better 

[Tomu]

I took am frequently blocked. Browser Chrome or Firefox

73.5.xxx.xxx

[PascalDragon]

1) Put the Wiki behind Cloudflare, which provides an initial layer of free protections that may be sufficient (this way, your firewall would reject any traffic that doesn't come from Cloudflare, and it would filter out some of the robot traffic; this could also be done through other cloud service providers; some exceptions would be necessary to preserve search engine indexing).

We will NOT use something like Cloudflare. End of discussion.

[n7800]

An interesting option to mitigate DDoS attacks by AI bots is to provide the ability to download the wiki archive. Of course, I understand that this is a naive assumption, since at least they should somehow know about it.

The copy may become outdated, but I don't think a "daily" version is important for AI. Regular publication no more than once a month will be enough. Or even a year, given the activity of changing the wiki.

I wonder why the bots don't use Wayback Machine or Google cache for training? These sites clearly have higher bandwidth.

[PascalDragon]

An interesting option to mitigate DDoS attacks by AI bots is to provide the ability to download the wiki archive. Of course, I understand that this is a naive assumption, since at least they should somehow know about it.

Yes, that's very naive. Those AI scrapers want the latest and greatest and they don't care what costs they incure for others.

I wonder why the bots don't use Wayback Machine or Google cache for training? These sites clearly have higher bandwidth.

They do for sites that block AI. And now that has the effect that Reddit blocked the Internet Archive in turn from archiving its content...

In the end we'll all loose with this AI scraping...

[ad1mt]

I know this is slightly off-topic....
I cannot login to the forum with Chrome. It says "cookies must be enabled", but I have cookies enabled.
I am posting this using firefox.

[PascalDragon]

I know this is slightly off-topic....
I cannot login to the forum with Chrome. It says "cookies must be enabled", but I have cookies enabled.
I am posting this using firefox.

Here is the on-topic thread for that.