I was used to normal fortinet filter (they block porn and warez in the work net which is acceptable) but this new one seems extreme, borderline unusable. You have one chance in the router lifetime to browse an obscure website and then next time it's already in blacklist. I think they pay some kind of subscription for ai autoclassification because there's no way with human support it can be updated this fast

for example, i browse to a completely new website that nobody ever visited. Nobody uses linux, so i go to gnome.org and first it loads because ther router/firewall never knew that URL before. Then in a few seconds gets blocked pending "ai classification", then it gets permanently blocked because that asshole of my boss decided to block every website that offers software downloads

or, for example, i take a linux distro that literally nobody uses, when i use pacman it firstly load packages because i'm the first one to access the download server, then in the middle of downloads the firewall learns what those server does, and blocks them during download (invalid SSL due to MITM that replaces content with the "access blocked" message)

and because the blocklist is infinite, even listening to offline music gives errors when using music APIs

i wonder what's the point of offering "free wifi" like this. It's cheaper and gives a better customer experience to just remove the access points and say "no, we don't have wifi" rather than "we have wifi but you need to login with facebook and you can only use it for whatsapp and to browse our website". I heard the boss was pissed that the upstairs neighbors were "stealing" the wifi, as he saw traffic during night time

They may or may not be used here. You could use LLMs to parse the content of sites being visited by web clients on your network. Then, ask the LLM whether the content includes certain topics or is work related. Based on the results of that, you add/remove the site from a blacklist.

Is this better than just string matching? I would say likely so, though more stochastic in the results. It would let the LLM provide summaries/context of the pages, and not by just confined to specific strings in a list. It might be better ramble to handle context and complexity of the desired outcomes.

For example, there was a paleontology conference at a hotel once that was stuck behind a firewall blacklisting all sites with the string 'bone' in them. Completely ridiculous. The string 'bone' has different meanings based upon context, which simple string matching cannot provide, but an LLM might be better and identifying and acting accordingly.

It's all blocked and I didn't find any single VPN to work (didn't try those with shadowsocks)

Tailscale worked but only if i login using 5G, then after choosing my exit node on my own server, connect to the crippled WiFi

Modern firewalls are very good at sniffing out what a host is doing.