I have recently cancelled my subscription with them. I have learnt that their setup is not HA (because their outage took my internet with it for a bit) and you can’t make it HA either, you get assigned to 1 endpoint and if that’s down your DNS is gone, even when they have loads of other working endpoints around you (and their suggestion to this is to add a different DNS as a fallback). Frustrating, as I did like their service for the price.
Kind of even more annoying was that while I was their customer I couldn’t get through their AI support bot to a person to just raise this as a problem and get a proper response, but as soon as I cancelled an actual human reached out to ask why.




The claim that you have (sufficiently) reviewed this change by change is insane at around 3k lines of code in a week.
I gave it the benefit of the doubt, and in about a cursory glance at 3-4 files it’s obvious that this is awful quality - and potentially malicious with intent too (join random communities/groups to post your ads), bordering illegal (masquerade a bot as a real person). Some of the logic doesn’t even do what the prompt said (e.g. don’t post in groups that say no self-promo in the rules, it finds the rules, then ignores them).
You are a great example why AI should not exist. It is frankly a disgusting project from intent to delivery.