I am running my own nginx proxy manager container to use across multiple applications as a reverse proxy using subdirectories so I can share a single SSL cert. The built in nextcloud app seems to have its own nginx container. When I look at the github page under applicaiton info (GitHub - nextcloud/docker: ⛴ Docker image of Nextcloud) it doesn’t say it comes with nginix but apache. I don’t see a way to specify the apache tag when I install it.
Chat-GPT tells me I can’t use two instances of nginx and there could be problems. I’m not sure if I’m overthinking this and I can use nginx as a web server and then use another instance as the reverse proxy?
As is often the case, either ChatGPT misunderstands the question, or it’s just outright wrong. There’s no inherent reason you can’t put one instance of nginx as a web server behind another instance as a reverse proxy.
AIs have literally zero actual intelligence - they have literally zero common sense and are absolutely unable to make any subjective assessments of the data they are given, and instead simply find patterns that match between your question of whatever information they have been trained on.
So if they are given garbage in, then they provide garbage out. And since they are fed the contents of the internet, which is full of factually inaccurate statements by non-experts, and since they have no way of weighing up the quality of what they are fed, AIs simply regurgitate the rubbish that they have been fed like some social media gossip group.
On top of that, their pattern matching is only vaguely accurate, so they mis-match similar pieces of data and mis-match the question to potential answers which lead to “hallucinations” where they literally make up information and present it as facts.
Nevertheless, the intelligence of these AIs can often far exceed that of the person asking it questions.
Thank you. I’m glad my intuition here was correct. It sounded like general advice, but like most things had way more nuance. I’m sure it could cause issues. Funny enough, the additional research I did with the AI led me to this conclusion as well, but wanted some meatspace validation.
This is what I fear with AI in general. If you’ve already developed some critical thinking skills and know to question it, it can be a great tool. Things can get problematic real quick when you don’t really question it at all. Our education system doesn’t seem to breed critical thinkers anymore, just crowd followers.
My Brother-in-Law is a lawyer who used to be a partner in one of the top 6 law-firms in the UK - but a nastier human being you have yet to meet, and having had a few legal run-ins with him I was able to run circles around him legally despite not being a lawyer myself.
But what drove my opinion of his legal abilities down even further was when his legal opinion had been obtained from ChatGPT.
