I couldn’t make it work whatever I did, whichever instance I used it seemed to get rate limited after a while or showing weird results…

I guess getting a flip phone and stick to an eBook reader would be a good solution

The AI thing is very cool, I think something like that exists, agentic browsers.

But I am scared this would be just the next abstraction, of this chains of abstractions… Corporations are already using AI to profile you even further, the internet will definitely adapt under this pressure, and I believe that in a few years agentic browsers will just become the new norm.

Search engine at first were more objective, now people have learnt to play the game of SEO to attract views, search engines have started to show targeted results, and stuff like Searx came out, or Yaci, claiming to get back a more objective web. There always have been ways to filter out or to try being more objective, but I think the evidence have shown that as a social momentum, this stuff doesn’t work.

Yeah Yaci (self hosted crawler) is a great project, but it’s stagnant, and the prevalence is still a shitty google or bing searche engine, and this is true for other aspects of the web.

Social media? There were more independent social medias while centralized stuff was rampant, now there’s the fediverse and decentralization. Which is super super beautiful, but most people are just unaware. The social momentum is not saying that we are going towards a world where every little server will be connected to other little servers and decide in which parts integrate one another, that would be great, and I’d love to see that, but it’s simply not where we are currently going as a society.

Now it’s the turn of AI, it can be a helpful tool for a while to avoid it all, stuff like agentic browsers can give us some freedom for a while when they will be actually usable and reliable, but in that time the web will have evolved again and pheraps we’ll need to take into account new ways to defend ourselves or to look through the bushes.

It’s a never ending hide and seek unless something really big changes. Linux, free software, open source is all great, but we are continuously pushed towarbalance mainstream in some way or another. And most people live the mainstream, not in the alternative, despite the alternative being objectively better. It is just unsupported by our culture.

It’s an abstraction built on another abstraction built on another abstraction… And the web is just the most clear example of that, I mean the very languages in which the web is built are an example itself: JS (which already is high level)>React>Next. You see? Abstraction on abstraction.

But when will we stop to play games and just stay in the present? Focusing on the core of things?

Do we strive to get to a sort of technological ecstatic point in which all will actually be clear? A sort of technological philosopher stone? And the way to do that is through collection of loads and loads of human data?

My perspective on this is quite pessimistic, because it’s a form of cruel optimism to say that one can solve this problem individually. To change this would require a coordination of consumers, programmers and people revolving around all things of the internet to fix it, unless we assume that AI is somehow sentient and can be better at solving our problems than we do, which I do not exclude: faster and better at looking and processing novelty than we are.

But that will mean that us, as humans, will just be obsolete.

I always come to the conclusion that the web maybe it’s not worth getting used as it is right now, and maybe to feel good we should stop trying to relate to machines and instead just living our own biological needs… Focusing on beings which we can understand better… Living in the present… And stop running, whether it means running away, or towards. Rejecting culture and just staying in our own spaces, cultivating simplicity and balance.

Sorry for the philosophycal rant lmao, I guess this was just more than a technical problem for me lmao, but thanks for your answer!

Use a reverse proxy to proxy everything through https, then you can install how many services you want. Caddy is super simple, you can reverse proxy with just 1 line.

For calendar and contacts (caldav, cardav) Baikal is extremely easy to install and use. And pretty minimal.

Slowly switch to Linux devices, or dumb phone, and learning to adapt happily living without some features, but also without control and censorship :)

As far as I knew reverse proxies could only reverse proxy stuff coming in from 443 or 80, I didn’t know they could listen other ports as well!

Main reason why I was using a reverse proxy at first is because I had everything behind cloudflare, and cloudflare can only proxy and give you an SSL encryption for stuff that goes through 443, so I could make Caddy listen to 443 and then forward to interested ports.

But this leaves out everything that needs to go in some other places than 443, and requires its own standalone ssl certificate, which is a bit cumbersome. Pheraps these can be proxied with other proxies than cloudflare, hopefully giving SSL to everything…

I’m not sure I understood the upstream ssh thing, what do you actually do?

Self hosting IS hard, don’t beat yourself too much because of it… After all you’re trying to serve services for yourself that are usually served by companies with thousands of employees.

A server requires knowledge, maintainance and time, it’s okay to feel frustrated sometimes.

Yes you’re probably right, I definitely have bias and the time spent tryna fix the bug influenced this…

Thanks

I know some basic Rust (currently at chapter 9) and a little bit of JavaScript.

I’m trying to work with headless CMSs and that requires some understanding on how APIs work…

Even tho I wouldn’t want to stick with JS, I don’t really want to dig into frameworks and dependency hells.

But I like the concept and I need to build a site that grabs some data from an external api, so a headless cms would be my choice to grab the data and structure them there in order to be rendered later in something like a static site generator (I’m quite good at Hugo). Or will learn some basic React and try to build a template on my own there…

Having a federated music platform would be great both for people and for artists!

How is the user experience compared to matrix? Is it easy to gateway towards matrix or other services? Can I easily join matrix or other communities servers?

I see a lot of people are now using matrix but not so many xmpp, but yeah it hoggs resources on my server too. Also I feel like it’s still pretty buggy…

Didn’t think about the 2 machines thing. But yeah it looks definitely easier than setting a transparent proxy… But I guess all of this has to be on the same network, I cannot use an external server to which I connect to via wan because at that point the connections would be already need to be unproxied going out right?

But can’t your setup be done on the same machine with a firewall?

Yes DNS and pihole were never thought as content filtering tools

I need to block IPs and unauthorized connections that are not http/https as well, I know about DNS filtering but it’s not what I am trying to achieve.

Yes it’s more something like that, making certain type of content a lot less accessible.

I think it’s all a problem of time: if we have more time to carefully think about what we are doing on our devices, we usually make better choices.

We need better tools to give us more time to actually evaluate and decide.

I’ll make an example: I installed an android device manager which let me set a block timer for each new installed app, that means that whenever I install something new I will have some time to reflect on whether I actually need that new app or not, and most often than not, the answer is no.

Enough focus to read documentation.

That’s really it. If your purpose is just self hosting learning bash could also be helpful. And yeah Linux would be a great choice.

But mostly, if you want to self host an instance of Nextcloud correctly and without having to deal with too many unexpected things, you have to read the documentation and do not rush. Most self hosted stuff isn’t “install and use”, because you’ll be your own server manager, and everything requires attention to be managed.

Docker or not docker you will have to deal with configuration, settings, requirements and updates.

So understanding how to read the docs/search and open github issues and taking time to read everything would be the most important skill for me.

Also writing down what you are doing would indeed be helpful too, in order not to lose track of what you’re doing on your server. (Check out Ansible).

Most apps out there simply need you to know about permissions, systemctl services and package managers.

Try to always find a specific package for your distro for everything you install (eg. .deb for Debian), and have strategies when this is not possible (aka using a Python venv when installing python programs).

Thank you so much for taking the time to answer!

I’m not sure how to get the N from session history, nor how to check my session history…

but this might be some relevant output I’ve found with journalctl -k -b

Nov 21 16:08:18 rpi kernel: usb 2-2.1-port2: cannot reset (err = -110)
Nov 21 16:08:19 rpi kernel: usb 2-2.1-port2: cannot reset (err = -110)
Nov 21 16:08:19 rpi kernel: usb 2-2.1-port2: Cannot enable. Maybe the USB cable is bad?

Nov 21 16:41:57 rpi kernel: I/O error, dev sdb, sector 2466347032 op 0x0:(READ) flags 0x3000 phys_seg 1 prio class 2
Nov 21 16:41:57 rpi kernel: EXT4-fs warning (device sdb1): ext4_dx_find_entry:1796: inode #75497968: lblock 42: comm apache2: error -5 reading directory block
Nov 21 16:41:57 rpi kernel: EXT4-fs error (device sdb1): ext4_journal_check_start:83: comm apache2: Detected aborted journal
Nov 21 16:41:57 rpi kernel: Buffer I/O error on dev sdb1, logical block 0, lost sync page write
Nov 21 16:41:57 rpi kernel: EXT4-fs (sdb1): I/O error while writing superblock
Nov 21 16:41:57 rpi kernel: EXT4-fs (sdb1): Remounting filesystem read-only

The output is from yesterday, when the device stopped working correctly.

I’m not familiar with linux kernel, but I can see there is definitely something wrong…

The HDD (old) is attached to a USB hub (new), I tried switching port of the hub but the same issue happened again, if I try to mount it with sudo mount /mnt/2tb, it says it is already mounted:

mount: /mnt/2tb: /dev/sdb1 already mounted on /mnt/2tb.
       dmesg(1) may have more information after failed mount system call.

sudo dmesg | grep sdb gives back:

[147776.801028] I/O error, dev sdb, sector 77904 op 0x0:(READ) flags 0x3000 phys_seg 1 prio class 2
[147776.815452] EXT4-fs warning (device sdb1): htree_dirblock_to_tree:1083: inode #2: lblock 0: comm ls: error -5 reading directory block
[147796.731734] sdb1: Can't mount, would change RO state

I wonder why we don’t have AI browser extensions that can recognise and obscure possible ads / unwanted content yet

I will add another amazing alternative i’ve found, currently working great: https://www.distractionfreeapps.com/ This was exactly what i was looking for.

Badum tssss