Except that they are separated. The clean side (model + implementors) doesnt have the original source code. The dirty side is the training room, which does have original source.
The model that you download from huggingface is basically the notebook passed between the two rooms, and as far as im aware, the model doesnt have the actual training data in it, just enough statistics to “hallucinate” something close.
Uptime Kuma, only really care if it’s up or not. Got about 30 services monitored, sending notifications to Signal.
Thats not a difference. Other clean room projects, for example ReactOS, have source projects that are incompatibly licenced. Its the entire point of clean room reversing, to bypass copyright and licencing.
The training data is on the dirty room side, it doesnt get passed to the clean side.
The clean room concept is violated if raw source code is in the model generated, but to my knowledge its not. There is no way to pull out a chunk of the model, and say “This is a direct copy of main.c from xyz project”. At best, you can ask the LLM to guess what it should be, and it’ll regurgitate something statistically close. I just tried asking my local qwen3.8, and did produce a close summary of the nginx main file, but it wasn’t even close to a 1 to 1 copy.
The clean room would also be violated if you consider the training, model and inference as a singular entity, but that would be tough to argue when the inference can happen on one server, the resulting model isn’t functional in any way, its a big box of data, and the inference can happen on any GPU anywhere else in the world.
Maybe they have shitty distros? Yeah, I don’t get it either tbh.
I haven’t had a failed kernel update in years. Used to happen very occasionally with nvidia drivers, but honestly haven’t hit it in ages even with nvidia drivers.
Clean room reversing is just laundering code, always has been.
Rewriting code into English/GGUF/Klingon/whatever, and then getting someone else to rewrite it back into code is just laundering it. Its been legally accepted for a while, and many projects we take for granted today started that way.
Whether this case is “clean” enough is up to the courts ultimately, but its laundering either way.
It is a good idea. Firmware updates can improve stability and performance. Just because it failed this time doesnt mean that all the other times it silently worked were also bad ideas.
Of course, but the problem is that new users come here, see a sea of repost bots, and then conclude that its a ghost town.
And then they get mirrored to fediverse via bots…
How do they know which trainis will open up to accept the other?
Or use an ide that shows the argument name inline for you. Its 2026, your editor can help you.
Smart home stuff is dominated by Tuya, which is Chinese. Even a lot of the american branded stuff is Tuya rebranded. Phillips hue is Netherlands based, and from a quick look isn’t tuya. IKEA is also not tuya. The rest, I would basically guarantee are all tuya based.
There is nothing nothing to be done. Casual users will literally install wiretaps in their own homes.
You cant help people who dont want or care for the help.
You kinda prove my point. Those who care will sell their bulbs or learn vlans. Everyone else won’t.
Casual users don’t need to setup vlans, because they simply don’t care.
There is no point hand-wringing about this, they don’t care, they don’t need our help for this.
User process makes sense, but login shell is probably limited value. If your service gets pwn’d the attacker will spin up a reverse shell, and that isn’t protected by the login shell. You ideally want to use selinux/apparmor to prevent execution, or containers to limit the available execution environment.
There are nicer ways to say that.
The hosted nextcloud solutions are not cheap. The base plan for librebit is 79EUR per month. After 3 months you’ve paid off a mini PC.
Yes, its more work, and a huge learning curve, but if the goal is to save money, paying 79EUR per month is insane, unless you need the SLA.
I agree that self hosting may not be the right path for OP, but there are valid tangible reasons to do it beyond “self hosting is fun”.
You only need your PC on when others are using the spreadsheet.
Keep an eye out for mini PCs on second hand market places, they can be found for a couple hundred dollars or less, and are perfect for running something like this 24/7. Youll want a fairly decent internet connection though.
You can get free VMs on oracle cloud. Yes, Oracle are a bunch of cunts, but free is free, and you can get a free ARM VM with plenty of RAM that should be adequate for nextcloud.
Make sure you setup https properly.
https://esphome.io/components/serial_proxy/
https://tasmota.github.io/docs/Serial-to-TCP-Bridge/
A esp32 dev board and some soldering might work.
cron/crontab are the general things you want to search for. They are used for running commands on a schedule, which is what you want.
You can also use systemd timers. Pick the one that your OS supports.