After working on our open-source Domain Monitor project, I started thinking about building another small tool and I’d really appreciate some input from the community here before going too far with it.

From my understanding the encryption process happens on the memory controller entirely and is Transparent to the CPU.

I've been running self-hosted Llama models on my own VPS for about a year now. Honestly the biggest factor is RAM – you need at least 16GB for anything useful, 32GB+ if you want decent speed.

I really don't see the future evolution of AI dependent on GPU and CPU ( x86 based ), sincerely, it is like Crypto, as the models get more and more complex, and they do, you will need ASIC's very soon

I get that it's still a valid concern. I had no idea it's possible to perform it also on soldered RAM. Obviously I have no experience with performing that kinds of attacks, but last time I heard about

Unfortunately all that matters is your own jurisdiction and I have no idea where you live.

Not all threat models that FDE is meant to protect against involve being stuck in the same room as an Mexican cartel who will stop at nothing to get at your data. That's like saying that it's pointles

in my own production, 4x3090 x 2. just some coder models along with it's embedded/reranker stack. being used for ~40 ish people.

Would need to be on your own IPs / network for it to be mostly usable. Digital Ocean, Vultr, Hetzner, and other big player's IPs are all mostly already seen as high risk IPs for most websites. Just re

Thank you very much for the insight. As you know, I'm trying to measure the viability and fiscal pressure experienced by hyperscaler, as they pushed hard for AI LLM, which caused the ramflation. Large

It's just general security if disk/s went somewhere and I don't own the hardware. I just rent it.

If you looking for subnets what start with 102, we own few subnets.

The sentence you wrote on its own for emphasis? Don't flip flop, stick by whatever you're saying rather than blaming the reader for reading what you wrote. FFS

Well I mean, also they could log/hack their own NoVNC or whatever, but all these cases are so hard to pull for a provider here on LET and most other providers, which makes it a not viable option. They

I used to rent one at Hetzner for app development, which worked well.

That sounds pretty HighEnd to me (Image)

Hello, I gonna try be straightforward, to the point no AI slop (Image)

There used to be an open issue on the github I believe where @jimaek would review requests for geoip updates. But I can't find it anymore.

Your example doesn't really make LUKS a privacy illusion and even when gov entities come to your door, you can just not decrypt it (Image)

I wanted to test how this affects the VPS servers as well.