floquant

OpenVPN, no. Wireguard, yes.

Idk, I think that even in late stage capitalism demanding constant “growth” when they used to be one of the most visited websites of the entire internet is just delusional.

Look into music assistant if you haven’t already! Makes casting to smart speakers and grouping even nicer than plain hass

A drone can fly at ~100kph which is ~30m/s, giving you a whole half a second to do anything about it after you have detected its presence lol

Whose premises? Release the GPUs and make them cheap again you motherfuckers

Way to miss the point

It could’ve been your brain but that’s apparently passé

Welcome to actual protests

90€/mo to self-host something? Lol

It was, but having offline access makes breaking it a matter of time, especially if you’re the CIA/NSA/DISA/whatever

That is indeed work, yeah

Yeah, Bloomberg though. Not saying that they are great, but it’s among those that Wall Street bros read to decide what to buy and sell

Lots of telling, not a lot of showing sources

Blacklists of what?

Curious about what kind of tracking segment you need to be in to have this surfaced lol

Is someone really arguing that prime numbers are a thing because we have 5 fingers?

Dimensional analysis is really neat

I don’t think anyone can in good conscience say that inches scale to small measurements. Sure, you can use them, but subtracting 8/1000 out of something that is 1 5/8 and noting the final length is just unwieldy.

You don’t need to get super existential about it, imperial was invented before we invented 0 and the decimal system and it shows. Sure the difference in basic units is irrelevant, no one cares what the meter or inch is defined as, the “superior” one is just what you’re used to and have an intuition for. But having units that relate to each other in round ways and can be easilt rescaled to the precision/range required can be nothing but a benefit

Is GLM too slow or not capable enough? It’s quite a big model, if the runtime is setup correctly you should have good quality but getting a throughput similar to cloud models is going to be difficult. Did you give Qwen3.8-27B-FP8 a try?

Also you mentioned Pi as your harness, but what are you using as the inference backend?

Why did you release alpha 6.0.6 then 6.0.1 beta? And how are you on 6.x for an unreleased hobby project started this year anyway? And why are they in different repositories?