Rhaedas

Lemmy and related places are still small enough where a regular name posting can become better known faster than large platforms. My only advice is to just review your posts before first submitting to make sure its message is clear, and if people ask questions about it, then clarify. If you want to engage and discuss things, this is part of it. You’re getting discussion. :)

Most models are going to require CUDA. There are some AMD ones out there, but it’s a totally different math and setup. As for the one I mentioned, it’s a pretty new idea so there are only a few out there, maybe just one (Qwen based). But I did get a 31B model to work on my 12GB, I just had to move from Ollama to llama.cpp to gain the control needed to set the parameters, and fine tune what it put on the CUDA to the max it would take. I had Claude help me along the way.

It’s new enough that there aren’t any good abliterated/uncensored models yet.

16GB is plenty for even older model setups. Now they’ve got a few models designed so you load just parts of the model onto the GPU (Mixture of Experts) and use the CPU for less referenced sections, so you get both reasonable speed and a much more complex model.

Dug your hole and threw out both the ladder and the shovel.

That’s the rub. Local models have some positives, but overall the source for them is a negative. And the worst thing is, we can’t fix that now. We can’t undo what’s been done, we can’t start over with new data that is fair, and we can’t seem to get AI out of where it’s been jammed in.

Agreed. The short is that AI was done wrong in so many ways, for the wrong reasons, and was the wrong direction for the goal they continue to insist they are trying to reach. The science and technology and what it can do, even the worst of it, is fascinating, but this is not what it should have been. Money corrupted yet another thing.

If you have a few things that work better without or don’t need the VPN, like some games, a good VPN will have a tunneling feature to let that one thing through while staying on for everything else.

Make the missing pieces. Various ways to do that, from hand drawing it in if you’re good or the picture there is simple. Do a reverse search to find the puzzle box photo or the original image it came from and use that. AI, if you’re not against that, could probably fill in that pretty well. Or just leave it, chances are depending where the pieces are most people may not even notice.

That’s just a matter of circumstance. How many of them would stay that way if you knew them?

I usually don’t have a problem with persons individually. But as a herd animal, people suck.

The why of mass is “just” more material within the same unit area.

What always gets me is how adding one more of the pieces of matter, specifically another proton, changes the very properties of that atom, sometimes dramatically.

There’s been a few videos on Youtube lately discussing using a particular Qwen model that lets you load only particular expert sections at a time onto the GPU and the rest in RAM. This one was the first I watched (https://www.youtube.com/watch?v=8F_5pdcD3HY), I haven’t tried it, but it makes sense on why it would work.

If they sit down at the table with them, or allowed them to sit there, they are one.

Or tell her she has a mix, and let her decide what she wants to be. I’ve seen plenty of people who look one way but claim roots and heritage from something I didn’t know about. She’s not just white, nor just brown. She’s Violet.

The first chapter of Larry Niven’s Ringworld mentions this. At that point they have had transporter technology for a while, where you step on a pad and instantly you’re somewhere else in the world. Opening up that level of ability to move around meant people mixed as well, and after a few generations everyone tended to be more similar than different. Probably not realistic, as there will still be culture and prejudices that influence pairing for a while, even if everywhere and everyone was just a step away.

“Lowest insurance in the state! Only $10 a month!”

In very small print - “Deductible includes total costs for medical and vehicle”

The rate that water returns to aquifers it was drawn from is very slow. Rainfall from the evaporation is only the first step of a long process. So it’s not contamination, just being used up faster than is reasonable.

I concur, it was faster than a 777. Everything is made with characteristics that fit its purpose.

Looking back at the problem, you say just build more. A 777 carries about 100 tons. The total for shipping crude oil averages around 2.5 billion tons. The world’s existing capacity for shipping oil is around 670 million, a rough number since there are various sizes of tankers. The 2000 777s (if that would even work) would replace one ship, not all those ships (thousands of them). You just don’t understand the scale.

Go ahead and email them, let them know they need to make the planes faster. I’m sure they’ll appreciate your consulting advice. Contact Airbus too, they have the same issue.

Yeah, they’re 2500 orders behind.

Boeing hasn’t built that many yet since it first became available.