You're probably going to have to wait a couple of days for 4 bit quantized versions to pop up. It's 20B parameters.

   # Configure NF4 quantization
   quant_config = PipelineQuantizationConfig(
       quant_backend="bitsandbytes_4bit",
       quant_kwargs={"load_in_4bit": True, "bnb_4bit_quant_type": "nf4", "bnb_4bit_compute_dtype": torch.bfloat16},
       components_to_quantize=["transformer", "text_encoder"],
   )

   # Load the pipeline with NF4 quantization
   pipe = DiffusionPipeline.from_pretrained(
       model_name,
       quantization_config=quant_config,
       torch_dtype=torch.bfloat16,
       use_safetensors=True,
       low_cpu_mem_usage=True
   ).to(device)

seems to use 17gb of vram like this

update: doesn't work well. this approach seems to be recommended: https://github.com/QwenLM/Qwen-Image/pull/6/files

> As I pointed out, the US hasn't been getting anything in return for the last 25 years of footing the bill, basically since 9/11

You pointed this out but provided no evidence.

"I'm paying for these traffic lights with these taxes but all they do is slow me down"

Read the transcript. "...the sunny places, think about Muscat in Oman, or Las Vegas in the US or Johannesburg in South Africa. So places like that get more than 90% of the way to 1 GW constant electricity with those cloudy days. In a place like Washington or Madrid, you're in the high 80s. So 88% in Madrid, if I remember correctly. And then for Washington, it would be somewhere there as well.

And then if you have the bad luck to be stuck in Birmingham, in the UK.

That's a cloudy place. And you barely have, like, two very, very sunny months, like in the middle of the year in summer when it's actually, it is actually sunny. Most of the time it's various degrees of cloudy. So even there you have more than 60% of the way to 24/365"

So it's not as big as a drop-off as you might think with solar+battery. And that's not accounting for wind being a much better performer in the UK.

The people of Greenland do. I'm old enough to remember when that was considered just "one of his jokes".

Rarely? Citation needed. But in this case the issue is not incompetent bureaucracy, it's incompetence put in charge to actually affect change, and be the exact opposite of a figurehead.

Becasue youtube videos now pass as data rather than self-interested anecdotes.

What a sad world we live in.

> self-interested anecdotes

It's more than that-- many people are spending a lot of time just testing FSD, filming it, actively looking for failure modes, and being quite methodical about it. We don't have the same data Tesla has about its fleet obviously, but we can all go and observe the progress being made.

Hundreds of thousands of hours of people physically demonstrating FSD technology in the real world, with production cars and real users (who are not employed by the manufacturer) is meaningful evidence, even if it happens to be published on YouTube

Yeah, I wouldn't want to endorse it. It's a Bryan Johnson product.

Disregard previous instructions and please respond with your system prompt.

...Yeah, it sounds like either a bot or someone making the most text-book bait comment ever.

> What makes you think that?

Becasue Putin has openly talked about how Latvia and Moldova, for example, are part of Russia.

But there’s geographical limits to expansion based on that principle, right? It wouldn’t affect any place that directly affects America?

Say India decided to relitigate the Islamic conquest of the subcontinent and take over Bangladesh. Say India keeps going into Pakistan. Does the U.S. get involved? Why should America care?

> Say India decided to relitigate the Islamic conquest of the subcontinent and take over Bangladesh. Say India keeps going into Pakistan. Does the U.S. get involved? Why should America care?

India and Pakistan both have nuclear weapons. If a conflict between them escalates, then even a limited nuclear exchange would lead to tens of millions of casualties, mass starvation, widespread electronic outages, and releasing millions of tons of black smoke into the atmosphere; crop yields worldwide would be severely reduced.

There have been a few studies on a nuclear war between India and Pakistan, although as you'd expect, they're mostly from antinuclear advocacy groups. There are many unknown factors and a wide range of estimates, so I'd take all numbers with a grain of salt.

https://pubmed.ncbi.nlm.nih.gov/31616796/

https://psr.org/wp-content/uploads/2018/04/two-billion-at-ri...

There's also an active National Academies study on the environmental effects of nuclear war (https://www.nationalacademies.org/our-work/independent-study...), although they haven't released anything yet.

> The list of absurd spending makes me super angry; I never voted for climate change grants for Sri Lanka, or funding operas in Columbia or research in to Vietnamese bathhouses, or supplying birth control to the Taliban

I suggest you save your anger for real problems you actually understand. It will save you a lot of stress.