My guess would be because the "we're trying to kill her" thing is an extreme aberration of the regular order of things so as soon as the narrow window where that was somehow okay is closed the default practice of trying to help people kicks back in. Especially once she's in the hands of actual medical practitioners who take their oathes seriously.

It made both of us richer. Spare us this Trumpian zero-sum view, it's normal for both trading partners to benefit.

It wouldn't have been if it was reliable.

Fundamentally, trade happens when you've got something that you don't want as much as your neighbour does, but he has something that you want more than he does. The result is that each side gets something they wanted more.

We can build a pipeline a lot faster than we can build refineries. By all means do both, but the refineries won't be online for many years after we're already selling the unrefined stuff.

Because trade is good for everyone involved? That's something the Americans have completely lost the plot on.

Your avatar image checks out. :)

Joke's on them, I'm almost done training FaceBot 9000.

Granted, it's a 1.2B model and the training data is entirely My Little Pony fanfics. It's insane and thinks it's Princess Luna. But it's going to dominate this benchmark.

Ironically, the guys who have tried so far have all been right-wing nutballs as well.

What emergency? Just generally an emergency?

That's not how supply and demand works. When a market is out of economic equilibrium like this, with demand sharply rising, that is literally what drives competitors to arise and to try to increase supply to match. This is fertile ground for exactly that outcome to happen.

Heh, I downloaded that one just recently, I read that it was good at natural prose and I've been working on a little pet project to make a framework for auto-writing short stories based on a simple premise. Haven't tested it extensively yet though.

Sadly they don't seem to have 36B-A3B models on their roadmap any more, they didn't do one for 3.8 either. I agree it was a nice sweet spot between speed and capability, I still use the 3.6 version of 36B-A3B for larger-scale local work. Maybe someone else will aim for that. Or Qwen4 will do something new with the architecture that makes it unnecessary.