> and the availability of good enough open weight models for their purposes
it is childish to believe Chinese are going to give up profits to provide such open weight models forever. the whole idea of being "open" is not compatible with the Chinese culture.
This almost racist read of other cultures has always seemed so bizarre to me. Even moreso when said as an argument on the side of completely closed competitors, some of which are outright seeking to ban open weights.
Chinese companies will continue to provide open weight models as long as it is profitable to do so. Chinese companies are on the more open end in many other industries despite the lack of meaningful foreign competition (for one, 3d printing) so there's plenty of reason to be optimistic as far as I'm concerned.
give yourself a break. F-35 doesn't even have long range a2a missiles comparable to PL-17 in service. its radar is 2nd class at best, a full generation behind the Chinese counterpart and heavily rely on Chinese rare earth.
it is 2026 now, partially de-industrialized US no longer has the best weapons in the world. that is why it is struggling to even beat Iran.
The radar might be second class (not sure I would go that far, it's not a Thales radar but it's still good), but the plane is built with EW in mind, and that makes it useful.
What makes it bad is how hard the cooling is, its range without range extenders (that break stealth) and the maintenance.
Yeah, I wonder why the F-35 doesn't have a couple of 20 foot long missiles hanging off the side of it. I truly can't think of a single reason.
Clearly the Americans are just dumb. Or, maybe, just maybe, a stealth strike fighter uses an internal weapons bay? While the planes flying with the PL-17s strapped to it have the radar cross section of a city bus?
Because that was 1985, and this is now. Entire supply chains, input costs, planning, etc. have been built around the assumption of the JPY trading in some sort of range.
Of course that will change over time, if it hasn't already. The Japanese government's messaging about this has been that they're not really worried about the currency weakening (after all they're massive net exporters! it should be a good thing), but rather the speed at which it's happening.
Many don't though, and even if they all did, you can't hedge forever. Hedges are also rolling, so as some hedges expire the companies need to set up new hedges, which are at a worse FX rate. So the hedges lessen the impact but they aren't perfect, otherwise they would not be called 'hedges'.
USD is falling vs euro so EURJPY is even more bleak. USD in general during those times was way stronger than it is today so those numbers are meaningless.
Because you cherry-picked 1985. It went below 160:1 in 1986 and hasn't been back since, until now. So it's the weakest it's been in literally 40 years; that's at least potentially a big deal.
will be great fun if one M5 Ultra with 512GB memory at 1.2T bandwidth capable of doing 3x smallish local model inferencing each at Opus 4.5 level of intelligence.
The memory bandwidth and size seems to be there, but what is the tokens per sec on like a qwen model? And you can basically do 3x opus 4.5 on the $100 a month claude plan. Your payback will be near infinity years after electricity.
you are assuming the idea, the data, the code being touched by opus worths nothing.
how about measuring returns in the sense that I no longer need to share my flagship idea with some random 3rd party just because it hosts the LLM I am using?
reply