BSD was mired in the uncertainty of a lawsuit over some of their code at the time that Linux was getting started, and the FUD around that gave Linux a head start that BSD had up until that point, so you can't infer much about the reasons Linux's early success over BSD through that fog. If Linux had been dealing with the same problem that BSD had instead, BSD almost certainly would be in Linux's place right now.
Before the AI craze, all the GPUs were being bought up by cryptocurrency miners, and I'm not sure that's better. Even as an AI skeptic I think AI is a better use of all this hardware than cryptocurrency.
You're right of course, I'm taking for granted that we've been there and done that. The question is, would we do it again and would that usage really be allowed to expand to soak up whole data centers? We might since I don't think a cynic has made a bad prediction in years..
IMO the most likely way to soak up the extra capacity is actually yet another iteration of AI rather than say, doing productive but boring work with any other techniques for curing disease or something. Still, a crash and a next iteration might be more likely to involve fresh ideas on architecture, or focus on smaller expert models that have less fake results and actually empower users. Right now I think there's a clear bias in research and execution. OFANG does want results, but also wants results that tech giants. Are subsymbolic techniques really the best techniques, or are they just the best at preserving the moat of big-data and big-compute?
I think the meaning was not "You can import used cars without tariffs", but "If you buy used cars already in the country, you don't pay the new tariff, so just don't buy new cars."
It seems shaky trying to apply labels like "low level" or "high level" to Rust, it covers a very broad range of these "levels", just short of C and Zig on the low level end (at least, without becoming substantially less ergonomic), and just short of Haskell at the high level end. Rust definitely gets more attention for its low level characteristics where its safety features set it apart from its peers in that space, but it's still suitable outside that space, just less distinctively suitable, and so less widely discussed.
Some other languages will certainly have more developed ecosystems and frameworks available, but the purpose of things like Dioxus is to provide that. Instead of trying to put things into boxes with labels and deduce from there, you'll get more insight from trying it out and seeing where it works and where it actually struggles.
They had two paths to comply with the law. Silently backdoor the worldwide cloud serving every Apple device, or loudly tell people in the UK they don't get to have security because their government prohibits them. Between these two options, this is clearly "making a stand".
It's not as much "making a stand" as telling a major government that you have substantial seizable assets under their jurisdiction who is a major market you want to be in, that you're not going to do the thing that their laws say you are required to do, but it's hardly simple compliance either, instead of doing what the government wants them to do, they are making sure there is blowback.
Whether to try to fight it in court likely depends on details of case law and the wording of the laws they'd be contesting, I imagine much of the delay in their response to the demand was asking their lawyers how well they think they would fare in court.
This doesn't affect only people in the UK. It allows access to all Apple users' data globally:
> No Heathrow connection necessary. “The law has extraterritorial powers, meaning UK law enforcement would have been able to access the encrypted iCloud data of Apple customers anywhere in the world, including in the US” [1].
As unhelpful as the other commenter suggesting atuin was, this is actually a thing that atuin does well, if you're using a shell it supports, it commits to history immediately.
I'm an atuin user and if this feature exists, I have not seen it and would love to have it. Could you expand on how to get per-directory history as something separate from global shell history?
