> Isn't the whole point of Satoshi's algorithm that those magic numbers it takes to commit a transaction are ever fewer and farther between?
No. I suggest reading the whitepaper. Also c.f. Google search keyword "difficulty adjustment". There are some fundamental misapprehensions in your model.
> The point of Stross' idea was that this ever-increasing scarcity can be provided by astrophysics
With ever-decreasing performance. Bitcoin performance stays constant while its inflation schedule decreases to zero.
> So Bit"coin" is millions of times as secure now as it was a decade or so ago?
Yes
> Wow, it must have had really crappy security back then.
Correct, but no one cared enough to invest the resources required to derail it. Now its security threshold is sufficiently high that it's economically infeasible for anyone to derail it. This is well understood and part of the fundamental economic design of Bitcoin's security model. Once again, I suggest reading the whitepaper. It will clear up most of these questions you're asking.
> If you're saying it was so insecure then that it was pretty much fraud to sell it for real money
Is this a real argument or sort of a joke?
> What do you then say to your equivalent (or yourself) ten years in the future, when they say the same about you now?
This argument will be equally silly ten years from now
> But it still hasn't, has it? And there is no such point in sight, or can you give us an approximate date?
Whenever and if-ever Bitcoin displaces the dollar, yuan, ruble, gold, etc. as the global settlement and reserve asset. I expect its stability to asymptotically improve as it approaches that point, or to exponentially blow up before that happens.
> Within six months is precise enough...
You could make a lot of money if you could predict this, so I invite you to try.
> the point where the actual resource usage exceeded what's reasonable to waste on silly games like this was passed long ago
If you think millions of people are wasting resources on something and you can't really explain why, what's almost certainly happening is that they're buying something you personally don't understand (which I am certain is true in your case). Ideally once you understand how Bitcoin mining works you'll understand the economic security model and this will all make sense.
> lest this fucking Paperclip Maximizer eat the world.
Don't worry, there's not really any compelling reason for Bitcoin-related energy production to consume more than a fraction of a percent of global economic output. It's strictly bounded by the social surplus associated with using Bitcoin over fiat alternatives, which I would estimate at perhaps a few hundred billion dollars a year in current terms. Maybe even just a few tens of billions. So probably no more than 10-50x what it is now. There is no exponential growth problem like with a runaway paperclip maximizer. The Bitcoin network's "utility function" also isn't contrived, unlike paperclip maximization.