The implementation may be more complex, but that's exactly the kind of thing that ends up encapsulated in a library once the implementation settles down.

What they've got here sounds likely to be an intermediate step. I'm reminded of Tim Peter's series of posts when he was developing what is now known as "Timsort". The implementation combines multiple sorting algorithms (similar to how this is combining multiple shortest path algorithms), and went through many iterations as it was tuned for various different datasets exhibiting different performance characteristics. This sounds to be at the stage where "it works better for these datasets, now we need to make it work better for *all* datasets". An important step showing that it's *possible*, even if not finished.

BTW since then Timsort has been further evolved into Powersort which has provable performance characteristics (and better performance).

How often do you successfully type a 48-character string without error?

If copying directly from a screen/printout?
If being listening to someone read it out loud? And how often do they correctly read it?

Now apply that to doing the same thing over and over.

In most countries vacation can be rolled over (employers can require that it be used or paid out if too much accrues) whilst sick leave doesn't roll over.

OTOH in most places I've seen that have combined PTO it can't be rolled over at all.

No really - it was the AI that shared the confidential governmental tax information that was then used by PwC partners for tax dodging.

https://www.abc.net.au/news/20...
https://www.theguardian.com/au...

What actually happened:
https://www.polygon.com/238737...
https://arstechnica.com/gaming...

San Francisco police told Polygon that the person who reported the threat to police said "an employee made a threat towards his employer using social media." The aforementioned employee reportedly works at an out-of-state office.

Police advised Unity to contact the out-of-state jurisdiction before taking a "courtesy report,"* San Francisco public information officer Kathryn Winters said.

Whilst I agree entirely that death threats are unacceptable, Unity's attempt to spin this so that people assume it was multiple angry third-party developers (spin which has been largely successful) is also unacceptable.

You know what would create even more jobs and keep even more people employed? Putting the same money spent on SLS into essential services such as health and education. Or even half the money spent on SLS.

SLS has been demonstrated to not be anywhere close to being able to keep to budget (time or money). Given that do you really think they will use even more money in a cost-effective manner?

Were the 2 Seagate drives bought at the same time (and for that matter, were the WD?).

It's a common thing when initially setting up a NAS to buy all the drives at the same time, which mean they're likely all from the same batch, which means if there's a flaw that leads to shorter-than-expected lifespan in that batch, all the drives will tend to fail after similar usage.

If you have to buy a batch of drives all at once, and you want them all to be the same model, best option is to buy from multiple vendors. You'll pay extra on shipping, but decrease the chances of them all coming from the one batch.

That's not entirely correct. Originally there were meant to be two suppliers of lunar landers selected (but only selecting one was not prohibited in the terms).

Of the submitted applications, only the SpaceX one had any chance of working or being delivered on time. In addition, NASA were only given enough money for it to actually pay for one (BTW the SpaceX application was also the cheapest).

And that's why only SpaceX were awarded a contract.

Damned lack of edit ...

Speaking of code, as a professional programmer I don't think that code should fall under copyright, but trade secret laws. Which would make open source automatically public domain.

And yes, I know there are complications e.g. web serials, programming, etc. So maybe multiple categories, where "complete published work" (and a movie is a complete published work whether it's part of a series or not ...).

I think there's a fairly simple solution.

Copyright may be held up to the current limits (so no one can complain that "their rights are being taken away"). But you need to pay each year, with the cost increasing exponentially every X years.

First year is free. For most people, this would include the vast majority of any income they will get from the work. Next say 5 years is a low cost. If you're making any kind of ongoing revenue then this payment should be easy. And every 5 years the cost doubles for the next 5 years.

So if Disney wants to keep copyright up to the current limit they can, but at some point they have to make the decision that it's no longer cost-effective. Meanwhile, most other works would have entered into the public domain much closer to the original 20 years (and in many cases after only a couple of years).

Year 0 - free.
Year 1-5 - $1000/year
Year 6-10 - $2000/year
Year 11-15 - $4000/year
Year 16-20 - $8000/year
Year 21-25 - $16000/year
Year 26-30 - $32000/year
Year 31-35 - $64000/year
Year 36-40 - $128000/year
Year 41-45 - $256000/year
Year 46-50 - $512000/year
Year 51-55 - $1024000/year
Year 56-60 - $2048000/year
Year 61-65 - $4096000/year
Year 66-70 - $8192000/year ...

0.75G seems a reasonable guestimate at what may be the low-end tolerance to allow gestation, and if so Mars (at ~0.38G) would be outside that tolerance range.

No, the GIL has always been seen as at best a necessary evil, not a desirable feature. The GIL is not considered part of Python but only part of the CPython implementation - other implementations are free to (and do) use other methods.

There have been many attempts to remove the GIL in the past, but they've all suffered from one major flaw - significant performance reduction on single-threaded code. This new implementation appears to not suffer from that same flaw, so for the first time in a long time people are getting excited about the possibility of finally fixing what most consider an undesirable (but practical) implementation in CPython.

Impressive - you managed to make one correct statement in there.

This is a Gwynne Shotwell-stated goal. Gwynne tends to be much more realistic about timeframes than Musk is. If she thinks that it's achievable by the end of July, then I believe it. Not that it's defnitely going to be achieved, but that if things go well it is possible. Remember that many of the items on your list are being worked on in parallel. There's a lot of people involved in this.

BTW, apparently they're looking to have ~70 raptors verified by (can't remember the exact timeframe, wasn't too long away).