Oh no, there is no victory. Your summary is pretty good here. But the idea that Linux is provably less secure because old bugs were found is flatly wrong. They were found late, but they were indeed found. How many ancient bugs are lurking in proprietary software that nobody has found for positive reasons and made full disclosures of so affected parties know they need to mitigate? Nobody knows!

Server OSes are more likely to have older versions of software and then backport fixes. Desktop OSes are more likely to default to doing automatic updates, or at least hassling the user to do them.

The only difference is that servers are more likely to have more outdated software.

Love of convenience, I guess. I often find it astonishing myself. The software might do 100 times more but it takes 1000 times as much memory...

If by simple, you mean simplistic, then yeah. What you're forgetting is that every time a bomb kills someone's mother, father, brother, sister, wife, son, or daughter, another America hater is born. So there's likely to be an endless supply of irrational leaders, so long as they are put into power by someone bombing the previous leader along with random military targets.

The only regime changes that are ever really positive long-term are regime changes led by the people of a country against their leaders. All other regime changes are statistically more likely to make things worse than better.

Linux underpins the internet. It's the primary server OS on the planet. High-value data is held on Linux systems. The idea that it's not profitable to attack those targets is silly. They're harder to attack. People still do it. That's why there are still ssh port scans for example.

Ontogeny recapitulates phylogeny.

It is, but IME a lot of software needs architectural changes to work on it, similar to when you're trying to build software for Windows in cygwin. That's one reason I decided it wasn't worth the hassle back when I was running it.

When it comes to being allowed to do what you want with your computer, it's a lot more like Windows than it is like Linux. And it's been getting worse.

Now tell us how many similar bugs are in Windows, and will be found even without the obscurity of closed source. You don't know, because you depend on Microsoft to tell you when they fuck up, but you're declaring this a victory for Microsoft anyway? Do fucking tell.

No, 2MB was never enough for a Linux desktop. I had 8MB on my 386 and it was only just sufficient.

Let me restate that. Most of what makes neighborhood streets dangerous is vehicles and pedestrians using the same space at similar times.

Who has priority is largely uninteresting, because ultimately if a car hits you, you're still probably dead whether you had the right of way or not.

Ostensibly, sure, if you got rid of all the cars, streets would be safer for pedestrians, but they would also be a huge waste of space, because pedestrians don't need huge roads to walk. Roads exist principally for cars. The fact that pedestrians have to cross them is just an unfortunate design constraint that's hard to avoid cheaply, and giving pedestrians priority is mostly just feel-good policymaking that doesn't solve any of the fundamental problems.

The only truly safe way to share the space is to ensure that pedestrians aren't in the road when cars are. The best approach, at least in cities, is second-floor walkways, so that pedestrians and cars are never vertically at the same traffic layer. A slightly less optimal, but still reasonable approach is to give pedestrians a separate walk cycle in which the entire intersection is theirs. Pedestrians have priority during that cycle, and cars have priority the rest of the time, and as long as everyone follows the rules, nobody gets hurt.

But none of those solutions work for neighborhood streets, which is why the presence of pedestrians on neighborhood streets without sidewalks and proper traffic control for pedestrians results in the roads being inherently more dangerous than other streets.

For about $3000 USD you can buy an AMD Ryzen AI 395 with 128 GB of integrated RAM, which I'm tempted to do to run coding models. Although it seems to me that 256 GB is more of the sweet spot for local LLMs that can do things at a decent speed. For that size of RAM, the only real game in town is the Mac Studio, will cost about $10k (and rising). Of course even $10k is cheaper than a personal assistant. Now with the true cost of agentic AI starting to fall on the customer, $10k doesn't seem so ridiculous.

Current spacecraft velocity relative to earth is 3697 kph according to the latest spacecraft status display.

Attacking nuclear facilities is at least moderately rational. Various countries have done that half a dozen times over the past few years. Attacking drone manufacturing and storage might also be reasonable.

But...

What does an illegal decapitation attack have to do with nukes? Do you think the new supreme leader is going to somehow be more rational than the last one? There is a fundamental difference between going after clear military targets to prevent Iran from developing weapons that threaten their neighbors and going after civilian and government targets.

No, they will keep doing this until they are a nuclear power. They've seen what denuclearization did for Ukraine, and it's hard to argue with their logic. Having nuclear weapons is a strong deterrent to invaders, who realize that the response could be swift and devastating at a scale that countries never recover from.

It's unclear what other things they will do at that point. We can only speculate. Mind you, I don't like the idea of a nuclear-armed Iran, but again, I see no evidence that anything happening over there right now is going to change anything, or even delay it enough to matter.

Is anything that the U.S. government is doing right now going to change that reality? The way you prevent them from laying mines is the same way that you prevent oil from leaving Iran — bombing ships the second they leave the harbor. If you're not willing to start with a full air and naval blockade, you've already failed, and the only thing continuing the war can do is increase the number of ways that you've failed.

If you set it to "85th percentile of observed traffic" you are selecting 15% to be targets of fines. Why 15 and not 20, or 10?

States with "reasonable and prudent" rather than "explicit speed limits" do a more logically consistent job here. Reasonable and prudent is what we're really looking for - everyone choose a speed that is safe for the conditions of the road, the vehicle, and the surrounding traffic.

The problem is that it's difficult to fine people for that, because it is partly subjective and different for every driver and weather conditions. It's much easier to set an explicit speed limit and then measure speeds. Explicit speed limits exist for the convenience of the courts, with safety of the road users as a distant secondary objective.

If you want to improve safety, then look into "traffic calming" measures. In particular those that cause drivers to perceive higher risk (and research into conditions where drivers falsely perceive lower risk). Even just drawing the lines narrower on a wide street can have an effect. If you design the road right, drivers will naturally choose the right speed for the environment without any need for a road nanny.