There are people on HN far better qualified than I to discuss both orbital mechanics and spacecraft safety assessments but I'll give it a layman's stab based purely on the high-level concepts (which is all I know).

They know there's little to no risk to aircraft or people hundreds or thousands of miles to the East of a Starship RUD in orbit because they know exactly what's inside Starship and how it's built. They model how it will break up when traveling at these insane speeds and how the metal masses will melt and burn up during re-entry. They actually test this stuff in blast furnaces. It's a statistical model so it's theoretically possible a few small bits could make it to the ground on rare occasion, so we can't say debris will never happen - but there's been a lot of history and testing and the experts are confident it's extremely safe.

The case of the MIR space station was very different than a Starship. MIR was built a long time ago by the Soviet Union and they used a big, heavily shielded power plant. That lead shielding was really the part that had a significant risk of not burning up fully on re-entry. Starship, Starlink satellites and other modern spacecraft are now usually designed to burn up on reentry. However, there are still some things in orbit and things we'll need to put in orbit in the future that won't entirely burn up on reentry. There will always be a very small risk of an accidental uncontrolled reentry causing a threat. However, these risks are vanishingly small both because we design these spacecraft with redundant systems and fail-safes and because Earth is mostly uninhabited oceans, much of our landmasses are unpopulated or sparely populated, even in the unlikely event one of the few spacecraft with a large mass that won't entirely burn up has failed and is de-orbiting out of control, we can still blow it up - and timing that at the right moment will still put it down in a safe place (like it did with MIR). There's no such thing as absolute 100% perfect safety. But you're far, far more likely to die from a great white shark attack than be injured by satellite debris.

More to the point, a huge number of meteorites hit Earth every year and it's estimated over 17,000 survive to hit the surface. There are a bunch listed right now on eBay. Do you know anyone injured by any of the 17,000 space rocks that crashed into our planet this year or any airliners hit by one?

There were quite large areas of airspace closed just for this reason via NOTAMS - with airlines grumbling about that even before launch.

As I said, I'm only familiar with the high level concepts of vehicle launch safety and not qualified to assess detailed scenarios. I'm just a guy interested enough to read some technical articles and skim a few linked papers several years ago when there was a lot of heat about launch safety from Boca and not much light. When there's a lot of heated rhetoric in the mass media, I find it's better to check directly in scientific and engineering sources.

I dove deep enough to a get sense that these questions have been extremely well-studied and not just by 2020s FAA and SpaceX but going back to the Shuttle and Apollo eras. The body of peer-reviewed engineering studies seemed exhaustive - and not just NASA-centric, the Europeans and Soviets did their own studies too.

Your question is reasonable and occurred to me as well. Components engineered to withstand the enormous heat and pressure of orbital re-entry should be more likely to survive a RUD scenario and subsequent re-entry burn for longer. From what I recall reading, this fits into a safety profile required to ensure very, very low risk because even if a tiny percentage of mass occasionally survives to reach the surface, the actual risk that surviving mass presents is a combination of its quantity, mass, piece size, velocity and, most importantly, where any final surviving bits reach the surface.

I recall seeing a diagram dividing the Boca orbital launch trajectory into windows, like: right around the launch pad, out over the gulf of Mexico, the Caribbean, Atlantic, Africa, Indian ocean, and so on. The entire path until it's out over empty Atlantic ocean has minimal land, people and stuff under it. The gulf of Mexico is by far the highest risk because the rocket is still relatively low and slow. A RUD there could potentially be a lot of stuff coming down. There's not a lot out there in the gulf, just a few ships and planes but the FAA closes a huge area because, while the statistical risk is very low in an absolute sense, it's still too high to take chances.

For later windows, they don't close the corridor underneath to plane and ship traffic because the rocket's much greater speed and altitude later in the flight allows more precisely modeling where the debris field will come down. There was another diagram showing a statistical model of a debris field impact zone as an elongated oval with color-coded concentric rings dividing the debris mass into classes. The outermost ring is the debris that breaks up into smaller, lighter pieces. It's the widest and longest but it's the stuff that's much lower risk because it's smaller and slower.

The smallest concentric ring in the middle is where the small amount of heavier pieces most likely to survive will come down, if any do survive. As you'd expect, that innermost ring is shifted toward the far end of the oval and is a much smaller area. The headline I took away was that there's a very small amount of higher mass debris that both A) is less likely to break up into tiny, lower mass pieces, and B) is less likely to completely burn up. This is the higher-risk mass and, due to its mass, it tends to stay on trajectory, go fastest, farthest and not spread out much. In short, the statistical model showed a very high probability of any higher risk stuff which survives coming down in a surprisingly tiny area. The overall safety model is based on a combination of factors working together so it meets the safety requirements in each window of the flight for each class of mass. The carefully chosen launch location, spacecraft design, component materials, flight path and a bunch of other factors all work together to put the small amount of higher risk stuff down somewhere that fits the safety profile of very, very low risk to people and property. Disclaimer: I've probably got some details wrong and left some things out but this is the sense I got from what I learned. I came away feeling that the safety work done on space launches is comprehensive, diligent and based on a long history of robust, peer-reviewed science backed up by detailed engineering tests as well as real-world data from decades of launches, RUDs and de-orbits.

A fun side story: a few months ago I was at the Hacker's Conference and Scott Manley ("Everyday Astronaut" on YouTube) was attending as he often does. He brought along some interesting space artifacts just to set out on a table for casual show and tell. I was able to pick up and examine a Starship heat tile that was fished out of the gulf of Mexico. It was surprisingly light weight. Sort of like a thick wall piece from a styrofoam picnic cooler. It had a very thin hard shell on one side. This shell was clearly very brittle as it had already been broken up and I was holding an index card-sized shattered piece that weighed maybe a couple ounces. This was clearly not something that was going to maintain structural integrity post-RUD. Once it wasn't packed tightly together into a smooth aerodynamic surface, it's gonna shred into tiny pieces. And that seemed by design - which apparently worked as intended because even without a RUD, at the low and slow speeds over the gulf and near the launch pad it did shatter into small, light pieces - assisted only by the rocket tipping over into the water followed by the relatively mild explosion of the remaining propellant (mild compared to an unimaginably violent orbital RUD, that is). Holding it I remembered the debris field oval diagram and thought, "this is smaller, slower, safer stuff in the outer zones."