“If you don’t like your job, you don’t quit! You just go in every day and do it really half-assed. That's the American way.” - Homer Simpson

I was not familiar with TAI until your question, but I googled it. It looks to me like TAI is the most accurate GPS time. The problem with GPS time is that it is not “earth time”. By that I mean GPS time is independent of the position of the earth. Since the earth rotation is slowing over time, GPS time deviates from it, running ahead of it as the earth slows. UTC time takes this into account. Leap seconds slow GPS time to align with “earth time”, or the time it takes the earth to rotate one time. Since satellites are often concerned with observing earth, or communicating with earth, it’s important for them to stay aligned with the actual earth rotation, so UTC time is more useful. One of my first assignments was making it easy for ground control to upload deltaUT1 and leap seconds to our satellite (Calipso) so that it’s science data was accurately tied to the ECEF reference frame.

That’s unusual. Usually the bird has a GPS receiver. In your case, the satellite clock usually has an epoch from which it counts time (time zero, set to something like 1958). This time is then adjusted by a parameter called deltaUT1, which is a tenth of a second resolution correction based on Earths rotation. The time dilation due to relativistic effects is minor, but an unsynced onboard clock that is not corrected will drift due to the clock a noticeable amount within just a few months. I would think Mission Control would have constant issues commanding it to take data at certain times (if it were an imaging bird for example) unless they synced the clock periodically.

I’m no Ayn Rand fan, but there was no change in her attitude to public assistance. She said of course she would take advantage of a system she was forced to pay into her whole life. She thought it should be completely dismantled, but she wasn’t going to pay in and not get back.

He doesn’t regret it, and he knew the facts when he wrote it. He said it was just a choice he made for dramatic effect and to get the book going. It’s hard to think of a scenario where astronauts would leave behind one of their own, so he invented one.

I tried to make the book as accurate as I could. The biggest place that’s inaccurate – don’t tell anybody – but if you’re in a dust storm on Mars, you’re not even going to feel it. Mars’ atmosphere is less than one percent of Earth’s. So a 150-km/hour wind would feel like about a 1-km/hour wind does on Earth. It wouldn’t do any damage to anything. Shhh …Most people don’t know how Martian dust storms work, that it’s not like being in a sandblaster. It’s just more dramatic that way. So I just made that concession. [shrugs and smiles] I know I’m a liar. I just … wanted that.

In the 90s I worked on hypersonic projectiles to intercept nuclear missiles. They were doing mach 6 out of the barrel and yes, the air around them was plasma. We had an on board control loop that used feedback from a LIDAR sensor. The window for the sensor had to be ruby so it wouldn’t melt.

The projectile was spun at 50 hz, so we got samples back at that rate. Even with the primitive chips at that time and a 50hz sample rate, we could control the projectile to within 1m of the target. This was with no ground communication and not even an inertial reference, so all control was in the frame of the projectile. My point is that while ground communication would have its uses, it’s not a fundamental problem with hypersonic projectiles. We can already get them where we want them to go.

Here is a computer simulation of two galaxies colliding, and they pause it at different points and show you Hubble photographs of 2 galaxies currently in that exact stage of collision. It’s really cool!

https://youtu.be/rs3CL930svo

edit: of course it’s two different galaxies in each photo