> I'm not even sure we can make a computer that actually lasts 100 years.

I'm pretty sure we can. I have a 40-year-old calculator that is still working just fine.

Well, it still has 60 years to go before you can contest my hypothesis.

Still, I was probably too ambitious when I said "computer". I was thinking more that we can't make a PC or server-style computer that would actually last 100 years, while in regular operation.

Also, 1980s calculators have much, much fewer moving parts than today's computers. They don't have any kind of active cooling, they don't go through rapid heating and cooling cycles, they probably have three orders of magnitude less electrical components than a modern PC.

For me at least, all of the PCs I have had so far have had a few electrical components giving out after at most 10 years (usually some condensers leaked).

But all that is just because computers are not designed for longevity, they are designed for capability. We could easily design them for longevity if we wanted to: lower the clock speed so that passive cooling is all that's needed. Make all the PC traces five times fatter than they need to be. Use non-electrolytic capacitors. The resulting computer would be big and slow but it would last forever.

In fact, it would not surprise me a bit if you could run a Raspberry Pi Zero for 100 years.

Given an energy source that lasts the duration of the trip, I'm sure we could figure out how to build a ship that we can keep repairing quicker than it falls apart.

Then we're back to the energy requirement, yes.

I wouldn't even be convinced then that we have the technology today. Any repair system would necessarily suffer from the same kinds of failures.

I think the amount of mass lost is small enough that extra provisions can be carried around. For anything else you just have to fight entropy. Given enough energy you can probably do that.