2 ms·
> A working VN probe literally has to reproduce a significant part of an entire civilisation's industrial capacity - mining, refining, all levels of assembly -
by TeMPOraL 4y ago
> A working VN probe literally has to reproduce a significant part of an entire civilisation's industrial capacity - mining, refining, all levels of assembly - while improvising with a mix of raw materials which may not be similar to those available at source.
The optimization target here is a minimum-complexity design that's able to reproduce itself using in-situ materials. This is still an unsolved problem, but I don't see a reason the probe would have to "reproduce a significant part of an entire civilsation's industrial capacity" - it doesn't have to build anything else other than its own copy (and whatever it needs for that). It has to do one thing, and thus we could cut corners in the design.
> And it has to do it in a way that is 100% guaranteed free from software bugs, hardware failures, and general space wear and tear. Over millions of years. (Assuming FTL isn't possible.)
This is not a hard requirement. You don't need 100% reliability - you only need enough so that, on average, one probe turns into more than one. Beyond that, more reliability only speeds things up. This is how biology does it: reliability is achieved by doing more things than necessary and rejecting (recycling) failures.