Whistle blower says non-human bodies recovered from crash

[UngainlyTitan]

I actually think one of the biggest issue is the cancer problem, or "evolution".

You make that many copies and your going to get variant code, eventually you might evolve a secondary machine that disobeys the prime directive, aka cancer. So controlling that is going to require a lot of same things the body has now, probably an immune system and the like, which would slow down the replication ultimately.

Cancer is essentially where a specialised cell in a multicellular organism begins to behave like a single celled organism. Not sure that an artificial replicator would have a single cell mode but otherwise your point stands. Given the environment it operates in there are plenty of room for transcription error to propagate.

 

[Stalker0]

Cancer is essentially where a specialised cell in a multicellular organism begins to behave like a single celled organism. Not sure that an artificial replicator would have a single cell mode but otherwise your point stands.

Ultimately a nanite in this scenario would need to have the ability to decompose and recompose elements its finds....which would include fellow nanites. So a nanite swarm would need some kind of instruction to prevent that from happening, and so if that instruction failed due to transcription error you would have nanites destroying other nanites and harvesting their resources to build more "bad nanites".

In other words....cancer!

 

[UngainlyTitan]

Ultimately a nanite in this scenario would need to have the ability to decompose and recompose elements its finds....which would include fellow nanites. So a nanite swarm would need some kind of instruction to prevent that from happening, and so if that instruction failed due to transcription error you would have nanites destroying other nanites and harvesting their resources to build more "bad nanites".

In other words....cancer!

I would not call that cancer that is a nanite predator. A cancer does not consume the host (generally ) the cancer grows until its growth interferers with something important.
However, the rogue nanite become a predator and now we have a classical evolutionary predator prey arms race. The original nanite can respond by evolving defences or moving to other environmental niches that are unsuitable for the prey.

 

[Umbran]

So controlling that is going to require a lot of same things the body has now, probably an immune system and the like, which would slow down the replication ultimately.

And increases size, which may be more the problem. The stuff is called *nano*technology because it is supposed to be down on that range scale - like viruses that are on the order of 20nm to 200nm.

If you have to include lots more mechanisms, you also talk size increase, probably up into the size range of cells, which is more in the micrometer range.

I have issues with the classic grey goo, in that I think the thermodynamics does not hold up.

What do you think is the thermodynamics problem? It rips apart some matter for energy to use to rearrange other matter. It has some waste material and heat. What's the problem?

 

[Rabulias]

Given the environment it operates in there are plenty of room for transcription error to propagate.

Ultimately a nanite in this scenario would need to have the ability to decompose and recompose elements its finds....which would include fellow nanites. So a nanite swarm would need some kind of instruction to prevent that from happening, and so if that instruction failed due to transcription error you would have nanites destroying other nanites and harvesting their resources to build more "bad nanites".

Nanites could be made to minimize transcription errors far more than a biological process. We have error correction and things like CRC checking for copying files and data transmission. I would hope something like that would be implemented in any machine that can replicate itself. It would slow the process down somewhat, but we can still copy and process data very quickly with those safeguards in place.

 

[Clint_L]

What do you think is the thermodynamics problem? It rips apart some matter for energy to use to rearrange other matter. It has some waste material and heat. What's the problem?

Wouldn't the heat be a huge problem? If it was converting matter at speeds approaching anything like the nightmare grey goo scenario, it would be releasing enormous amounts of heat - probably way too much for it to withstand.

Life is already very efficient at converting matter, and I'm not sure how the grey goo would get many orders of magnitude faster.

 

[Umbran]

Wouldn't the heat be a huge problem? If it was converting matter at speeds approaching anything like the nightmare grey goo scenario, it would be releasing enormous amounts of heat - probably way too much for it to withstand.

Well, it depends on what gives you nightmares - it being fast, or it being relentless.

One thing that gets in the way of (terrestrial) life at high temperature is that proteins aren't very heat-tolerant. Grey Goo doesn't have to rely on fragile protein folding, so we can imagine its operation being more tolerant of high temperature.

Also, Grey Goo doesn't have a body. Animals have to maintain bodies to contain complex structures, and that gives us issues moving heat into, or out of, those structures. Grey Goo does not use macro-structures, and we'd assume that all the chemical action would happen where the Goo touches other materials. Any other bulk is of no use to the goo - so it maximizes its surface area to maximize that interface, and spreads out thin. Really thin, like a millimeter or less, where radiative cooling is more sufficient and efficient.

Life is already very efficient at converting matter

Compared to what? And efficient in what sense? What doesn't life waste in converting matter?

"Efficiency" does not actually mean anything in and of itself - efficiency is always measured in some specific sense - like car fuel efficiency is usually listed as distance per unit of fuel used. But generally speaking, increasing efficiency in one sense sacrifices some other measure of performance. A fuel efficient car typically isn't fast, so it isn't time efficient, for example.

In a Grey Goo scenario, we can imagine the Goo winning because it is horribly inefficient in some sense. Say, for example, that the Goo is extremely ineffecient in terms of material waste products - the Goo is grey because it only converts a small amount of the material it consumes into more goo - most of it it leaves behind as ash. Perhaps even hot ash, dumping some of that extra heat into the physical waste.

 

[Stalker0]

Nanites could be made to minimize transcription errors far more than a biological process. We have error correction and things like CRC checking for copying files and data transmission. I would hope something like that would be implemented in any machine that can replicate itself. It would slow the process down somewhat, but we can still copy and process data very quickly with those safeguards in place.

I've been trying to find current accuracy rates for hard drive data copying (which would be the rough equivalent of copying data into a new nanite)

But my google fu has been failing. I did find that Dna polymerase replication has about 1 error per 1 billion nucleotides (which includes error checking and correction). A nucleotide is about 2 bits of info considering the 4 possible nucleic acid options in standard dna, so ~1 bit error per 2 GB copied.

 

[Stalker0]

Compared to what? And efficient in what sense? What doesn't life waste in converting matter?

"Efficiency" does not actually mean anything in and of itself - efficiency is always measured in some specific sense - like car fuel efficiency is usually listed as distance per unit of fuel used. But generally speaking, increasing efficiency in one sense sacrifices some other measure of performance. A fuel efficient car typically isn't fast, so it isn't time efficient, for example.

I would argue that a Grey Goo swarm would likely have two key efficiency %.

1) What % of total energy is used when replicating (in other words how much energy is wasted).
2) % of total energy is used when performing atomic rearrangement (aka its matter conversion purpose).

Likely you would have a "standard operating" efficiency as the nanite will have to spend some % of its lifecycle replicating, and the rest doing matter conversion (similar to how a cell has to stop and divide every so often). So a combination of the two %s above based on the standard replication rate of the swarm would give you its overall efficiency.

 

[Umbran]

I would argue that a Grey Goo swarm would likely have two key efficiency %.

1) What % of total energy is used when replicating (in other words how much energy is wasted).
2) % of total energy is used when performing atomic rearrangement (aka its matter conversion purpose).

For the archtypal grey goo, the purpose is replication. Like, you released a bunch of self-replicating nanobots to perform some major task, and in some failure mode they stop with the major task and are only replicating.