Can a Commander build itself?

I'm not sure that it is necessary, and it might not realistically possible. It's one of those strange paradox problems. "Can a set of all sets contain itself?" Rather, can a Commander contain all the information necessary to build himself? Let's say "sure". For the sake of this argument I use TotalA logic.

Normal units are produced from a nanite slurry that "congeals" into a killbot. Nanites transform into circuits, motors and armor, creating a robot down to the molecular level in less than a minute. Flaws like molecular errors are basically irrelevant, as the process is about quantity rather than quality. Glossing over a molecular crack is okay if it gets 2 more bots on the field. Since the killbots do not replicate, any flaws in their design die with them.

Let's use the same process for the Commander. The first Comm uses its nanite slurry, creating another copy of itself. The process creates a number of flaws and errors as a matter of course. Those flaws become part of the new Commander; its designs might have weaker sensors, or a factory might give units too much armor. That Commander spawns a new one, carrying its flaws and introducing several more mistakes in the process. What you get is a form of mutation, where an originally perfect machine is getting new errors and flaws with each generation. Robot designs mutate even more, and patterns start becoming lost or unfixable.

Eventually you get a generation of Commanders so flawed, they can no longer replicate themselves. Antimatter chambers develop a flaw, or the d-gun doesn't quite focus properly, or a control nanite malfunctions, and all that remains of the copy is a crater. As early generation Commanders die off, the war loses its ability to create new Commanders, and their numbers begin to diminish. This process might take hundreds or thousands of years. But eventually, the war can no longer sustain itself, as the original Commander design has become ruined beyond repair.

There are ways to limit the damage. Using more advanced factories, it might be possible to create a perfect atomic copy of a Commander. It may be a slow and painful process, unsuited to mass production and highly vulnerable on its own. But it does reduce the number of flaws introduced per generation. Unfortunately this once again depends on a high quality original, and further flaws may render a Commander incapable of repeating this process.

The only way to truly fix the accumulated flaws is to break everything down, reverse engineer it, and completely redesign the Commander, atom by atom, from the ground up. No single unit could do this, especially one on the field. It requires such an insane amount of processing ability, that even attempting to describe the processing needs wouldn't do it any justice. The systems would occupy an entire planet at least, an unparalleled "central consciousness" dedicated entirely to the task. Only then, could the flaws in a Commander's design truly be discovered, analyzed, repaired, and redeveloped into a remake of the original.

This planet would likely be an entirely artificial world. It would be built entirely out of the heavy metals and raw materials needed to mass produce the new design, a simple task by comparison. When completed, a new wave of 1st generation Comms would hit the field in minutes, spreading through the galaxy like a plague born anew. In effect, a re-perfected Commander would crush every other faction and single handedly end the war; and would be a threat to the entire universe!

WOW! Just got to say thats a wonderful idea!

You're supposing that the nanite creation method is akin to an analogue copy, with no error checking. Like recording on a VCR. I see no reason why this would be. I think the idea of nanites introducing flaws into a design, through mistakes, is pretty unlikely. So I have no problem with the building of the ACU machine, I think that can be done.

What I do think may take time, would be to create a new machine consciousness. A clone of the current ACU's sentience could be made in the time it takes to copy the data. If it wished to birth a new entity, that could take a significant amount of time. (relative for the machine, it's likely they would be operating on a much smaller timescale than us, nanoseconds or shorter certainly)

Yeah, nanites being flawed wouldn't really make sense. Each nanite would be programmed to go to a particular spot from a blueprint - even if the copy was missing bits, the blueprint copy would be a data file, not a copy of the actual commander. Therefore you'd more likely have a perfect commander being built by an imperfect one. Corruption in the blueprint file itself would be unlikely.

Of course, the commander may be the only unit with an organic brain - Machine AI's solution to the non-linear thinking required for military leaders. So you may not be able to just replicate a commander anyways...

I also find it highly doubtful that a commander has an organic brain. They're the ultimate culmination of technology after an essentially endless war. There is absolutely no physical reason why a more intelligent consciousness than even the most augmented human brain possible cannot be created. Tell me why you'd keep an obsolete evolved system in play, long after you've designed something far more capable. The only reason I can think of is incompetence.

Which is why ARM suck.

No, no no no no nononono. You all got it all wrong. Gosh, what are they teaching in commander school these days anyway?

Bobcules, I think it's about time I taught you all about the peepers and the finks. You see, when a girl commander and a boy commander love each other very much, they cross their nanoparticle streams and-

Flaws by tech are improbable. Flaws by design are probable - i.e. while building new ACU old ACU may decide to build more efficient version by replacing some parts, creating more efficient short-term battle commander, but less capable long-term battle commander. But anyway, it won't lead to "reproduction inability" =)

I'd rather agree with Zordan - it's more about conscience, not about tech behind, if you believe that robots are sentient. But probably they are not sentient - they are automatic systems left without turn-off timer.

Also, it could be other way around (which is better, imo) - initial design is a the most advanced tech, but not perfect. War is perfect evolution model so only strongest commanders survive. Commander-copy mechanism may explicitly create few random modifications to develop most powerful and efficient commander ever by trial&error.

And this process may lead non-sentient automatic system to then sentient robotic race =)

Well have you noticed how effective sentience has been at escalating a rather average animal directly to the top of the food chain?

Yes, but it also make this average animal more keen and inventive on it's ways of killing other animals. Endless war of chasing each other is neither "keen", nor "inventive" =)

That's why sentience is rather end-of-game.

Some would say that is actually a conflict between the higher sentience and the baser instincts that results in that.

Zordon has a point, Commanders wouldn't simply copy themselves, it would be more logical that they use a data blueprint, before copying the blueprint (along with those of all units) to the created Commander ; and good error checking and handling would make the corruption of the blueprint very unlikely.
They may use mutations (be it because of the rare unchecked error or voluntary modifications), but after thousands of years of war and subsequent evolution, they have to be pretty optimal now to be called 'pinnacles' of war.

Having the consciousness be the hard part is interesting, and I'd see another one, justifying the entire artificial planet stuff : the FTL. If we keep Galactic Gate-like fast-FTL being the advantage of the Commander, then making the Commander Gate-capable may be extremely difficult, justifying giant installations and explaining why Commanders aren't keen to create new Commanders by the dozen.
That and the backstabbing, individualistic tendencies of each Commander that think that the Universe would be a better place without everyone else.

Aren't they vonn neuman probes?

If the commanders are capable of reproducing themselves with no more knowledge that they themselves possess, then yes (there may also be a clause in the definition that means they must be able to mutate/improve upon themselves, I would need to look it up).

(Note: all non-commander units are not vonn neuman probes, as the require additional knowledge provided to them from the commander)

Even if they did having errors when reproducing, your idea that they would slowly get worse wouldn't happen.

In fact, some mutations could end up in a more efficient commander, and natural selection (through robots killing worse commanders more easily) means these better commanders would live longer and pass on their "genes" in the same way cells do.

Do you know how many years it passed since life on Earth emerged? And we still has nothing optimal. Problem with evolution is that it's very bad with finding long-term optimal solutions, as path to long-term optimal solution may lie though a lot of sub-optimal intermediate phases. And evolution is about finding short-term optimal solution.

As war is constant and there is constant fast-paced evolution process, I presume that ACUs evolution will run in circles, or spiral, but far from being "optimal".

If you have blueprints that produce a robot capable of traveling though gates, so it's not much difficult to produce it. So you may simply state "ACUs are very-very expensive. So why spent resources for ACU when you may build a 1000x planet-scale KEWs using this resources?".

But possibility of safe self-consciousness transfer though teleportation (but not FTL, actually) produce an philosophical problem. So you can say that gates are teleportation, not FTL, so to transfer ACU for long distance you need other tech (FTL), which is much-much more expensive. But that's rather weird explanation.

Backstabbing == consciousness. Mostly =)

If you can make something flawless, then you aren't making it fast enough. The nanolathe isn't so much about building flawless machines, as it is about producing the MOST machines. Similar to the purpose of biological stem cells, the nanite is generic and can be pressed into an enormous number of different roles. They might not create ideal forms or perfect designs, but for the sake of speed they too are sacrificed to create electronics and armor for the robot. It's quick, it's dirty, and the unit doesn't even have a chance to cool off before another one is on the assembly line.

Commanders who neglected this need for speed simply would not have been competitive on the field of battle. They would have fewer robots, slower economic growth, and would just die off.

When used in this way, it is easy to see that a mass assembly process just isn't suited to producing a Commander. They are designed from the ground up for mechanical perfection on an atomic scale. A generic nanite slurry just won't cut it, as the nanites themselves interfere with the process and reduce fidelity. What's left from a copy attempt is a flawed design, identical in appearance but lacking the full capability of the original. The clone would have less processing power, a weaker lathe and armament, and consequently less ability overall. Subequent clones would be based off of that design, losing further quality by the lathing process, perhaps even rendering its experimental tech inoperative from the accumulated flaws and incomplete inheritance. Eventually it reaches a rock bottom, where the Commander is no better than a high tech nanolathe unit. Useful, expensive, but by every measure a shadow of its former self.

A Commander is not quite a 100% von neuman machine, and it can't be. It is not reasonable or competitive for the Commander to have the tools needed to replicate itself. To put it another way, the Comm is a .png, and the lathe only makes .jpg units. It's great, but there's always some bit decay. A ".png" lathe can certainly copy a new Comm, flaws and all, but bit decay can only be removed by a complete remake of the Comm's systems.

The problem here, is that a Commander starts from "perfect". Mutations at the top tend to only go downhill.

I think it's been fairly well proven on the modern battlefield that less numbers, higher quality, can be more effective than more numbers. So your entire argument that nanolathes would make inferior copies kind of falls apart. Not to mention that the limiter isn't construction speed but rather available resources, indicating a need to use those resources as efficiently as possible.

Not only that, but with complex systems (as is the problem with evolving units) a discrepancy in the design is going to make the system fail. Good designs take these failures into account and have redundancies in place, so a 'bad' commander copy would perform exactly the same as a good one, except with less redundancy.

And not to mention that nanites can obviously break down as well as build components, so it's not unreasonable to suggest that they have a built-in component check mechanism that automatically checks a component once it is complete, and disassembles it if it fails the test. It would be silly to send out a robot without a functioning leg, or arm, or targeting sensors just in the name of 'speed'.

On the topic of biological brains, it's intuitive thinking and leaps of logic that can bring out the really unique and winning strategies, and these are areas that are traditionally very hard for a programmed machine to do. Augmenting the AI core with a biological system wouldn't be too far fetched if it gave the commander the ability to think outside the box and potentially win more battles.

We're talking insane future tech pawz. You can't on one hand assume awesome perfect material replication with nanoscale manufacturing and on the other that biological processes are the only way to create intuitive thinking and leaps of logic.

*edit* totally a word.

Perfect doesn't exist. There's no single perfect item, ANYWHERE. It all depends on the context. Sure a commander we have right now may be perfect for a generic planet, but what about extreme heat or cold? It might need to trade off armour for more stable electronics inside. Maybe something on a planet ravaged by natural disasters would need to be heavy (at the expense of being slower) so that it could survive floods etc without being washed away.

Perfection doesn't exist.

Not necessarily, there may still be the factor of adapting to the current situation, resulting in mutations that are better in some scenarios etc. (Did i just describe factions?!?)