Robots of the Imperial Army

[samuelvss]

This is hardly the first instance where the OTU description of the universe as-it-exists and the economic implication of the physical rules as-they-are-described do not mesh when subjected to extensive scrutiny. I always remind myself that the rules exist to facilitate gameplay, not to model a perfectly functional universe (OTOH, heaven help any poor fool who tries to make a Dungeon and Dragons economy work, etc., etc.). And when the system breaks, the solution is to provide an explanation that most quickly and easily explains how the PCs travelling together in the ship that they have doing what they are doing makes perfect sense.

To that end, I always rule that at Imperium standard tech level, factory robots are economical. The parts of the ship that are effectively robots, from the autopilot to the cargo-bay loaders to a targetting computer, are economical. The robot who takes over for a crew member and therefore should justifiably replace a PC is in fact either more expensive than a human crewman (perhaps their parts are so expensive, that the maintenance cost exceeds a salary and life support), or is simply not reliable enough to trust in that position (the same way that self-driving cars are right now. They are getting reliable-enough to use, but not reliable enough to replace the driver's decision-making capacity outright).

A good thread which has some examples in it, and some discussion on "Semi-autonomous," in LBB8: http://www.travellerrpg.com/CotI/Discuss/showthread.php?t=17244&highlight=AstroMech

I think the economics do mesh, and as you point out, it's a game not an economic simulation. Culture drives law and economics. Cannibalism arguably makes much more sense economically than our pricey funereal industry, but is culturally unacceptable to the point that it is illegal.

I would say that a robot who can take over for a crew member could not justifiably take over for a PC. A PC who is the Engineer is going to decide whether and when to fix the jump drive. Safety, money, personality, and myriad other factors will play into this. A robotic engineer is going to fix the jump drive according to its last instructions. A semi-autonomous LBB8 robot with Mechanical-4, Engineering-4, and an Apparent Intelligence of 13 is going to do a very good job of it; it's not going to tell you that it was probably sabotaged by that weaselly customs inspector, however, or that maybe we would be better off trying to go over to the Corsair, get it's powerplant online, and jump out without maneuver drives. That is a good reason I would not have a robot as a chief engineer. I might have a Chief Engineer with Engineering-1, and his able-bodied genius-trashcan assistant, if I can afford it.

LBB8 gives a great opportunity to allow robots in, and a justification for limiting their proliferation in the 3I, however weak we may judge it. IMTU, I do both. As Hans' signature states, it also has to make sense; to me, it makes enough.

A robot never takes over for a PC, though. If a PC was gunner, he could not be replaced by a robot; I could have a robot gunner, but there would be a hole on the "crew." A robot could not even replace Jayne, the "man-ape-gone-wrong-thing." He might shoot better, but he would never be trusted to make the call as to when to start shooting Patience's men, o: let alone have a song written about him...

Under LBB8, you're going to be able to buy a gunner for cheaper that you can house, feed, and hire one in LBB2; you may actually get your gunner confiscated by some local bigoted customs official, though; your gunner is not going to help buy groceries, "keep an eye on things," or come up with a solution to your next dilemma. Since it is not a "required" crew position under LBB2, that may get me through legal regulatory issues by not claiming it as "crew."

Anyway, for me it fits. I can't tell anyone else why it should fit for them, but I have tried to explain why it fits for me: The costs of robotic crew are higher, these costs are just not quantifiable economic costs.

 

[rancke]

This is hardly the first instance where the OTU description of the universe as-it-exists and the economic implication of the physical rules as-they-are-described do not mesh when subjected to extensive scrutiny.

But wouldn't it be nice if they did mesh even under very little scrutiny? Is there any advantage to refrain from adding an option to make autonomous brains more expensive and stating that this option applies to the OTU? Perhaps even give examples of universes where one price scheme applied and universes where the other applied?

I always remind myself that the rules exist to facilitate gameplay, not to model a perfectly functional universe (OTOH, heaven help any poor fool who tries to make a Dungeon and Dragons economy work, etc., etc.). And when the system breaks, the solution is to provide an explanation that most quickly and easily explains how the PCs travelling together in the ship that they have doing what they are doing makes perfect sense.

So do I. But I don't think there is any possible explanation why, at the canonical prices, robots are not used extensively and routinely as ship crew. Except the one that angry mobs will destroy any robot employed to supplant a live being, and I don't like that explanation because I quite like the option of employing robots in situations where the cost is not the most important consideration.

To that end, I always rule that at Imperium standard tech level, factory robots are economical.

Do you mean automated factories or do you mean factories that could be operated by live beings but are instead operated by autonomous self-propelled mechanical entities (ASMEs)¹?

¹ It puzzles me that in a Traveller discussion, the SF-tropish form of robot is not the default assumption for what is meant by the word 'robot', but as this is evidently not the case, I feel compelled to spell it out.​

The parts of the ship that are effectively robots, from the autopilot to the cargo-bay loaders to a targetting computer, are economical.

The automated parts that would be there whether the crew were live beings or ASMEs. Sure, why not? If they weren't economical ships wouldn't be built that way.

The robot who takes over for a crew member and therefore should justifiably replace a PC is in fact either more expensive than a human crewman...

That would work for me if robots capable of performing the work of a crewman were not, in fact, less expensive than a human crewman.

(perhaps their parts are so expensive, that the maintenance cost exceeds a salary and life support),...

Then the robot rules should include separate robot maintenance figores.

...or is simply not reliable enough to trust in that position (the same way that self-driving cars are right now. They are getting reliable-enough to use, but not reliable enough to replace the driver's decision-making capacity outright).

Then the robot rules shouldn't state that they are capable of replacing people with the same skill set. Also, please note that no one has suggested ships crewed entirely by robots; there'd be a few humans along to captain the ship and make the decisions the robots can't.

If a player wants to play a robot, that should be a higher tech level model whose probably too expensive an investment for a merchant ship's owner to buy, but most of the cost is a sunk cost.

But according to the robot design rules robots are NOT too expensive an investment for a merchant ship's owner to buy. If they were, there wouldn't be a discrepancy between robot design and ship design. You can get a very capable robot for the Cr500,000 you save on the stateroom you can omit if you use a robot.

Once the robot is there, it costs roughly as much to keep going as a salary + life support.

That would work if it was true, but the problem is that according to the rules this is not true. And, really, if robots did cost about Cr200,000 per year to maintain, you'd have the opposite problem that robots wouldn't exist anywhere outside research laboratories.


Hans

 

[WistfulD]

But wouldn't it be nice if they did mesh even under very little scrutiny? Is there any advantage to refrain from adding an option to make autonomous brains more expensive and stating that this option applies to the OTU? Perhaps even give examples of universes where one price scheme applied and universes where the other applied?

Yes it would. It would be fantastic if it had all been worked out perfectly by an economist (or at least a pedant who was very good at math) back in the 70's (although given that running cargo for all but J1 routes is not economical by-the-book, it's clear that the system was very much not stress-tested). It's certainly reasonable to say, "wouldn't it be nice?"

But after that, were do you go? How do we "[add] an option to make autonomous brains more expensive and stating that this option applies to the OTU"? The Official Traveller Universe is currently using a 5th edition that very few people are accepting as even truly fully released, and the majority of Traveller players play discontinued versions of the game. How do we "officially" fix an out-of-print gamebook?

I don't know if you were on the hydrogen jump bubble thread, but we've clearly seen what can happen to otherwise brilliant people when they get too worked up about the fact that a four decade old rulebook doesn't meet our expectations.

Do you mean automated factories or do you mean factories that could be operated by live beings but are instead operated by autonomous self-propelled mechanical entities (ASMEs)¹?

¹ It puzzles me that in a Traveller discussion, the SF-tropish form of robot is not the default assumption for what is meant by the word 'robot', but as this is evidently not the case, I feel compelled to spell it out.​

Automated factories. Outgrowths of what we have today. My base point was that automation=fine, replacing crew member=more expensive or not reliable enough to trust with the role. That's my explanation as to why the majority of a merchant ship's crew are not robots.



I don't think we're in disagreement on anything here. You are right about the economic implications of the rules as written. I'd love to see your fixes. I'll try to share mine.

 

[Golan2072]

This thread might get me to get back to LBB8, after all these years in which I did not like it much due to over-complexity... The designs here are intriguing, to say the least.

 

[flykiller]

in discussing robots and their costs and capabilities one may wish to distinguish between ai and expert systems. in the traveller books robots clearly are nothing more than expert systems, more or less anthropomorphized, while the robots being discussed here clearly are ai, more or less thought of as mere machines.

 

[rancke]

in discussing robots and their costs and capabilities one may wish to distinguish between ai and expert systems. in the traveller books robots clearly are nothing more than expert systems, more or less anthropomorphized, while the robots being discussed here clearly are ai, more or less thought of as mere machines.

Maybe you're right. I'm relying on the opinions of others when it comes to LBB8 robots (ASMEs) and possibly I'm conflating them with MgT robots.


Hans

 

[samuelvss]

In discussing robots and their costs and capabilities one may wish to distinguish between AI and expert systems. In the Traveller books robots clearly are nothing more than expert systems, more or less anthropomorphized, while the robots being discussed here clearly are AI, more or less thought of as mere machines.

I would heartily agree! I believe that LBB8 opened a small :CoW:, but attempted to avoid the bigger :CoW: by kicking AI out to TL16. I can buy, in LBB8, for the price of a stateroom, a TL13 robot that can fly where I tell it to fly, shoot what I tell it to shoot, and perform other specific tasks according to instructions or programming: a system of expert systems, physically present together, but not truly integrated enough in executive function to be fully independent. It is not going to be a crew member, any more than the computer is. It can't think.

It will, however, be much better at certain discrete tasks than a human; these tasks may normally require a crew member. The robot performs that task, and can reduce the crew size, IFF the owner is willing to capitalize the thing, and deal with various hassles. It's not crew, though.

If we want to bash LBB8 on assumptions, we might much more readily look at whether true AI would not be present at something like TL10. If we look at the advances to date, project Moore's law out a few centuries. That to me is the most unrealistic. But I play Traveller, so I take the "science" that comes with my fiction with more than a grain of salt.

 

[Meeko100]

I would heartily agree! I believe that LBB8 opened a small :CoW:, but attempted to avoid the bigger :CoW: by kicking AI out to TL16. I can buy, in LBB8, for the price of a stateroom, a TL13 robot that can fly where I tell it to fly, shoot what I tell it to shoot, and perform other specific tasks according to instructions or programming: a system of expert systems, physically present together, but not truly integrated enough in executive function to be fully independent. It is not going to be a crew member, any more than the computer is. It can't think.

It will, however, be much better at certain discrete tasks than a human; these tasks may normally require a crew member. The robot performs that task, and can reduce the crew size, IFF the owner is willing to capitalize the thing, and deal with various hassles. It's not crew, though.

If we want to bash LBB8 on assumptions, we might much more readily look at whether true AI would not be present at something like TL10. If we look at the advances to date, project Moore's law out a few centuries. That to me is the most unrealistic. But I play Traveller, so I take the "science" that comes with my fiction with more than a grain of salt.

Eh...

Moores law hasn't been proven true for the past 5-ish years. Computer development is slowing down sadly, issues like heat.

Of course there is the ever snappy synapticss that apparently means computers will be able to process computer code by having complex sophont. Cause we process code better...

Yay for speculation!

 

[aramis]

Eh...

Moores law hasn't been proven true for the past 5-ish years. Computer development is slowing down sadly, issues like heat.

Of course there is the ever snappy synapticss that apparently means computers will be able to process computer code by having complex sophont. Cause we process code better...

Yay for speculation!

Heat, and electron tunneling. Electron tunneling looks to be the big reason... We've got the architecture down too far, from what I understand, to make it any smaller and still have reliability.

 

[Meeko100]

Heat, and electron tunneling. Electron tunneling looks to be the big reason... We've got the architecture down too far, from what I understand, to make it any smaller and still have reliability.

Yeah, I heard that as well. Forgot the word for it, was going to say something about voltage breakdown, or a lack thereof. Voltage from the current in the chip isn't being broken down/blocked by the insulators within as efficiently enough to take the voltages we are applying. Jumping ship to the next pathway over.

I suppose that if we invented some brand new semiconductor to make chips out of, then it'd work.

 

[whulorigan]

Yeah, I heard that as well.
...
Jumping ship to the next pathway over.

I suppose that if we invented some brand new semiconductor to make chips out of, then it'd work.

I would guess that even a new semi-conductor material will make minimal improvements at best, as the problem is that our chip-circuits are approaching the atomic-level size-range. Quantum tunneling is going to be a problem for any semi-conductor material at the sizes in question.

An improvement in computing is likely going to require a paradigm shift in our computer hardware and/or architecture, comparable to how semi-conductors replaced vacuum-tubes.

Quantum Computing has been experimented with and gives significant improvement over standard computers in terms of database search-tasks and encryption functions, but does not seem to be much faster in terms of normal computing tasks.

Very basic Organic (nucleic-acid based) computing has also been experimented with, but i am not sure how much of a potential speed-improvement it might be over semi-conductors.

 

[aramis]

Quantum Computing has been experimented with and gives significant improvement over standard computers in terms of database search-tasks and encryption functions, but does not seem to be much faster in terms of normal computing tasks.

It's also paying a hell of a mass penalty - quantum computing requires a lot of thermal support (Chilled to below -200°C), as well as massive power use (both for the cooling and the computing)

Very basic Organic (nucleic-acid based) computing has also been experimented with, but i am not sure how much of a potential speed-improvement it might be over semi-conductors.

Got a citation for that one? I'd like to read more about that...

I suspect it's slower...

I've read that micro-vacuum-tube has potential for analog computing (bad pun). They're big, by comparison (individual tubes up to several cm wide, but a fraction of a mm tall), and can be used with much higher voltages than can semiconductor units. And they're nearly immune to EMP. But still smaller than quantum units. They hold a lot of potential for neural network use

Still, speed comes from short paths and fast switching.

 

[whulorigan]

Very basic Organic (nucleic-acid based) computing has also been experimented with, but i am not sure how much of a potential speed-improvement it might be over semi-conductors.

[FONT=arial,helvetica]Got a citation for that one? I'd like to read more about that...

I suspect it's slower...[/FONT]

I am remembering it from an article in Science News within the last 2 years, but I unfortunately do not have a reference off the top of my head. If I can locate it, I will post it.

 

[flykiller]

Heat, and electron tunneling.

and manufacturing limitations. churning out simple chips with a 10% defect rate is one thing. carefully laboring over highly complex and expensive machines to slowly turn out max capacity chips with a 75% defect rate is another matter altogether.

probably we'll find that a human mind to make decisions, coupled with good processors to carry out those decisions, is the best possible resolution.

I once ran a story where multiple expert systems, capable of operating independently in their expertise, were also subject to the rotating at-will control of a single human mind - that mind being imprisoned and having no other operational outlet other than the machines ....

 

[rancke]

I would guess that even a new semi-conductor material will make minimal improvements at best, as the problem is that our chip-circuits are approaching the atomic-level size-range. Quantum tunneling is going to be a problem for any semi-conductor material at the sizes in question.

Ah, how different things might have been if the people of the Traveller universe had access to technology more advanced than that of 21st Century Earth.

Oh, wait...


Hans

 

[flykiller]

if the people of the Traveller universe had access to technology more advanced

they do. but they still have the same periodic table and the same laws of physics ....

 

[rancke]

they do. but they still have the same periodic table and the same laws of physics ....

Sure. But they also have laws of physics that we haven't discovered yet.


Hans

 

[aramis]

Sure. But they also have laws of physics that we haven't discovered yet.


Hans

And something about their lanthanides does strange stuff instead of melting when a high enough current is run through it...

 

[rancke]

And something about their lanthanides does strange stuff instead of melting when a high enough current is run through it...

So perhaps the Traveller universe does not actually have all the same physical laws that the real universe does?

(Though I admit that that is a concept that should be used only with the utmost caution. Essentially only if it allows for a better game.)


Hans

 

[aramis]

So perhaps the Traveller universe does not actually have all the same physical laws that the real universe does?

(Though I admit that that is a concept that should be used only with the utmost caution. Essentially only if it allows for a better game.)


Hans

When enough energy is passed through lanthanum, it heats up, melts and/or vaporizes...

So, either the lanthanum hull grids are not element 57, or the physics of the OTU are different.

Perhaps Lanthanum is vilani for element 164 (unhexquadium), especially the isotope ⁴⁸²Uhq...