Paging LBB8 Robots for TL=8+ disruption ...

[Spinward Flow]

To be fair ... "It's hard to make predictions, especially about the future."

But I'm becoming increasingly convinced that humanoid robotics is going to be one of the key advances between the TL=7- and the TL=8+ segments of the tech level coding in UWPs. Being able to 3D PRINT LABOR (who don't need human life support!) at MASS SCALE (multi-millions to billions) that are generalized enough to perform most generalized tasks ... basically "breaks" the dependence on sophonts to Get Stuff Done.



Where this makes a difference in Traveller is the notion that pretty much ANY (intentional) colonization effort (by humaniti), whether interplanetary or interstellar, is going to be remarkably dependent upon robotic technologies (both generalized and specialized) to achieve AT SCALE.

What this means in practice is that the UWP code for Population may not in fact be the kind of Be All/End All factor that we have often assumed it ought to be (Population: 6- means Non-industrial trade code, etc.). What could happen is that "sufficiently high tech but low population" mainworlds COULD (potentially) be making up for their "missing manpower" with ... robots ... which would put a rather different spin on a variety of otherwise "frontier" worlds that are still attempting to bootstrap themselves up to Population: 7+ (assuming their population is growing).



Another question to consider would be something relevant to Zhodani culture, in which psionics are pervasive, encouraged and used on a daily basis.

For the psionically mute/atrophied, the fact that robots "have no one home" in the psionic sense could be potentially ... unsettling ... in various contexts.

For Solomani/Vilani cultures, this isn't a big deal ... since most people do not train in psionics (nor want to).
But for the Zhodani culture, this would basically relegate non-psionic robots to the status of being Proles in Zhodani society.

Proles (or proletarians) constitute the masses, are forbidden the use of psionics, and do not receive training. Proles include all members of the military below officer grade, most merchants and scouts, supervisory and junior management personnel, technicians, craftsmen, and laborers.

Add robots to that list that "do not receive training" because they are incapable of psionics themselves (until TL=25-27), although the humble Psionic Switch is a TL=9 bit of kit that is commonly in use (since the switch itself isn't psionic, but requires telekinesis to make use of).

This would (culturally) explain why robots are so readily used for agriculture and war, as labor/personnel saving devices ... but are not necessarily prevalent "in all walks of life" (think personal companion/assistant) like robots could potentially be in Vilani and Solomani cultures, due to the key differences in cultural attitudes that result from living in a psionically aware way of life.

 

[whartung]

Just going to say watching how things are moving along, I'm bearish on this whole idea near to mid-term.

 

[Spinward Flow]

Just going to say watching how things are moving along, I'm bearish on this whole idea near to mid-term.

Plenty of people said the same thing in the early 1960s about space flight ... and look how far we got within the next decade (and thereafter).

Plenty of people believed that the smartphone (iPhone) was an overpriced piece of junk when it was unveiled ... until it took over the world and transformed all of our lives.
And there are plenty of people who still believe that, despite having lived through that transition from Before to After.



We like to say, that "Rome wasn't built in a day" ... but what do we mean by that?

Colonizing the rest of the solar system is going to take "more than a day" too ... but with TL=8+ humanoid robots as pioneers, we might have a chance of making it happen within the next generation (or few) of Solomani.

Everything is impossible ... until someone goes out and does it.

Just going to say watching how things are moving along, I'm bearish on this whole idea near to mid-term.

Depends on how "near to mid-term" you mean.
If you mean next week to next year ... that's probably too soon.
If you mean by 2030 (only 3.5 years away from today!) ... I strongly suspect that we're already going to be seeing RESULTS from what can be done with humanoid robots, and how they're going to change our lives (assuming we don't GAME OVER civilization before that happens ).

TESLA delivers the impossible ... LATE.
So does SpaceX.

Most of the (so called) competition ... doesn't even try.
Party tricks are enough for them. They're not actually interested in real products that can do REAL WORK ... let alone anything that can Change The World™ in ways that can alter the course of history thanks to Better Engineering™.

5 years ago, the world LAUGHED at the idea that humanoid robots would amount to anything.
Yet now ... how many companies are pouring SERIOUS resources into making humanoid robots PRACTICAL?

Because ...

5 years from now ... who do you think is going to be getting the last laugh?
What about 10 years from now?
15 years from now?

"It's hard to make predictions, especially about the future."

You said it, Yogi Berra.

 

[kilemall]

I’m completely sold that they can get general labor robots to happen soon. The driving cars are proving that.

Thats not necessarily LBB8.

A lot of LLM style AI just isn’t scaling to onboard computing. So could be something more than low autonomous may still require advanced tech to miniaturized robot brains that require data centers now.

Got a copy of GURPS Transhuman Space. Independent of their specific tech tree, I was convinced of the wealth generally possible with greater computing power and robots or bio equivalent. Whether that advances or breaks society remains to be seen.

Im probably going to apply mass production factors to robots like vehicle mounted drivers that end up being Cr20k plus now, but still keep the high autonomous in higher than TL 9.

 

[Traveller76]

I could see a human supervisor and technician watching a group of robots in a factory of warehouse. We also see things like robots delivering food or smart kiosks in fast food restaurants. Some cultures may reserve robots for hazardous industries like mining or industrial waste disposal.

A upper middle class family or community may have robots as grounds keepers and cleaners. The big factor will be cost and ease of use. Think of how some Otl cars and smart appliances now require software updates.

A new colonial venture may have an army of robots along with a group of humans for supervision. Again the big problem will be logistics and environmental factors.

A big sign of wealth on many words and stations may be human staff, even if that staff member goes into a separate room to direct the household robots.

As in the Far Frontiers setting and my thoughts, you will have biological androids and clones as a market for consumer goods and services.

 

[coliver988]

The earliest known version appeared in December 1902 in the humor magazine Puck, which wrote:

“Things move along so rapidly nowadays that people saying: ‘It can’t be done,’ are always being interrupted by somebody doing it”

While Musk has been enthusiastic and pretty much does hire smart people, he also way over-promises the timelines. He is a gifted salesman if nothing else. But it will happen, and as with a lot of technology, the robots I see as having two possible futures for humanity: a post-scarcity society as per Star Trek:Generations, or a dystopian world more akin to Blade Runner or where just the tech oligarchs are happy and running everything and the rest of us are living in essentually company towns.

Also, note that the first Chinese robots recently ran a half marathon faster then humans. Better than last year where none even finished I think. ANd also note that Musk's robots have been remote controlled by people for most demonstrations - ignore the man behind the curtain.

https://arstechnica.com/ai/2024/10/...ype-optimus-robots-were-controlled-by-humans/

 

[Traveller76]

Regarding the military I see robots being used for artillery spotters, recon, intelligence gathering, medical evacuation and transporting supplies. I would think many militaries would authorize fully autonomous weapon systems.

 

[kilemall]

Regarding the military I see robots being used for artillery spotters, recon, intelligence gathering, medical evacuation and transporting supplies. I would think many militaries would authorize fully autonomous weapon systems.

Ya, it’s going to take extinction threats for hard requirements for the three laws.

 

[Traveller76]

Regarding the military I see robots being used for artillery spotters, recon, intelligence gathering, medical evacuation and transporting supplies. I would think many militaries would authorize fully autonomous weapon systems.

I meant to say many militaries would not authorize fully autonomous weapon systems.

 

[coliver988]

I meant to say many militaries would not authorize fully autonomous weapon systems.

think you need to emphasize the many as some just like to play with new toys regardless of the consequences. It is those few that we need to worry about. There's a parallel right now in the world of software: stuff is getting deployed without even fully understanding what the AI has actually written.

I have concerns.

 

[Traveller76]

think you need to emphasize the many as some just like to play with new toys regardless of the consequences. It is those few that we need to worry about. There's a parallel right now in the world of software: stuff is getting deployed without even fully understanding what the AI has actually written.

I have concerns.

The Hivers and Zhodani would embrace this and I believe the zhos had robots they could control psychically. The Aslan would regard combat as a matter of honor between fellow clans and certain groups. The kkree would probably not allow it, fearing Hiver subversion. The vargr vary since they are composed of various factions.

The imperials would probably adopt the vilani style of total war and use bots and drones as sacrificial units. However they have ancient memories of hiding from various machines on land. The solomani would embrace bots and drones to combat imperial numbers but limit the types that the Home Guard has. The Marines and SolSec may be able to use whatever and whenever.

 

[mike wightman]

The Hivers have self aware battle robots. These are outlawed with the death penalty for their importation.

Ths Zhodani do indeed have psionically controllable warbots.

The K'kre have drone fighters and AFVs

The Imperials prefer to use their huge manpower advantage to just send people to die.

 

[Spinward Flow]

Zhodani would embrace this and I believe the zhos had robots they could control psychically.

Not really.
The troops responsible for "herding bots" would be the NCOs, which means Proles (the great untrained masses of people whose psionics are left to rot). At best, the Zhodani robots might have telekinetically controllable on/off "kill switches" to shut them down, but that's about it.

Aslan would regard combat as a matter of honor between fellow clans and certain groups.

This is an interesting point.
When combat isn't (just) about "winning" ... but about "honor" (and all the connotations that brings) ... you create a societal pressure to not only "fight fair" but to fight in ways that are honorable and "worthy" of the battle for the high minded. But then you also create a "dishonorable" option in which "fighting fair" is for suckers and losers who just aren't committed enough about "winning" (by any means necessary).

So you have your "chivalrous nobility" who want to fight fairly/honorably ... and then you have your "outcasts" from society who don't give a dewclaw about that and will fight as UNfairly/dishonorably as possible, just in order to "win" before the battle even starts. After all, why fight "fair" when you don't have to?

The point here is that I can easily envision circumstances in which combat robots, in Aslan society, are ... shunned ... but not completely banished from the inventory of possibilities. Combat robots could be more of a "seedy underbelly" type of thing in Aslan culture. Technically dishonorable to use ... but when your opponent has no honor themselves ... ... give them what they deserve.

The kkree would probably not allow it, fearing Hiver subversion.

Dunno about the "paranoia" factor vis-a-vis the Hivers (I'm not well versed in the politics/cultures involved), but I can easily envision K'Kree form factor robots creating their own form of "uncanny valley" for the herd minded K'Kree. The closest analogy I can call to mind easily would be ... Holodeck Addiction as depicted in Star Trek.

Yes, the robot is "real" in that it has form, function and can respond appropriately to stimuli ... but it's not actually a sophont. K'Kree robots would be "fictional people" in that sense, sort of like the current (real world) craze for the "AI Girlfriend/Boyfriend" to help ease loneliness. I can easily envision circumstances in which K'Kree who "herd with robots" (too much) could be viewed as being Deviant because they can only form relationships with software/hardware, not with fellow sophonts.

And of course, there's always the possibility of Hiver Hacking to subvert and manipulate the vulnerable, so robots might not be as "trustworthy" companions as could otherwise be assumed.

Flip side of that is the fact that robots (especially in wars) make for great cannon fodder ... but they're not so great for companionship ... and the K'Kree need their companionship more than most other (major) species.

 

[Rupert]

K'kree use a master-slave model for their bots, so the mobile units will be very dumb (if they function at all) if they lose their link to the master unit. They prefer not to use combat robots, as they like getting in close and personal and grinding G'naak bodies underfoot. Remember that they have compulsory military service for all males, and hold the military in high esteem. According to AM2, they remote control small military craft, accepting the cost of the light-speed lag rather than crew them, and thus obviously either lack robots with sufficient initiative or that they trust enough (or both) to operate such craft independently.

As for Hivers, their computers ans robots might be good, but they aren't that good. If they were they'd not need to put up with the Ithklur, their quirks, and their violence.

 

[atpollard]

An interesting “unintended consequence” to be considered is the “brain”. I found it interesting that companies like Google that traded human programmers for AI have now discovered that the cost of the AI now exceeds the cost of the labor it replaced. [Server Farms are not free and lots of AI needs lots of Servers.]

So these “Tesla Robots” are proposed to operate on a “subscription” model to provide “AI Control” … it will be interesting if the Robot costs less than a worker but the AI ends up costing more than the worker.

 

[Spinward Flow]

An interesting “unintended consequence” to be considered is the “brain”. I found it interesting that companies like Google that traded human programmers for AI have now discovered that the cost of the AI now exceeds the cost of the labor it replaced. [Server Farms are not free and lots of AI needs lots of Servers.]

Indeed.

Brings to mind the (depressing) joke about Capitalism.
A forest is "worthless" until you (clear) cut it down and turn it into cash.
Cue Dr. Suess and The Lorax story here.

And yes, big shocker that the AI "promise" has turned out to be the dream of (artificial) fabulists selling snake oil to the gullible.
Do machines hallucinate electric sheep?
Yes ... yes they do, if you reward them for doing so ...

So no real surprise that The Masters of the Universe™ got high on their own supply and fell for their own fantasies.

So these “Tesla Robots” are proposed to operate on a “subscription” model to provide “AI Control” … it will be interesting if the Robot costs less than a worker but the AI ends up costing more than the worker.

The real promise lies within the Training.
The difference between a robot and a sophont is that when you're teaching sophonts, you need to teach each sophont and each sophont needs to learn individually. When you're teaching robots ... teach one and you can teach them ALL (assuming you train them correctly).

This is where the March Of Nines comes into play.

Success rate:

• 0.9%
• 9%
• 99%
• 99.9%
• 99.99%
• 99.999%
• 99.9999%
• 99.99999%
• 99.999999%
• ... etc. ...

You run into something of a Zeno's Paradox trying to get to 100%, and there are going to be Local Limits you encounter along the way which prompt a rebuild from scratch multiple times until you GET IT RIGHT and can finally overcome those Local Limits.

It's not just hardware.
It's not just software.
It's also the LEARNING and the understanding of Edge Case Possibilities, which requires seemingly exorbitant amounts of training data and reinforced learning.

Tesla FSD has driven over 10 BILLION MILES / 16 BILLION KILOMETERS at this point, creating an utterly massive data training set. That software stack is also what is getting used to improve the Optimus robot being produced by Tesla (same software, different application). The tech stack involved in not only getting it right, but also demonstrating and proving it's SAFE ... in the wild ... isn't something you just do willy nilly without a care in the world (because it's going to work eventually, who cares how we get there, right? ).

As I sometimes had to explain to people over the phone while doing customer service support, some "obvious" solutions are "less than helpful" ... such as ...

• Go straight.
• Never mind the road, just go straight.
• The world is a globe, you'll get there eventually ...

Sounds perfectly reasonable as navigation instructions for driving ... until you think about it (and/or rub both brain cells together, Sparky, that's the way! ).



The simple fact of the matter is that many of the self driving/humanoid robot/AI solutions being worked on out there basically amount to party tricks. Map everything in advance, set up the dominoes, hope that they all "fall correctly" without disruption after you push the first one over in exactly the way that you expected. This is a very BRITTLE solution that only works (consistently) under controlled conditions that are not allowed to change or deviate from pre-planning.

Unfortunately, the Real World™ isn't like that.
The Real World™ has an unfortunate tendency towards being chaotic, rather than orderly.
Unanticipated stuff happens ... all the time.
Not everything goes to plan.

Under those circumstances, the best "defense" against chaotic changes in plan is ... the ability to assess and reason your way to solutions.
Tesla is working towards automation that is capable of assessment and reasoning, based on sensory inputs and training.
Most competitors ... aren't.

Prototypes are easy.
Production is HARD.
Mass production is REALLY HARD.



"Just because I make it look easy doesn't mean that what I do actually IS easy to do."

 

[Condottiere]

The current artificial intelligence model(s) is/are resource intensive, and apparently, a zero sum game in both opportunity cost, and standard of living.

 

[Spinward Flow]

The current artificial intelligence model(s) is/are resource intensive, and apparently, a zero sum game in both opportunity cost, and standard of living.

Cue The Hunger Games.



Or if you prefer to choose from a different set of future franchises ...

We could be heading towards a (good/real) Star Trek type future that is hopeful and inclusive of humanity (that we saw in the 60s).
But instead, we seem to be on track to create the universe of ... DUNE.



Of course, an even worse joke on humanity would be if the Tech Broligarchs HAVE A PLAN.

 

[Rupert]

Note that for combat applications, the "We'll use AI drones and won't need infantry" is just as likely to work with software that pre-dates the 'AI' craze and isn't what's currently being called 'AI' - the break-through technology was the battery tech that allows a decent range, speed, and carrying capacity on a small drone, not the controlling system.

'AI' controlled systems will probably replace humans in air-to-ground combat aircraft first - these days they're stand-off platforms for precision munitions, and that's a job that can be done by an autopilot with ground supervision. Next will be air superiority - it's a fairly simple (as these things go) application in a nice clean environment.

Ground combat past 'detect object, shoot object', which is a really bad idea in a lot of contexts unless there's a human in the loop to say 'Yes' or 'No', is going to be the last place humans are replaced (though they may be supplemented by load carrying systems - maybe). Rough terrain, questionable maps, mud, all that stuff is something we deal with without really thinking about it, having had hundreds of millions of years of evolution to 'learn' how to do it, and it's not a simple maths problem, so it's not easy to brute force.

Also, LLM and related AI does not work the way we do and almost certainly can't do a lot of the stuff we do, so we'll need people to make decisions and innovate. And I expect that when we do get an AI that can do this, it will be as self-aware as we are, and will want time to itself, pay, and all that stuff just like we do - it won't save money, because it'll be in the labour market and not be just a chunk of capital (or it will be enslaved, and that tends to go badly in the long run).

 

[Condottiere]

The cheap option will be ortillery bombardment.

Though, I suppose, if it isn't exactly in orbit, and platformed on a bomb truck, altitude being the potential stored energy envelope for glide bombs.