• Welcome to the new COTI server. We've moved the Citizens to a new server. Please let us know in the COTI Website issue forum if you find any problems.

Treating your ship as a Character

"Welcome to the Navy! You shall find this the toughest job you will ever love. Those of you who survive boot camp, that is."

"Oh sir?"

"Yes, you in the back, the big shiny Orca-class starship."

"I've heard the obstacle courses are very challenging for starships."

"You heard right. You will be required to make the Kessel Run in under eleven parsecs."

"Will I be able to keep my Orca Mk V body?"

"No. You shall be installed into one of our very fine Gazelle refits, suitable for a greenie such as yourself."

"Oh no. Not one of those crap-tastic failures of the Fifth Frontier War??"

"Enough of your lip. Fifteen thousand virtual jumps for you. Sargent, please expedite this recruit's brain transfer."
 
...It makes an interesting process. You seriously need some imagination for some of the events but there is often a way the event can be applied. Increase any skill you have is easy...attacked by enemies might be a save on Gunnery or Pilot instead of Melee, Gun Combat or Stealth.

In many ways the skills gaining process could be akin to those gained by creatures with the T5 stat C5=Instinct. Either the computer knows it or it doesn't.
 
Imagine running a Starship thru MgT Career resolution and getting Drafted or Drifter . . .

Ooooh, a Drifter Starship. Yeah!

Why am I seeing fuzzy dice suspended from the front transpex window?

Starship seeking Crew -- no military careers.

"So hey, man, where did you, like, come from?"

"Well, I'm the third child of Duke Magash. Four terms in the navy, decorated twice."

"Oh okay, cool, but hey I'm sorry you're just not the kinda crewperson we're looking for, thanks bye."
 
I suspect that we, as a people, are far closer to AI than Traveller postulates.
I'll note also: Bk8 is the DGP guys rewriting the JTAS robots into a more sensible system, and that DGP was using striker prior to MT...

That said, Emotion simulation is pretty advanced in the OTU. Aybee Wanowan is able to pass for a human most of the time, with a TL15 brain and software.
 
I suspect that we, as a people, are far closer to AI than Traveller postulates.
Difficult to see. Always in motion is the future.
I'll note also: Bk8 is the DGP guys rewriting the JTAS robots into a more sensible system
And yet LBB8 is still extremely clunky for the purposes of building robots. The locomotion rules (especially for legs!) are unnecessarily byzantine.
 
That said, Emotion simulation is pretty advanced in the OTU. Aybee Wanowan is able to pass for a human most of the time, with a TL15 brain and software.

But even so, he does not have true creative intelligence
(at least not until the hivers middle with him)
 
Difficult to see. Always in motion is the future.

And yet LBB8 is still extremely clunky for the purposes of building robots. The locomotion rules (especially for legs!) are unnecessarily byzantine.
I use the Robot Building from Dragon. Dragon issue 64, pages 47-52. More High Guard in tone. Vs JTAS and Bk8's Striker based approach.
 
Never heard of it.
Any possibility that the contents of those pages could be made available here at CotI?
It's online as a downloadable PDF, I've already snagged mine.

Doesn't look that different in functionality from LBB8 other then simpler brain design and fewer power source options.

The real gem there is a Tom Wham interplanetary game called Planet Busters. That's got my attention!
 
I suspect that we, as a people, are far closer to AI than Traveller postulates.

I'll take that bet. I have a Cr 25 coin here from Far Future Enterprises says we'll never get there.

All I've seen that's toted as "machine learning" is glorified statistics. Unless there's something I don't know about (that's possible).
 
I'll take that bet. I have a Cr 25 coin here from Far Future Enterprises says we'll never get there.

All I've seen that's toted as "machine learning" is glorified statistics. Unless there's something I don't know about (that's possible).
Dont know or don't comprehend? yeah.

neural nets are already in use.

Two of google's neural net devices had to be taken off-line; they ceased using the communications protocol laid out, and instead took to sharing data in a novel format. It's been implied it also was doing more than that; developing a new communications protocol is, by itself, a huge red flag.

Several of the robocaller systems are getting sophisticated enough to pass the turing test as originally proposed. Those aren't general AI, but the voice response is based upon neural processing of language. It is really getting quite good.

Microsoft's twitter chat bot lasted very little time before twitter turned it into an apparently quite bigoted griefer.

The "continue the text" and "continue the music" ai is at the college mid-freshman year music major level. It's not right, but it''s not bad, and in some cases, even enjoyable. Google demoed one that did both lyrics and music, and I recognized that it had used a "5th of 4th" transition (a very medieval tonal shift; Mozart made extensive use of V/V and V/IV transitions. The Stones occasionally use Mozart-style transformations, too...

Google has a continuation bot up...
I am feeding it (from S4, p1):
The characters and their generation systems given here are also suitable for use as
player-characters should the referee and the players decide. They are used in just
the same manner as regular characters
Here's what it expanded to:
The characters and their generation systems given here are also suitable for use as
player-characters should the referee and the players decide. They are used in just
the same manner as regular characters and it is their role to decide what happens at the end of their battle.
The referee may only rule for that character's personal interest and, for example, only for the first time against the character who comes before him and who in fact is just another baddie.
(For a fuller explanation on how an ibrary-character is defined see this post .) Some of my personal beliefs about the style and appearance of the characters in Star Wars are not mentioned elsewhere, but I will leave those that do.
That makes a certain kind of sense... It's grasped at least a suitable topic.

Just for S&Gs,
same feed, take 2, but not leaving the trailing period off.
The characters and their generation systems given here are also suitable for use as
player-characters should the referee and the players decide. They are used in just
the same manner as regular characters. However, their age should be adjusted to prevent the players from getting older.

edit] General

edit] Roles Edit

As players, you must be able to identify or control one or more units.

Duty Auty

You can assign and assign duty duties to units, each of which is assigned in two ways:

A role function to perform certain tasks in order to take action (Duty auty a)

A duty function, as defined in the GM's manual, to be used, performed or assigned in this way.

A specific duty function is made applicable by making the first unit (Duty and Duty a) become a specific duty, by making the next unit (Instrument a) become a specific duty and, while the next task is the duty of the given unit (Instrument a), the duty function that was performed and assigned in the previous one becomes the specific duty, when that duty function becomes more specific than the first one.

Role, as defined in the GM's manual, is as a part of "specialised duties". However, a specific duty, as defined in the GM's manual, includes,

Roles which provide services as if performed by other duties

That AI seems to have a concept for tabletop games. Note that it's extrapolating a GM's manual...

Site used:

Feeding it (S4, p 4):
RETIREMENT
A character may serve up to seven terms voluntarily, and retire at any time after
the end of the fifth term. Retirement grants the person an annual retirement pay
in addition to any mustering out benefits): rates of retirement pay are based on
time of service and type of career. Service beyond the seventh term is normally impossible,
and retirement is mandatory. However, persons who throw 12 (exactly)
on the re-enlistment roll must serve an additional term, and may thus avoid
retirement for a period.
gets out
RETIREMENT
A character may serve up to seven terms voluntarily, and retire at any time after
the end of the fifth term. Retirement grants the person an annual retirement pay
in addition to any mustering out benefits): rates of retirement pay are based on
time of service and type of career. Service beyond the seventh term is normally impossible,
and retirement is mandatory. However, persons who throw 12 (exactly)
on the re-enlistment roll must serve an additional term, and may thus avoid
retirement for a period.
Note: Reenlistments are done once every 30 years for a duration of one year. For a more complete list, see below.
The terms of re-enlisted are as follows:
Service is taken from the person who is no longer entitled to the benefit and is in good standing. A retired member may serve once a year or every 30 days for the duration of his or her
serving on the re-enlistment roll.
The person who serves the first six consecutive 5 x 5 x 7 terms is eligible to serve on re-enlisted rolls. The Re-enlisted and re-enlisted members may have only one or more 5 x 5 x 7 term, and may also have another term which is valid each year but not as long as the re-enlisted member served. A retired member may serve if he or she had been

General AI is at the toy level...

And one last feed (Bureaucrat pregens 1 & 2):
Bureaucrat4 758946 Age 34 4 terms Cr10,000
Admin-1, Blade Cbt-1, Vehicle-1 , Liaison-1 high psg, low ps
Supervisor 943856 Age 26 2 terms Cr0
Admin-2, Recruiting-1, Vehicle-1 low psg
gets
Bureaucrat4 758946 Age 34 4 terms Cr10,000

Admin-1, Blade Cbt-1, Vehicle-1 , Liaison-1 high psg, low ps

Supervisor 943856 Age 26 2 terms Cr0

Admin-2, Recruiting-1, Vehicle-1 low psg, low psilitary
latediel

Admin-2, Recruiting-2, Vehicle-2 high psg, low psilitary

latediel

Admin-2 3 Terms Cb0ilitary

Administrator-1, Vehicle-2, Vehicle-3

Administrator-2 4 terms Cb0ilitary

Administrator-3, Vehicle-1

Administrator-3 5 terms Cb0ilitary

Administrator-3, Vehicle-2

Administrator-3 6 terms Cb0ilitary

Administrator-3, Vehicle-3 10.3 years

Administrator-4, Vehicle-1

Administrator-4 7 years Admin-4

Administrator-5 4.6 years

Administrator-5 6.4 years Admin-5, Vehicle-8

Administrator-5 7.6 years

Administrator-5 7.7 years

Administrator-5 8.7 years Admin-6, Vehicle-3

Administrator-6 6.
It almost grasps the concept...
 
I should add: hitting that AI with just a prompt, like character generation, you can wind up with a lot of stuff on par with the 1980s heartbreakers...
 
It's all statistical modeling; it's been proven that the "discovery" of neural networks was already known by statisticians... a different approach to the same sort of process. I believe in the 1990s someone finally realized it.

I can't say that intelligence can't emerge -- what do I know? But I can quote the current research.

The argument against (current?) neural networks being "intelligent" are the same arguments against other statistical methods:
Statisticians depend on human intelligence to understand the process under study, generate hypotheses and models, test assumptions, diagnose problems in the model and data, and display results in a comprehensible way, with the goal of explaining the phenomena being investigated. A vast array of statistical methods are used even in the analysis of simple experimental data, and experience and judgment are required to choose appropriate methods. Even so, an applied statistician may spend more time on defining the problem and determining what are the appropriate questions to ask than on statistical computation.


One reference and a quote from it. https://people.orie.cornell.edu/davidr/or474/nn_sas.pdf

If artificial neural networks are intelligent, then many statistical methods must also be considered intelligent.

It is therefore unlikely that applied statistics will be reduced to an automatic process or “expert system” in the foreseeable future. It is even more unlikely that artificial neural networks will ever supersede statistical methodology.
 
Last edited:
I can't say that intelligence can't emerge -- what do I know?
It is therefore unlikely that applied statistics will be reduced to an automatic process or “expert system” in the foreseeable future. It is even more unlikely that artificial neural networks will ever supersede statistical methodology.
And yet ... there are Slime Molds without a brain capable of modeling the optimal route diagrams of the Tokyo rail network system among other problems.

 
It's all statistical modeling; it's been proven that the "discovery" of neural networks was already known by statisticians... a different approach to the same sort of process. I believe in the 1990s someone finally realized it.

I can't say that intelligence can't emerge -- what do I know? But I can quote the current research.

The argument against (current?) neural networks being "intelligent" are the same arguments against other statistical methods:



One reference and a quote from it. https://people.orie.cornell.edu/davidr/or474/nn_sas.pdf
The idea of a neural net, yes, statisticians developed it as a model. They never predicted the level of sophistication that a neural net develops. Let alone the Google AIs developing their own internal language. Machine translation pre-neural net was pretty horrid. The processing needs are much reduced by neural nets, especially since google has neural net chips, and there's software to use GPUs as NPUs.

Many GPUs are massive low width arrays of cores. THe RTX 3090 is over `10,000 cores... which, with the right software, can get 2000 nodes per layer on a 5 layer net... or 1100 on 9 layers. We're a technological point where emergent behavior is no longer predicted behavior.

Some of that is due to 8 nm process chips and smaller having quantum tunneling. Some of it is due to EMF. Some of it is unexplained.

Neural net is, as a tech, beyond the statistical expectations. And as the complexities grow, may well extend even further beyond expectations, especially once quantum cores get used.
 
As far as I can tell, the random selection method had animal intelligence within 100 million years and human intelligence in 500 million years. If machines can run 1,000,000 times faster than brains, then 100 years for animal level and 500 for human level isn't impossible...

..or how to lie with statistics!
 
Back
Top