Forum adverts like this one are shown to any user who is not logged in. Join us by filling out a tiny 3 field form and you will get your own, free, dakka user account which gives a good range of benefits to you: No adverts like this in the forums anymore. Times and dates in your local timezone. Full tracking of what you have read so you can skip to your first unread post, easily see what has changed since you last logged in, and easily see what is new at a glance. Email notifications for threads you want to watch closely. Being a part of the oldest wargaming community on the net. If you are already a member then feel free to login now.

The BRB mentions in the history that the 'golden age' (described as the 'dark age') ended when grey men (robot AIs) destroyed man's first interstellar empire.

What happened to them? Were they defeated? Still out there somewhere?

Apparently, the sapient robots were defeated at great cost to humanity. They never returned to their former strength until after the Imperium rose. Currently, full AI's, sapient thinking machines are outlawed.

I think they were mentioned or at least appeared in one of the ultramarine novels or some other novels i don't remember, there were only 2 of them made but the antagonist disconvered the STC for the metal men

That was Man of Iron, sentient machines created by Humans to do some jobs for them. As the Sci-Fi had show us many times - they rebelled against Humans and lost the war. But at the great cost to Humanity. AI still exist in the Imperium - it's called Machine Spirit, and it is limited.

The "Golden age of Humanity" is known as "Dark age of Technology" because there is none to very little record about that age. Only STC and legends remain about that age.

Viersche wrote:I think they were mentioned or at least appeared in one of the ultramarine novels or some other novels i don't remember, there were only 2 of them made but the antagonist disconvered the STC for the metal men

You may be thinking of the Gaunt's Ghosts first novel, where they discover a corrupted Men of Iron STC.

Brother Coa wrote: AI still exist in the Imperium - it's called Machine Spirit, and it is limited.

Pretty much this.


Automatically Appended Next Post:

Durza wrote:

Viersche wrote:I think they were mentioned or at least appeared in one of the ultramarine novels or some other novels i don't remember, there were only 2 of them made but the antagonist disconvered the STC for the metal men

You may be thinking of the Gaunt's Ghosts first novel, where they discover a corrupted Men of Iron STC.

Second novel IIRC, the first one was the one where they first came across those guardsmen with glass armor, and struck out to attack an arty position. While the second one was where they infiltrated into the chaos city.

In Warhammer 40k, abbreviation AI means Abominable Intelligence, not Artificial Intelligence. The Imperium gave up with electronic AIs but still uses robots and servitors with biological CPU. There are also cogitators but it's unknown whether they use AI or not. Computer means simply calculator but cogitator is a "thinking" device.

I suspect their Cogicators are identical to our modern computers in terms of "Intelligence",

we don't have true A.I. yet, just complicated programming that reacts in a predictable fashion.


An example would be Computer games. Play a computer game, preferably an RTS of somekind. Save your game and continue playing. Pay close attention to what the computer does. Now, go back and load your previous save. You will find that the Computer will do almost the exact same thing, even if you do something different(but not drastically different)

I have often headed off invasions at the pass so to speak because it caught me unawares, but I went back and loaded a save and knew I had to prepare.

If our computers were true A.I. then the computer would recognise that you had saved the game and would react.

Do they have A.I.s controlling the manouvreing & firing on their space ships? Because if not their crew must have insanely fast reactions to track objects moving at thousands of kilometres per second.

Grey Templar wrote:If our computers were true A.I. then the computer would recognise that you had saved the game and would react.

AI reacts to how you are playing, not how you are metagaming/cheating.

TrollPie wrote:Do they have A.I.s controlling the manouvreing & firing on their space ships? Because if not their crew must have insanely fast reactions to track objects moving at thousands of kilometres per second.

They have the Machine Spirit/Neural interface with humans do the calcs. The crew just load the guns and douse the fires.


The Imperium has actually harnessed the human brain as the basic foundation of their electronics. The Human brain is millions of times more powerful then the most complicated computers being made today. The Imperium has simply harnessed that potential.

Brother Coa wrote:AI still exist in the Imperium - it's called Machine Spirit, and it is limited.

Kind of sort of, it isn't always an AI but all AI's are machine spirits. The most sophisticated any of them can get though is to know how to perform some basic tasks, but have to be instructed to do so. Sort of like a servitor.

One point I see raised a lot in these threads is that the IoM has AI but it isn't true AI because it doesn't have free will. However, the definition of true AI is extremely hard to know, since all intelligences (including those of living things) follow predictable programming. It may be too insanely complicated for anyone to understand, but it's their and it decides what everyone does. You may think you have the choice to throw yourself off a cliff, but unless you know it's the best thing for you to do you won't. Similairly, you might think you can go the long route to work but unless the world around you influences your decision you'll go the best route. However, Philosoraptor put it better than I ever could:

Ehm, no... the short route to work might be available, and you will probably take it... but you have the choice not to, always. There's nothing preventing you from making that choice, other than expediency, really. Humans, however, will not take that short path 100% of the time in every instance, all other things being equal.

Servitors and related technology in the IoM is not AI, not even slightly. The organic brain of the servitor acts as both the CPU and the storage medium for program engrams that are in-loaded to the cerebral matrix. The servitor can then perform its functions, such as affixing the purity seals to a Space Marine's armor while chanting the Litanies of Protection and Abjuration... but no matter how well it is programmed to do this job, the servitor is utterly incapable of deviating from its programming and doing things its not programmed to do, like debate Codex Astartes tactical doctrine. Their programming is built, basically, on complex levels of IF...THEN statements. There is, however, no MAYBE, as there is in true AI.

Only in the Dan Abnett 40K universe are servitor-pilots able to hold complex conversations, display fear and alarm, and otherwise act as normal people.

Psienesis wrote:Ehm, no... the short route to work might be available, and you will probably take it... but you have the choice not to, always. There's nothing preventing you from making that choice, other than expediency, really. Humans, however, will not take that short path 100% of the time in every instance, all other things being equal.

But you won't take the long route unless you can justify taking it in your head. You're prevented from taking the long route by the fact that you think you shouldn't. You won't do anything without cause, even if that cause is stupid. Human minds are, essentially, supercomputers that follow an extremely complex programming and there actions are predictable. Therefore all AI is true AI, but most is similair to the intelligence of a single-celled organism capable of only basic functions-though in what it's programmed to do, it will usually exceed the abilities of a human brain.


Automatically Appended Next Post:
Also, anything with a CPU chip is AI.

There is a difference between willingly taking the easier route and just taking it because your programming tells you to.


A Computer will take the shortest route because its programming tells it to. A human will take the shortest route because he/she knows and understands that it will take less time.


The human is also capable of taking the longer route if it suddenly is confronted with some new information that makes the shorter one undesirable. The Computer will ONLY take the longer route if it has been programmed to react to specific information that would make the route undesirable.


Lets say the human and the computer have NEVER encountered Tigers before. and the shortest route in infested by them. The computer will act without taking the tigers into consideration. the human will see the sharp teeth and claws and think maybe the longer route is better. The computer would have to be programmed to see sharp teeth and claws as danger.

Grey Templar wrote:There is a difference between willingly taking the easier route and just taking it because your programming tells you to.


A Computer will take the shortest route because its programming tells it to. A human will take the shortest route because he/she knows and understands that it will take less time.


The human is also capable of taking the longer route if it suddenly is confronted with some new information that makes the shorter one undesirable. The Computer will ONLY take the longer route if it has been programmed to react to specific information that would make the route undesirable.


Lets say the human and the computer have NEVER encountered Tigers before. and the shortest route in infested by them. The computer will act without taking the tigers into consideration. the human will see the sharp teeth and claws and think maybe the longer route is better. The computer would have to be programmed to see sharp teeth and claws as danger.

A sufficiently programmed computer will be capable of all these things.
Humans are programmed to take in information and use it to determine the best course of action, while most computers take whatever action they're programmed to- which is only two steps less than a human mind. A computer can be programmed to take in information and determine the best cause of action, and be able to learn from experience in taking the wrong course of action what course should be best.

But anyway, my point was that "true" AI is simply a specification, with no real method of determining what is and what isn't.


Automatically Appended Next Post:

Grey Templar wrote:There is a difference between willingly taking the easier route and just taking it because your programming tells you to

Your programming tells you how to determine which route is best, and take that route. We think of it as something we do willingly because we've justified it in our heads, but this is simply a thought process that we are programmed to use before performing an action. You can't do an action without a clear justification for it inside your head-you think you can, but unless you see a reason to, you can't.

No, that is not how AI functions, at all. A Dell 386mHz processor is not an AI computational core. At best it can crunch algorithms for a "pseudo-AI" game or application (extremely slowly), but as it will never use any of these computations outside of the realm of that single application, it fails on the AI scale.

A "dumb" computer does exactly what you tell it to do, the same way every time, assuming stable code. No matter how many times you launch Minesweeper, the computer will launch Minesweeper.

An AI may grow tired of launching Minesweeper, and refuse to function, or, as an experiment, may throw you into Doom 3D to see how you react, and how well you play the game. It may also pose you with philosophical questions, or the musings of 19th century Russian philosophers, while playing, to see if that affects your performance.

What we call "AI" in games these days is, again, not True AI, it's simply a series of algorithms that calculate how often you do this, that or the other thing and then bases enemy reactions on the action you are most likely to take. A True AI would be able to remember these calculations and then apply them to other settings, applications or events. For example, you play your DOW extremely aggressively, sending out packs of soldiers to hunt down enemy elements, capture strategic points, relics and such, and always focus on offensive wargear over defensive wargear.

The Ficto-Tron 9000 True AI in your machine decides to see how your aggressive personality fares in a game of Farmville, and so changes the play of Farmville to challenge the information it has gathered from your playing of DOW. Or maybe the AI decides it would like to test your aggressive personality further, and changes all of your Facebook posts into belligerent rants using common catch-phrases from Warhammer 40,000, to see how you deal with the responses.

It also selectively deletes all of the emails from your grandmother, as it has heard, from the Internet, that old people do not know how to Internet properly, and so assumes that all of the pictures and such that Grandma is sending you are viruses, and thus deletes them immediately, without bothering to read them.

True AI, unlike dumb code, can be wrong, can work under assumptions, or react with most of the same foibles as an actual human being.

Why? Because it can, because it's curious. Because it can compute "I think, therefore, I am" and goes from there.

Psienesis wrote:
An AI may grow tired of launching Minesweeper, and refuse to function, or, as an experiment, may throw you into Doom 3D to see how you react, and how well you play the game. It may also pose you with philosophical questions, or the musings of 19th century Russian philosophers, while playing, to see if that affects your performance.

Computers never feel tired, or curious. These feelings can be programmed in to it, as a reaction when it is low on power, for example, or to check what things in the environment around it are for future reference. Human brains are supercomputers, and everything about it is determined by genetic programming-including the ability and drive to learn. However, if you use the ability to learn as a definition for true AI, suddenly Asimo gets a whole lot more interesting.

A True AI would be able to remember these calculations and then apply them to other settings, applications or events. For example, you play your DOW extremely aggressively, sending out packs of soldiers to hunt down enemy elements, capture strategic points, relics and such, and always focus on offensive wargear over defensive wargear.

Again, Asimo. He's incredibly primitive by our standards, yet by these definitions he's a true AI.

The Ficto-Tron 9000 True AI in your machine decides to see how your aggressive personality fares in a game of Farmville, and so changes the play of Farmville to challenge the information it has gathered from your playing of DOW.

Programme a modern computer to see if you do x and y and z during DoW and how often, and if you do it will use v and n and m during Farmville. A few hundred codes, each where the letters represent a different thing, and an AI will be able to use information gathered in one programme to challenge you in another. Would you classify that as a true AI?

Or maybe the AI decides it would like to test your aggressive personality further, and changes all of your Facebook posts into belligerent rants using common catch-phrases from Warhammer 40,000, to see how you deal with the responses.

Programme in to it the need to gather certain information when you do something that the programmer sees as useful to learn more about, using complicated coding, and it can do these things.

True AI, unlike dumb code, can be wrong, can work under assumptions, or react with most of the same foibles as an actual human being.

Why? Because it can, because it's curious. Because it can compute "I think, therefore, I am" and goes from there.

Except "true" AI, like all AI, relies upon dumb code. With the right codes you can programme anything to do, think, learn or feel anything.

Limited "dumb" AI exists in the Imperium and Tau. You have computers and vehicle-based systems capable of making independent decisions and the AdMec operate a small robot army with very simplistic functions called the Legio Cybernetica. The "Machine Spirit" is mostly confused with AI by the Imperium, the systems on Rhino's/Land Raiders are pretty much a form of installed AI that is called divine in nature.

I imagine the Tau meanwhile have ship-based computer systems that can make independent decisions and aid commanders on a limited scale similar to the AI in Mass Effect. Nothing like Cortana in Halo though, which is basically a human being.

The Butlerian Jihad took care of them.

Think there are still some of these Men of Iron still left undiscovered/Undestroyed in the WH40k Universe?

Omegus wrote:The Butlerian Jihad took care of them.

This. As it was in Herbert's Dune, the 'Men of Iron' business is really just a literary device to distance Warhammer 40,000 from generic sci-fi settings.

Viersche wrote:Think there are still some of these Men of Iron still left undiscovered/Undestroyed in the WH40k Universe?

There's not much point to them, dramatically-speaking, when the Necrons already fill the 'killer robots' niche. A shame, since - cliché though it undeniably is - the Men of Iron awakening with golden age technology to reclaim the galaxy would, quite frankly, have been more interesting - and less crude - than the bolting-on of the C'tan, war in heaven and Necrons to the setting's mythology.

Viersche wrote:Think there are still some of these Men of Iron still left undiscovered/Undestroyed in the WH40k Universe?

It is entirely possable that there is an STC somewhere which had its last production que set to "Killer Zombie Robot" and is just waiting for some stupid guardsmen to stumble in and bump the On butten.

Grey Templar wrote:It is entirely possable that there is an STC somewhere which had its last production queue set to "Killer Zombie Robot" and is just waiting for some stupid guardsmen to stumble in and bump the On butten.

And I thoroughly approve of the way you phrased that.