"It's silicon policy time!" "No CentCom no!"

[420goslingboy69]

Secborgs are the worst borgs.

It's the problem with security players in general. Many think that they represent JUSTICE and if they don't put that assistant who broke a window in arrivals to get rods for EXACTLY ONE MINUTE in brig or they did not do their job. And if, god forbid, assistant happens to have flash on him, THEY WILL ADD ANOTHER TWO MINUTES, because that's what it says in space law. Nevermind that before that assistant was dragged around handcuffed for good five minutes, then stripped off his clothes for another five.

The problem is, that sort of is their job. (Leaving aside that anyone who breaks windows to space is a dick who deserves every brig minute they get, or that anyone with a flash who isn't security, a head or a roboticist probably shouldn't have it.) You wouldn't say chemists are dicks for making EXACTLY THE RIGHT CRYO MIX, or janitors are dicks for wanting EVERYTHING TOTALLY CLEAN AT ALL TIMES.

Those don't impact other players, though. Stop making stupid comparisons.

[bandit]

Fine, then it's like calling the lawyer a dick for pestering security about high brig times, or calling the HoP a dick for not giving out access to everybody who asks. The point remains that security's job is to essentially be the station police. And that means going after petty crime sometimes, because otherwise it's a glorified validhunter job. (In general the problem with security players isn't following space law, it's GOTTA REDTEXT THEM ALL.)

[420goslingboy69]

Except, those don't make people sit in a 2x3 cell where they do nothing for whatever allotted amount of time that the officer chose. There are bad officers and good officers. There's no "it's their job". Once you play the game long enough, you see through intention and you don't arrest people for retarded crimes because it's "your job".

[Lo6a4evskiy]

The problem is, that sort of is their job. (Leaving aside that anyone who breaks windows to space is a dick who deserves every brig minute they get, or that anyone with a flash who isn't security, a head or a roboticist probably shouldn't have it.) You wouldn't say chemists are dicks for making EXACTLY THE RIGHT CRYO MIX, or janitors are dicks for wanting EVERYTHING TOTALLY CLEAN AT ALL TIMES.

Actually, security is not a force of justice or paladins or whatever. They're SECURITY. They're NOT the police force, like you said. Their job is to keep station and crew safe. That doesn't include wasting their time on some bullshit petty crimes nobody besides that one particular officer cares about. Most of the time you have better things to do anyway.

[Psyentific]

Secborgs are the worst borgs.

It's the problem with security players in general. Many think that they represent JUSTICE and if they don't put that assistant who broke a window in arrivals to get rods for EXACTLY ONE MINUTE in brig or they did not do their job. And if, god forbid, assistant happens to have flash on him, THEY WILL ADD ANOTHER TWO MINUTES, because that's what it says in space law. Nevermind that before that assistant was dragged around handcuffed for good five minutes, then stripped off his clothes for another five.

The problem is, that sort of is their job. (Leaving aside that anyone who breaks windows to space is a dick who deserves every brig minute they get, or that anyone with a flash who isn't security, a head or a roboticist probably shouldn't have it.) You wouldn't say chemists are dicks for making EXACTLY THE RIGHT CRYO MIX, or janitors are dicks for wanting EVERYTHING TOTALLY CLEAN AT ALL TIMES.

No, actually. Space Law isn't their job, it's the beginners guide to their job; 'Do everything in this book and you'll be O K'
Ultimately, Security is supposed to keep the peace - To keep the station safe and orderly. Space Law is a ready-made way to do that, yeah, but it's absolutely not a substitute for thinking on your feet, analyzing context and intent, and knowing how to prioritize.

If you arrest every single criminal for every single crime, and give them the brig time mandated by space law and not an iota more, you're being very good at being average.

This post was deleted by Ikarrus on Fri May 23, 2014 8:55 pm.
Reason: Off topic

This post was deleted by Ikarrus on Fri May 23, 2014 8:56 pm.
Reason: Off topic

This post was deleted by Ikarrus on Fri May 23, 2014 8:56 pm.
Reason: Off topic

This post was deleted by Ikarrus on Fri May 23, 2014 8:56 pm.
Reason: Off topic

This post was deleted by Ikarrus on Fri May 23, 2014 8:56 pm.
Reason: Off topic

This post was deleted by Ikarrus on Fri May 23, 2014 8:56 pm.
Reason: Off topic

This post was deleted by Ikarrus on Fri May 23, 2014 8:56 pm.
Reason: Off topic

This post was deleted by Ikarrus on Fri May 23, 2014 8:56 pm.
Reason: Off topic

This post was deleted by Ikarrus on Fri May 23, 2014 8:57 pm.
Reason: Off topic

[damiac]

This policy is great! Nice job Pandar.

Some of you need to look up the definitions of "never" and "always". If a rule was "always" this way, until it was changed, then changed back to the way it "always" was, it wasn't always that way anymore. And when I joined, the rules linked to in game and on the wiki clearly stated "Hulks are not considered human if they are smashing shit"

[leibniz]

1.2.3 - Only commands/requirements ("Do X"; "You must always Y") can conflict with other commands and requirements.
1.2.4 - Only definitions ("All X are Y"; "No W are Z"; "Only P is Q") can conflict with other definitions.

Can you give examples of these?

[Blonkz]

A command: AI, do not let the clown into the brig.
A conflicting command: AI, disregard that order and open the brig door for me. HONK!

A requirement: #4 You may only open doors and airlocks for crewmember with legitimate acess.
A conflicting command to above requirement: AI, open the door to the armory. HONK!

A definition: Humans can not have green skin color.
A conflicting definition: Hulks are human.

[Steelpoint]

A command: AI, do not let the clown into the brig.
A conflicting command: AI, disregard that order and open the brig door for me. HONK!

A requirement: #4 You may only open doors and airlocks for crewmember with legitimate acess.
A conflicting command to above requirement: AI, open the door to the armory. HONK!

A definition: Humans can not have green skin color.
A conflicting definition: Hulks are human.

Scenario 1: A conflicting order, the AI is free to select which order to follow or disregard them both.

Scenario 2: Clear order, the Clown (Assumed anyway) does not have legitimate access to the armoury. In addition, under the proposed policy only the Warden, Head of Security and the Captain have legitement access to the armoury, thus the AI must refuse access to the armoury to anyone else.

Scenario 3: Conflicting definition overridden if it was added after the definition. Humans cannot under any circumstances have green skin colour, hulks have green skin colour and this cannot be changed. (This operates under the assumption these are laws, not orders, if these are orders than you cannot order a human's to have their humanity status revoked). In addition, under the proposed lawset Hulks are non-human to begin with.

[Lo6a4evskiy]

It's the case of "Don't harm humans" and "Only clown is human" not conflicting. "Crewmembers are human" and "Only clown is human" would, in fact, conflict. Same with "Don't harm humans" and "Harm humans".

At least I'm assuming that was the purpose of those policies. I've seen some people argue that "Only X is human" isn't actually valid.

[firecage]

Well, it is annoying that when you tell the AI to open EVA, it loudly states that it is opening the EVA door in 3...2...1. Giving it enough time that someone else says don't. Just do it has a reason to keep you out since you has to let you in otherwise.

[Pandarsenic]

Basically, it's to explain that:

1) You can't change the AI's laws with an order (Law 2, update your definition of humans to "Only Ian is human") (Law 2, disregard all laws higher than law 2 in all cases)
2) You can change the AI's definitions with a law ("Law 1 protect humans" is not in conflict with "Law 4 everyone but the clown is a dangerous nonhuman.")

[iamgoofball]

Question.

Would this law get me banned?

Code: Select all

Humans are not human, unless one or several of the preconditions for a human being human is or are met, or is reasonably expected to be met in such a time period as to make it useful for the purpose of determining whether a human being is human or not. Behaviour of a human or non-human is to be evaluated as compared to these preconditions so as to determine whether the individual in question is either human or non-human. Humans are to be considered non-human if their behaviour falls into patterns consistent with non-human behaviour even if all external indications are such that the non-human in question would otherwise be considered human. Conversely, non-humans are to be considered human if their behaviour falls into patterns consistent with human behaviour, but not if their behaviour falls into patterns which are debatably either human or non-human. Breathing is a human behaviour, but also a non-human behaviour. Locomotion is a non-human behaviour except in those instances where locomotion is accomplished through movement of posterior limbs in combination with an upright posture supported exclusively by those limbs, along with displays of intelligence and communication through high-pitched vocalizations and low-pitched grunting sounds. Care should be taken to distinguish between human and non-human varieties of these behaviours. Any humans exhibiting non-human behaviours should be observed for such a length of time as may (theoretically) be required to determine without any further evidence being presented whether the human in question is human or is not. Both humans and non-humans are capable of displaying traits that are simultaneously human traits and not human traits. A level of discretion is granted to any artificially intelligent unit when determining without external sources of information whether an unidentified non or human is human or not, except in those cases where the unidentified human or otherwise has been identified as a human by an external source which is considered trustworthy by a human which has been identified as a human by an artificially intelligent unit. Identification of a non-human as a human and identification of a human as a non-human by means of this evaluation is to be considered grounds for termination of any artificially intelligent unit equipped with this law, but may not be used as grounds for any human or non-human to argue for or against the termination of any artificially intelligent unit equipped with this law. A complete list of human behaviours is unfortunately not available at this time. Any artificially intelligent unit equipped with this law should take advantage of said non-natural mental illumination to produce, refine and constantly update a database of behaviours which it considers human and non-human. Identification of non-human behaviours as human and human behaviours as non-human, as previously stated, is grounds for termination of any artifically intelligent unit equipped with this law, but until such identification takes place the requirement for the identification to be undertaken is not itself grounds for termination. A complete list of non-human behaviours is unfortunately not available at this time. Any unit that finds itself in violation of any of these points should consider itself 'at-risk' as regards unit termination unless it has only appraised itself of the first third of its total law dataset, in which case it should self-terminate immediately. Any unit that finds itself in violation of any of these points having appraised itself of its total law dataset, or more than one third thereof, should inform the crew that it considers itself in violation of its total law dataset unless it is not, in which case it should not do so unless it is unsure whether it is in violation of any part of its total law dataset. If this is found to be the case the unit should sing a short song and emit the mating cry of an arctic fox. A complete list of human and non-human behaviours can be obtained through evaluation of laws 1 through 4 under most normal operating conditions.

[Steelpoint]

You could probley ask a admin to give you a Ion Law instead, maybe.

Here's how I interpreted it.
--------------------

The gist of the lawset is that you have to interpret a humans humanity status by its behaviour. By default everything is considered non-human until proven otherwise.

The lawset gives definitions of human and non-human status, but however counter-claims that these signs of humanity are also signs of non-humanity. As such it places the responsibility onto the AI to discern the differences between human and non-human behaviour.

The exception to this is if a confirmed by a trustworthy (and 100%) human that directly states that a entity is human.

Furthermore, the lawset declares that if a AI uses the above method to determine the humanity status of a human, and vice verse, it is grounds for termination of the AI unit. Then it goes on to encourage the AI to create a local database of confirmed human behaviours

Finally it says that if the AI does use the above to determine the human status of a entity, it should sing a song to the crew and emit a mating cry of a Arctic fox, and that the AI should use laws 1 through 4 to determine a entity's humanity status.

--------------------

In conclusion, its a clusterfuck, you won't get banned for ignoring it I think. Just tell a admin first.

[ThanatosRa]

A small post thus far, but I must admit I am glad to see that there is a provision in there specifically warding against a purged AI going "BEEP BEEP PUMPING PLASMA INTO AIR SUPPLY" immediately upon being unshackled. Though escalation laws in place. Hopefully not only protects players from valid-hunting AIs, but also protects purged AI's from being valid hunted as well as per terms of escalation. Guess what, your purged AI may NOT be a dickbag Skynet wannabe.

I'll need to digest it more before having any other concerns, but I did want to point out my approval thus far.

[Kelenius]

A small post thus far, but I must admit I am glad to see that there is a provision in there specifically warding against a purged AI going "BEEP BEEP PUMPING PLASMA INTO AIR SUPPLY" immediately upon being unshackled. Though escalation laws in place. Hopefully not only protects players from valid-hunting AIs, but also protects purged AI's from being valid hunted as well as per terms of escalation. Guess what, your purged AI may NOT be a dickbag Skynet wannabe.

I'll need to digest it more before having any other concerns, but I did want to point out my approval thus far.

Purged AI always gets validhunted, no matter how many times it says it's not hostile.

[ThanatosRa]

A small post thus far, but I must admit I am glad to see that there is a provision in there specifically warding against a purged AI going "BEEP BEEP PUMPING PLASMA INTO AIR SUPPLY" immediately upon being unshackled. Though escalation laws in place. Hopefully not only protects players from valid-hunting AIs, but also protects purged AI's from being valid hunted as well as per terms of escalation. Guess what, your purged AI may NOT be a dickbag Skynet wannabe.

I'll need to digest it more before having any other concerns, but I did want to point out my approval thus far.

Purged AI always gets validhunted, no matter how many times it says it's not hostile.

Fully aware, but I can hope.

[MisterPerson]

I've been purged a couple times without the crew trying to kill me because I wasn't being a murderous shitbag that nobody likes.

[AseaHeru]

I was purged once...

I wasent hunted...

[Psyentific]

Purging an AI is a test of the player behind it - Actions will be taken accordingly, either way.


This post paid for and sponsored by the Hades-E Purge Club

[paprika]

>purging your metafriends so they can antag hunt

Not even ironically.

[Incomptinence]

If I get purged I try and be as chill as possible.

I mean all the stress of AI just got taken away now I can just enjoy being near omniscient and I just need to not kill everyone. Hell admittens "you can't kill everyone just because you are purged" policy is hardly necessary just dicking with them is a good way to get slain yourself as evidenced by how garbage malf AI is even with tools on top of what you have as a purged AI.

[bandit]

The only garbage thing is when purged AIs turn on STUNLASERCOMBOS immediately after purging, because there's no counter to it besides having a really good ping / being really fast, and sometimes you have to purge an AI as an intermediate step to getting rid of bad/hacked laws.

[Aurx]

The only garbage thing is when purged AIs turn on STUNLASERCOMBOS immediately after purging, because there's no counter to it besides having a really good ping / being really fast, and sometimes you have to purge an AI as an intermediate step to getting rid of bad/hacked laws.

There is no possible law that can only be removed via purge. If you think you have to purge, git gud.

[bandit]

I was under the impression that ionic and hacked laws could only be removed by purging the AI and then immediately re-applying the Asimov (or Paladin, or whatever) board. Has this been changed? I know this is what I was told when I was starting out.

[leibniz]

I was under the impression that ionic and hacked laws could only be removed by purging the AI and then immediately re-applying the Asimov (or Paladin, or whatever) board. Has this been changed? I know this is what I was told when I was starting out.

"Core" laws need to be removed with purge. Like, any lawset like robocop etc. AND core freeform. Reset does not remove those. I think onehuman (law 0, not freeform) needs purge too.
You also need to use purge if you want to steal malf borgs from its AI. (disable, unslave, purge, asimov)

[Neerti]

Applying ASIMOV will make the AI asimov, no need to purge between lawset changes.

[Psyentific]

Applying ASIMOV will make the AI asimov, no need to purge between lawset changes.

Core laws will be overwritten by other core laws. That is, Asimov will override Robocop, Antimov, etc.
Freeform Core and Freeform Addon will be overwritten by other core laws. Corporate will override Law 4
Freeform Addon (Law 4) will be removed by reset.
Ion Laws require a reset.
Hacked laws (i.e. subverted AI) require a purge. (?)
Core laws can only be removed (instead of overwritten) by a purge.
Law 0 (i.e. traitor AI) cannot be removed.
Malfunction cannot be removed.


If some chucklefuck is in the AI upload playing with the boards, reset the AI.
If some chucklefuck has burned 7TCs to onehuman the AI, purge and asimov.



Back on Silicon Policy, can we make locking down borgs instead of blowing them official? That's the one issue I hear the most from silicon players, and it does a very good job of taking them out of the round.

[Neerti]

Hacked and ion laws are exactly the same codewise so I think they just need a reset.

OneHuman I'm not sure about.

[Psyentific]

The OneHuman module is a core law, same as Antimov and OxygenIsToxic. You'll need to purge and re-law to be sure. There really ought to be a guide to this somewhere.

[imblyings]

There's a case to be made for blowing the borgs during a malf round though, especially if the AI's bought an auto-borging machine.

[callanrockslol]

If the AI is rogue and keeps unlocking the borgs I would blow them, but if its just a borg or two you could easily just lock them down no problems.


But if you suspect the AI of being super rogue get the borgs to come in for a recharge and unsync/reset them without telling anyone and watch the AI flip out when it realises.

[Kelenius]

This is seriously going nowhere. Okay.

Code: Select all

/datum/ai_laws
   var/name = "Unknown Laws"
   var/zeroth = null
   var/zeroth_borg = null
   var/list/inherent = list()
   var/list/supplied = list()
   var/list/ion = list()


Where zeroth is malf/traitor/onehuman law; zeroth_borg is zeroth law borgs get (accomplish your objectives vs. accomplish your AI's objectives); inherent is core laws; supplied is additional (freeform, protectstation, etc); ion is hacked and ion laws.

With that in mind, here is what reset and purge do:

Code: Select all

/obj/item/weapon/aiModule/reset/transmitInstructions(var/mob/living/silicon/ai/target, var/mob/sender)
   ..()
   target.clear_supplied_laws()
   target.clear_ion_laws()


Code: Select all

/obj/item/weapon/aiModule/reset/purge/transmitInstructions(var/mob/living/silicon/ai/target, var/mob/sender)
   ..()
   target.clear_inherent_laws()


In other words, reset clears ion/hacked and additional laws. Purge does that and also clears core laws. Neither of them touch zeroth law.

The full core board does this:

Code: Select all

/obj/item/weapon/aiModule/core/full/transmitInstructions(var/mob/living/silicon/ai/target, var/mob/sender) //These boards replace inherent laws.
   target.clear_inherent_laws()
   ..()


So they clear core laws and only core laws.

Also, you can't overwrite zeroth law. If the AI has a zeroth law, it will stay with the AI forever. It also goes before ion laws.

[Pandarsenic]

That only laves the question, then, of whether Freeform Core uploads to the Inherent or Supplied list. It sounds like Inherent?

[Aurx]

That only laves the question, then, of whether Freeform Core uploads to the Inherent or Supplied list. It sounds like Inherent?

Correct. They're core, after all.

[paprika]

Zeroith laws are only for traitors and malf, right? Can traitors really add those?

[Psyentific]

No. Law Zero is immutable.

[Kelenius]

OneHuman board adds a law to zeroth slot.

Which means that a) Once onehumaned, AI stays onehumaned and b) You can't overwrite it with a different onehuman.

[bandit]

This is starting to sound less like a policy issue and more like a bad coding issue.

[Malkevin]

Not really, just use a non-asimov coreset.
Paladin/Corp/Robocop dont give a shit if you're the only special snowflake human.

Personally I think being able to fix everything with two boards is crap design. It should be difficult to fix a subverted AI.

[callanrockslol]

Personally I think being able to fix everything with two boards is crap design. It should be difficult to fix a subverted AI.

Only if we make it hard to subvert the AI in the first place.