|
Infinite Monkeys posted:Nakar what exactly is your suggestion? The obvious answer is to add a suicide module in the upload but like I said people would just space it. I don't want to add a traitor item because you shouldn't have to spend telecrystals to kill the AI in addition to getting head access. Anyone in the first group is a dick who should be punished regardless of whether their intention was to prevent the AI from being destroyed or just to prevent stupid laws from being uploaded, so that problem is dealt with by admin intervention once we hear from somebody that Joe Soandso is spacing modules. Because you're not supposed to be doing that. Anyone in the second group is either uploading a "funny" law and trying to preserve it and should be warned away from spacing poo poo for the sake of their terrible gimmick, or is intending to subvert the AI and wants to avoid it being reset or disrupted in which case (as long as they're an antagonist) they should be allowed to do so as that's just good planning. If you intend to subvert the AI and kill it later, you'd use the law modules, space them, and hold on to the termination module so you can go back later and upload it to kill the AI. If somebody manages to stop you, the only thing they can do is kill the AI because you cleverly ditched the law modules, so now they must decide between an AI that is actively assisting you in your crime or destroying it, because they can't just override your law with another one. It seems entirely beneficial to me.
|
#
?
Dec 27, 2013 20:35
|
|
|
|
| # ? Jun 14, 2024 13:30 |
|
|
If the idea is to add in a more direct kill the AI method, there seems to me to be a simple solution: In the computer core, on the robotics terminal, add in a kill switch option for the AI. It seems to me that kill switches as they're implemented RIGHT NOW would satisfy every requirement we're talking about, with extra entertaining stuff like AI monitors blowing up optional after we're more sure that it's a good solution. The advantage to having it slightly delayed is that it's substantially easier to enable the killswitch routine without alerting the AI, since it can't hear you over in the Core.
|
|
#
?
Dec 27, 2013 20:36
|
|
|
The thing is that the ways a paranoid AI and a paranoid anyone else try to keep safe are completely different. Which would be fine except most of the ways that a paranoid AI can try to defend itself are poo poo. So if the AI doesn't want to be poo poo it's more likely to rules lawyer because choosing not to take steps to defend itself (in the interest of not being poo poo) made it so someone could kill them, and in a really boring way no less. That's just a different kind of poo poo though. The way things work now is frustrating for the AI, and the traitor, and that's why this argument has already broken out a half dozen times or something in this thread. Disallowing AI suicide laws in some way (throw a destruct console in the computer room across from the AI core or something), making AI death more interesting, giving the AI the chance to say something but not neccessarily know who did it, or allowing the AI to be revived would probably help to stop the arms race of shittery that people are complaining about. I also like the idea of stealing the AI's brain and slapping it into some syndicate machine instead of just taking it's brain and leaving it at that.
|
|
#
?
Dec 27, 2013 20:37
|
|
|
amuayse posted:On the subject of Vampires, how do I counter them if they both have their screech and flash?
|
|
#
?
Dec 27, 2013 20:40
|
|
|
If we're worried that people will space AI modules why not make a module printer?
|
|
#
?
Dec 27, 2013 20:41
|
|
|
Admiral Funk posted:I also like the idea of stealing the AI's brain and slapping it into some syndicate machine instead of just taking it's brain and leaving it at that. Could maybe give Traitors an optional item to buy which prevents a reset, or something along those lines? Maybe just prevents a set number of resets/hides laws so that they cannot be changed immediately. Giving them a safer way to get around AIs without needing to take them out. Also, I rather like the idea of putting the AI's brain in some sort of Syndi-bot.
|
|
#
?
Dec 27, 2013 20:43
|
|
|
Dr Cheeto posted:If we're worried that people will space AI modules why not make a module printer? If you could print off a reset module it would make it too risky to ever subvert the AI because it's impossible to know how many reset modules exist. If there's a limited supply (1-2) then if I as a traitor own both of them I know the AI can't be reset.
|
|
#
?
Dec 27, 2013 20:43
|
|
|
Ok, this might sound silly, but hear me out: We add a Big Red Button to the bridge. The Big Red Button is ensconced within a glass case that is impact/emag/explosion-proof. There are two card swipes on either side, that will only accept Head IDs. You have to swipe two separate IDs on either one to open the glass case, which causes klaxons to blare and all the emergency lights to flash. Hit the button, and the AI is loaded into a mass driver and fired into the debris field.
|
|
#
?
Dec 27, 2013 20:46
|
|
|
I haven't played in a while, but last I remembered, you could scan the modules. Is this not true anymore?
|
|
#
?
Dec 27, 2013 20:47
|
|
|
Zaldron posted:I haven't played in a while, but last I remembered, you could scan the modules. Is this not true anymore? Not necessarily so with a manufacturer being able to produce them.
|
|
#
?
Dec 27, 2013 20:48
|
|
|
What if there was a syndicate AI housing that could be dragged around and attached to wiring allowing it to travel through the stations wires and gently caress with things on a smaller scale than a regular AI? I feel like that could be neat.
|
|
#
?
Dec 27, 2013 20:49
|
|
|
Captain Bravo posted:Ok, this might sound silly, but hear me out:  .
|
|
#
?
Dec 27, 2013 20:52
|
|
|
Better yet, hit the button and the AI is flung into Belt Hell and then out the mass driver into the asteroid field. You'd need a special bypass so it doesn't get stuck in the crusher area though. And attaching a barcode causes it to override and instead shoot the AI to the department indicated, allowing for AI jacks.
|
|
#
?
Dec 27, 2013 20:57
|
|
|
The only issue with alternative methods of AI destruction is that I can't think of ways to then fence off freeform modules. You would need the carrot and the stick for something like this, a way to get people to stop adding in suicide laws into the freeform. Sure, you could just make a new rule that AI's don't have to obey suicide laws, but that seems kind of lame. You could mandate that suicide laws are against the rules, and maybe bannable, but that seems kind of finicky, and difficult to enforce? I don't know, I'm out of ideas for the stick portion of the idea.
|
|
#
?
Dec 27, 2013 21:02
|
|
|
If destroying the AI through other means is made funny and awesome enough, nobody will want to bother with freeform suicide laws. Why do that when you can fire it out of a cannon into an asteroid? EDIT: Obviously I see your point with respect to your proposal requiring head IDs, as not everyone will have two. But you could lessen or even remove that requirement and just have some other means of accessing the button. Or no means, just put a big loving button there.
|
|
#
?
Dec 27, 2013 21:04
|
|
|
You should have to fight in virtual space a la Lawnmower Man or whatever
|
|
#
?
Dec 27, 2013 21:08
|
|
|
Poland Spring posted:You should have to fight in virtual space a la Lawnmower Man or whatever
|
|
#
?
Dec 27, 2013 21:13
|
|
|
Nakar posted:Or no means, just put a big loving button there. I would love this, just for the fact that the week after implementation it would be utterly impossible to live longer than three minutes per round as an AI. 
|
|
#
?
Dec 27, 2013 21:15
|
|
|
Actual real fix: The access required to open the Big loving Button panel is simply any ID that also has access to AI Upload. If you could get into the upload to upload a "kill urself" law, you can swipe open the button on the Bridge that kills the AI. It's essentially the same thing, except you need not walk into the AI's chamber and risk it tattling and it's in the Bridge, which is an area with way more escape routes than the AI Upload.
|
|
#
?
Dec 27, 2013 21:18
|
|
|
Coolguye posted:I realize you're technically agreeing with me here but this sort of thick-headed flippancy is why a lot of people don't like you. Try to not do that. I explained my reasoning to why I don't agree with him and he retorted with a "No, you're wrong  .", so I responded in kind.Anyway, I don't understand the sudden uproar for changing something that works and has worked for years. In order to get into the AI upload you need either a head ID or a willing AI. If the guy is not a head, you as the AI can prevent them from uploading a law very easily by either stunning them or turning off the power to the console. Why make the whole thing more confusing and convoluted by having an extra module, or having to use the mainframe or other equally far-fetched ideas when what we have works? When playing as the AI, just as much as when playing human, you run the risk of either getting murdered horribly or getting subverted. It's just how the game works and people are reaching pretty far for legitimate complaints about how uploading laws work. I'm really sick of people who play synthetic roles and think that they don't have to follow their laws if it inconveniences them at all. My biggest annoyance and I'll repeat it every chance I get. Nakar posted:Actual real fix: The access required to open the Big loving Button panel is simply any ID that also has access to AI Upload. If you could get into the upload to upload a "kill urself" law, you can swipe open the button on the Bridge that kills the AI. It's essentially the same thing, except you need not walk into the AI's chamber and risk it tattling and it's in the Bridge, which is an area with way more escape routes than the AI Upload. How the gently caress is that any better for the AI than uploading a law to it? That would just make it way easier to kill the AI.
|
|
#
?
Dec 27, 2013 21:26
|
|
|
Even better: The Big Red Button does not kill the AI. Hitting the button causes the AI to slowly lose camera functions. Something like:quote:Warning: System Lockout Implemented. After that, it loses radio connection, (So it's had about 15 seconds to see who's on the bridge and scream bloody murder) then it disconnects from the power grid and switches to its reserve battery, all the turrets in the room power down, and finally the mounts unbolt and its blast doors open. At that point, the AI is only able to connect to its two shells, and can no longer interact with the station. through cameras. If someone wants to break into the AI room, steal the AI, drag it over to a terminal and bolt it down, it can have limited camera and door functionality again. If someone wants to drag it out of the room and load it into a mass driver to fire it on spess adventures, that's fine too. This fixes the main reason for killing the AI, which is that some jackass protected by his walls and turrets can spy on you all over the place and alert the crew to your hijinks, or bolt you inside. It fixes the "Bullshit suicide laws!" complaint because you're not dead, you're just no longer the Station God. And, it still allows for easy access to the AI so you can fire him at spess invaders like a big, boxy torpedo.
|
|
#
?
Dec 27, 2013 21:27
|
|
|
Klayboxx posted:Anyway, I don't understand the sudden uproar for changing something that works and has worked for years. Remove from the equation any possibility of the AI player interfering in its own death in a way that violates the rules. Believe me, this will make most AIs less cagey and less angry about being killed, and the ones it doesn't will just have to adminhelp it after being dead instead of while stubbornly refusing to suicide, meaning that for you the traitor the AI is dead and it's not your problem anymore. As it should be.
|
|
#
?
Dec 27, 2013 21:47
|
|
|
Good god, you're all children in here, aren't you? The only solution I've seen so far in this thread that I've liked is the format module idea posted a couple pages back: Captain Bravo posted:To neatly sidestep all the "Grr suicide laws!" chat, what if we add in a new traitor item? Costs 1 telecrystal, a Format Disc. Insert it into the AI Upload, and the AI has a countdown before it's erased. A window pops up on the AI's screen "Warning: Format impending. Emergency shell download?" Clicking yes downloads the AI into one of it's shells, and instantly shuts down the main AI. Boom, you've given traitors an easy, simple way of removing the AI, given the AI a way to survive someone who pops into the upload hell-bent on their destruction, and also prevented the AI from completely bugging out whenever he wants. Neat and simple. Change it so it's not a traitor only item, it's on the station, and limited to 1 or 2 in well guarded locations. Go a bit further: Once the format happens, the AI has its speech disabled, and can only communicate through a few emotes, like nodding yes or no, and the like. The AI core can be reset by using the reset module, which slowly reprograms the core back to factory defaults over, say, 1 minute. Once the core is reset, the AI can jump back into it and resume play, or a new brain can be inserted in to start with a fresh AI player. If the reset doesn't happen, Once the AI is either borged, humaned/monkified, or reinstalled into the AI core, it regains its abilities to communicate. Furthermore, it should be a miscreant objective to forcibly get the AI into a borg or "living shell" and take it with you on the escape shuttle. With this, if an antag wants to take out the AI, they can format them and either hunt down and beat up the shell, or just grab and space the reset modules. If the AI is clever and fast, they have a chance of getting back into their core, just like how humans have a chance of being cloned. The silence effect on their shells prevents them from running to the nearest intercom and screaming out their aggressors name.
|
|
#
?
Dec 27, 2013 22:52
|
|
|
(Disclaimer: I don't play SS13, I just watch it and read this thread as a spectator sport until the standalone comes. I hope I'm still allowed an opinion  )Klayboxx posted:Anyway, I don't understand the sudden uproar for changing something that works and has worked for years. Suicide-succumb is functional in that it is a mechanically valid way to kill the AI. However, it seems like a lot of people are unhappy with it and that it could be improved One of the problems I'm seeing here is that the AI has no way to tell the difference between someone subverting them for fun times and killing them. That's really poo poo. It makes the entire problem of people suicide lawing lovely AIs self-perpetuating. It means AIs who don't want to die are going to defend their core vigorously, pissing off people who just want to upload gimmicks or subvert them, which means those people are instead going to upload suicide laws, which makes the AI pissy and super defensive, ET CETERA ET CETERA and it's a cycle of lovely play. Klayboxx posted:How the gently caress is that any better for the AI than uploading a law to it? That would just make it way easier to kill the AI. I had an idea for implementation for the big red KILL AI button. Pardon my (loving awful) MSPaint skills, and if I get any of the exact details wrong. Picture assumes upload currently has two blast doors and 6 turrets protecting it, walls would all be whatever the walls of the AI core currently has - R-walls, if I'm not mistaken.  (Maybe even only protect the upload with two turrets and put the four in with the killswitch) Essentially, same amount of security to kill the AI, but it's actually easier to subvert or gimmick law the AI. AI players seem to love getting good gimmicks or getting subverted and sneakily helping traitors - encourage that poo poo. Also, it means if someone starts breaking into the upload, you can PDA message them saying "hey if you're not going to go after my killswitch, I won't turret you in the face and scream that you're a traitor. Touch that god damned killswitch door and I'm screaming bloody murder though." It gives AIs a way to defend themselves without getting called funhating shitlords because they're blocking gimmicks and subversions. Obviously, suicide laws would need to be declared invalid on a meta-game server rules level for this to work - the upload has to stop being the primary way to kill the AI for them to stop defending it like it's the primary way to kill them. Also, if someone's going to be a bitch about getting subverted or gimmicked, they're probably the kind of person to scream bloody murder about someone going near their upload in general - once they start screaming, you have perfectly valid grounds to tell them to go gently caress themselves and killswitch them for being a funhating fuckface. The best part? If you're a traitor, you then turn around to the crew and say 'guys I was trying to upload a gimmick and he was being a shitlord. Do you really want to have a lovely AI this round? I killed him because he was being a piece of poo poo  ' and talk your way out of getting brigged for popping the AI.Edit: The heavily guarded reset module seems like a reasonable solution too. If it's easier to subvert the AI than kill them, people are gonna subvert more and AIs are gonna chill out a little Magres fucked around with this message at 22:55 on Dec 27, 2013 |
|
#
?
Dec 27, 2013 22:53
|
|
|
Klayboxx posted:I explained my reasoning to why I don't agree with him and he retorted with a "No, you're wrong Nakar posted:Actual real fix: The access required to open the Big loving Button panel is simply any ID that also has access to AI Upload. If you could get into the upload to upload a "kill urself" law, you can swipe open the button on the Bridge that kills the AI. It's essentially the same thing, except you need not walk into the AI's chamber and risk it tattling and it's in the Bridge, which is an area with way more escape routes than the AI Upload. It also makes a corrupted AI a lot less threatening, and a corrupt AI is already about a 3rd or 4th priority on an average round. I'd go so far as to argue that you upload a corrupting law to the AI really to corrupt the borgs, because borgs are robust as gently caress. I don't really understand what the problem was with the killswitching idea from the Robotics console that we decided to appropriate the exact same idea and move it to the bridge with less kill time? Killswitching from the Robotics console is more stealthy than uploading a law (since the AI can't hear the noises in its house), has identical access and security requirements to a suicide law (head access, turret in the floor), and gives people time to react so you're not just magically instagibbed the way you are with a suicide law or The Button. Also if you REALLY want privacy in there, I think there's a grand total of 1 camera that can see the core, so snip that and the AI can't even scream your identity as you skulk away. The bridge has a good handful of cameras you'd need to disable to give yourself full privacy there.
|
|
#
?
Dec 27, 2013 23:16
|
|
|
That's why I suggested the button merely isolating the AI, instead of killing it. The way I see it, you have two dynamics that you need to balance here. AI can't be revived, so death should be a last resort <> Sometimes the smarmy gently caress needs to shut the gently caress up The AI needs to be fairly decently protected <> Dealing with the AI should not be impossible A red "Oh poo poo!" button on the bridge that's nonlethal would strike a balance with all four of these. The AI is still decently protected against deadly attackers. What's more, if someone hits the button to try and make him easier to hit, it becomes ridiculously blatent, and the entire crew will probably rush in there anyway just to gently caress with the AI. However, this also works excellently as a way of neutering the AI, if only for a while, and in the event of a rogue AI with the reset boards spaced, the command staff has a decent shot at containing the situation. What do you guys think? Anything obvious I'm missing here?
|
|
#
?
Dec 28, 2013 00:18
|
|
|
You'd probably fix a lot of issues by just making it possible to reboot the AI with a little bit of work (say, getting an engineer to jump-start it with a rewire and some decent power from the engine). They'd stay dead unless someone actually came to the computer core to fix it, and if a traitor really wants the AI dead they'd have to actually put some effort into it. Adding an automatic announcement might work too. It'd be harder to sneak in and tell the AI to suicide itself when it's going to make  "WARNING: THE AI HAS SHUT ITSELF DOWN" blare over the radio when it happens.
|
|
#
?
Dec 28, 2013 00:53
|
|
|
Put a nuke in the AI, make it explode when the AI shuts down. You can still upload suicide laws but you'll have to pretty drat careful or else it will be a little too eager to comply and erase you and most of the station in radioactive fire.
|
|
#
?
Dec 28, 2013 00:54
|
|
|
Here's the thing, though. We could "fix" some of the issues people are having with a small change that might work out better or worse... OR we could add a giant button of fun that causes fun things to happen when you hit it, and almost certainly will result in hilarity. (for better or worse...)
|
|
#
?
Dec 28, 2013 00:58
|
|
|
Nakar posted:It doesn't work. That's the issue. It's creating lovely conflict that doesn't need to exist and needlessly complicating what amounts to "if you can reach the computer, the AI dies" because the "feature" is something not originally contemplated. But if we're okay with the principle (that the AI doesn't need to be physically destroyed), then just let the AI die if you reach Computer X, wherever it's decided that computer should be and whatever module, button, or game of checkers is needed to activate that computer and kill the AI. There is no conflict it's all in your brain. How it works is totally fine, making it more convoluted doesn't make it any better. I don't know what you mean when you say that it wasn't originally contemplated because an admin has come in and stated that it's totally OK to upload a suicide law if you're a traitor. Making the AI completely blind and unable to do anything with a button is literally like locking it inside of a locker and welding it, something I've been yelled at for doing even as a traitor. How are any of these any better than just simply uploading a "kill urself" law? You guys are throwing out ideas that are adding more layers to something that isn't easy to do in the first place, in fact, some of them are making it easier to kill the AI. What makes you guys think that future AI's won't just start locking down the room the button is in like some already do with their upload? The idea of making a traitor only kill module is absolutely awful. This leaves every other antagonist at a large disadvantage. Having it not a traitor item makes it literally the same thing as the freeform. It would probably be kept in a room with a turret, with a single door leading into it. The AI would probably get sensors that tells it if someone goes into said room. Whoa, sound familiar? Well it should because that is literally the upload room. The unknown of "Wow is this dude picking up the freeform to murder me horribly or to make me a murderer?" goes really well with the theme of the game and I don't understand why you'd want to squelch that. Don't change game mechanics in order to serve the lowest common denominator, and in this case that would be lovely AI's who stun literally everyone who goes into the upload. Those are people who should banned from synthetic roles, not pandered to by changing something that works perfectly fine. You have to think of the AI as just another player and not some almighty god that should get every benefit ever. Neddy Seagoon posted:You'd probably fix a lot of issues by just making it possible to reboot the AI with a little bit of work (say, getting an engineer to jump-start it with a rewire and some decent power from the engine). They'd stay dead unless someone actually came to the computer core to fix it, and if a traitor really wants the AI dead they'd have to actually put some effort into it. Adding an automatic announcement might work too. It'd be harder to sneak in and tell the AI to suicide itself when it's going to make Yeah lets do this for people too but make it so we know when and where the person was killed! People are so defenceless against traitors they should have some kind of defence, you know???
|
|
#
?
Dec 28, 2013 01:03
|
|
|
Klayboxx posted:There is no conflict it's all in your brain. Dr Cheeto fucked around with this message at 01:10 on Dec 28, 2013 |
|
#
?
Dec 28, 2013 01:05
|
|
|
Captain Bravo posted:Here's the thing, though. We could "fix" some of the issues people are having with a small change that might work out better or worse... Make VR bleed over into reality  Edit: I think it's time for another AI challenge If you don't put AI as your favorite job for a week you're a scrub
|
|
#
?
Dec 28, 2013 01:06
|
|
|
Poland Spring posted:Make VR bleed over into reality "Your mind makes it real... for everyone"
|
|
#
?
Dec 28, 2013 01:08
|
|
|
Klayboxx posted:Yeah lets do this for people too but make it so we know when and where the person was killed! People are so defenceless against traitors they should have some kind of defence, you know??? The problem is the AI isn't a typical player. An average death for a crewman generally means that they can be cloned, and even then they still have the ability to avoid it if they go hide themselves in a locker or something. Hell, even if they get drained by a Changeling they can still get their brain stuck in a Cyborg. Death for the AI in any form means their round is over, and the AI being dead in a crisis really does hamstring the crew if they don't know it's gone yet.
|
|
#
?
Dec 28, 2013 01:13
|
|
|
Klayboxx, Nakar is an admin too, so I don't think playing the  an admin has weighed in on this issue card will work super easily.But you've got to understand, dude, obviously these dudes think it's an issue. Sure, it's probably all in their heads, and I don't honestly think anything needs to be changed, but if there's a simple solution which can possibly make the game more fun for more people, it's worth hashing out. Tradition is a poor reason not to consider alternatives, just because it's always been one way doesn't mean it can't be made better. You'll notice that I was careful to mention a lockout button would still allow the AI to deploy to it's shell. There's two bodies the AI can use to move around and interact with the crew, plus if it can convince someone to drag it out and wire it back in over a network terminal, it would regain limited functionality. That is a hell of a lot better than just dying, while accomplishing many of the same purposes, and allowing anyone who actually does have a hard-on for murdering the AI to continue working towards that goal. Instead of making the AI "easier" to kill, it makes killing the AI more complicated. It adds in alternative methods and possibilities. Adding in the button would, in no way, prevent someone from teleporting a lion on top of the AI, or RCD'ing in through the walls and uploading a suicide law, it just makes those less attractive propositions. Since some of the people that originally complained it was too easy to kill the AI have said they like this plan, I don't think it would be a huge problem? I agree with you on the freeform thing, but since we have multiple people posting in the thread that it's a problem, both from the perspective of the AI and from the perspective of the traitors, maybe you might wanna consider the possibility that some people actually do have legitimate problems with it? If a solution can be found that makes the game better for them, and doesn't negatively affect you, why the gently caress do you give a poo poo? Also, this "We don't bargain with terrorists" schtick is kind of crap, dude. Yes, if people can't understand how to play a job the way it currently is, they probably don't need to be playing that job in the way that it currently is. But if a portion of the playerbase doesn't like the way a job currently is, sometimes the answer isn't always to ban that portion and soldier on. Sometimes, maybe we could take a look at changing things if there's a possibility it could work out better for players. Just because someone is poo poo at a portion of the game in play doesn't mean their opinion is to be ignored. Amazing things have come out of figuring ways to make the game more fun for people that don't work well with certain portions of it. There is no such thing as a stupid question. Captain Bravo fucked around with this message at 01:19 on Dec 28, 2013 |
|
#
?
Dec 28, 2013 01:17
|
|
|
Klayboxx posted:There is no conflict it's all in your brain. Klayboxx posted:How it works is totally fine, making it more convoluted doesn't make it any better. I don't know what you mean when you say that it wasn't originally contemplated because an admin has come in and stated that it's totally OK to upload a suicide law if you're a traitor. I don't care if it's emergent. I don't care if it's not against the rules. It's idiotic and it should not be how AIs are killed. The purpose of suicide as a command is not to murder players. Period. I mean think about this: Murdering the AI shouldn't be dependent on the AI's player being a good sport about killing itself. That's not murder, that's a sternly-worded request that the AI die, and to no one's surprise people try to find any excuse to not do it. If you could just killswitch it, the AI can't debate it. It can't fight it. It can't whine. It can't out you and gently caress you over. It dies, you took care of it, end of story, let the admins deal with the AI whining that it was killed just like any other player who gets killed. Why force the admins to intervene in a battle of wills (and proofreading) when you could just have a Terminate AI module that shuts it down without worrying whether your AI is going to be a  bastard and try to weasel out of it while ratting you out?
|
|
#
?
Dec 28, 2013 01:23
|
|
|
PoptartsNinja posted:I can't say with any certain (because I don't mess with the engine at all, Engineers are allowed to have fun too!) but it may be because a hotwired engine starts zapping the AI pretty much immediately, so they've either got to shut off the APC to the AI chamber and go to internal battery (theoretically limiting the amount of time they can potentially have fun) or leave the APC on and pray they don't get instantly killed by electricity. Engine wasn't hotwired at all in this instance, he was just one of the AIs that try VERY hard to find a way to ruin people's fun/gameplay with loose/literal interpretations of their standard laws.
|
|
#
?
Dec 28, 2013 01:26
|
|
|
Honestly, this was an issue that needed to be addressed back when killing the AI was a traitor objective. I've had rounds where I did everything and still failed at the end, because I uploaded a suicide law and the AI was afk. But since there's no longer an objective to kill the AI, it's not as important an issue as it used to be. I think that due to the concerns other players have voiced, if we're going to go through the trouble of changing up the dynamics of eliminating the AI, we might as well go whole hog and see if there is an acceptable alternative to killing at all. Leave killing as a definite possibility, for anyone with a grudge and a PDA fulla pipebombs, but add in other nonlethal options to accomplish the same goals.
|
|
#
?
Dec 28, 2013 01:29
|
|
|
Just make it so you can steal the AI brain, and then shove it into a PDA in place of a cartridge. Then you can have a kidnap AI without needing telecrystal costing tools. Also have it so the AI can then only send messages to that PDA so it can talk to its captor, and if the PDA gets abandoned maybe someone will pick it up as they hear a PDA beeping and investigate leading to the AI being revived.
|
|
#
?
Dec 28, 2013 01:37
|
|
|
|
| # ? Jun 14, 2024 13:30 |
|
|
Or the AI is dislocated from the mainframe into a network ghost that can only talk in packets, terminal commands and e-mail messages.
|
|
#
?
Dec 28, 2013 01:40
|
|
