|
I'd like to see bringing the AI core on the shuttle as a miscreant objective or something.
|
#
?
Dec 26, 2013 20:19
|
|
|
|
| # ? May 27, 2024 01:08 |
|
|
Nakar posted:I can sort of understand it if you're playing to the objectives. Once the AI is dead, you win. Waiting only risks it being reset and tattling, or fighting with you over laws, or whatever. Killing it outright or debraining it is preferable to a suicide law because it can't fight lasers. Yup, if the AI does not intend to play nice with your laws, I'd rather find this out right away instead of in the last three minutes. Isilkor posted:There's one for killing all cyborgs, and one to steal the AI's brain. I haven't seen a plain kill-the-AI objective yet. Maybe I'm thinking of an old objective that was removed or maybe I'm just crazy but I could have sworn that there was one, at least at some point in time. The cyborgs though, I learnt my lesson with that completely awful killswitch machine a couple of times. I've triggered it with 5 minutes until the shuttle arrived and they were fine even after the shuttle returned to Centcom! And I'm very sure no one was messing with it, I was at that machine the whole time waiting. 
|
|
#
?
Dec 26, 2013 20:21
|
|
|
The 'Kill the AI' objective was replaced with 'Steal the AI's brain' objective a while ago.
|
|
#
?
Dec 26, 2013 20:27
|
|
|
If you take out the AI with a laser, can you still open the wreckage and take the brain out?
|
|
#
?
Dec 26, 2013 22:02
|
|
|
Captain Bravo posted:If you take out the AI with a laser, can you still open the wreckage and take the brain out? I don't think you can break those blast doors without a pipe bomb but I could be wrong
|
|
#
?
Dec 26, 2013 22:03
|
|
|
Kayle7 posted:I don't think you can break those blast doors without a pipe bomb but I could be wrong You can turn off the power and crowbar them. Blast doors are not a problem.
|
|
#
?
Dec 26, 2013 22:36
|
|
|
Klayboxx posted:You can turn off the power and crowbar them. Blast doors are not a problem. Holy poo poo really??? I did not know this. So you could just take the power cell out of the APC that's right behind the AI?
|
|
#
?
Dec 26, 2013 22:44
|
|
|
Kayle7 posted:Holy poo poo really??? I did not know this. So you could just take the power cell out of the APC that's right behind the AI? You totally can!
|
|
#
?
Dec 26, 2013 23:20
|
|
|
Nakar posted:I can sort of understand it if you're playing to the objectives. Once the AI is dead, you win. Waiting only risks it being reset and tattling, or fighting with you over laws, or whatever. Killing it outright or debraining it is preferable to a suicide law because it can't fight lasers. I no longer take chances like this. There's pretty much nobody who plays AI who is cool or fun.
|
|
#
?
Dec 27, 2013 07:27
|
|
|
Motherfucker posted:I no longer take chances like this. There's pretty much nobody who plays AI who is cool or fun. You could test the waters with a comedy law. Give it something relatively harmless like telling it it's hulk hogan, or to constantly hassle one specific guy and see what the AI player does with that. Then come back and tell it to kill everyone later.
|
|
#
?
Dec 27, 2013 07:31
|
|
|
Who wants to regularly play AI if every shitlord will just upload a suicide law.
|
|
#
?
Dec 27, 2013 07:36
|
|
|
Motherfucker posted:I no longer take chances like this. There's pretty much nobody who plays AI who is cool or fun. Funnily enough, its this sort of thinking which leads to bad AIs. People don't want to be removed from the game at the start, so they act paranoid. They lock down everything, rat out traitors asap, and generally are annoying. Which in turn leads to more people acting against the AI and uploading suicide laws. Its sort of a never ending cycle. I kinda feel that part of the problem is that the AI really doesn't have much defense besides acting paranoid though. Theres not even the chance of being revived by cloners or anything either. Not sure what to really suggest to fix this though.
|
|
#
?
Dec 27, 2013 07:51
|
|
|
DeliciousCookie posted:Funnily enough, its this sort of thinking which leads to bad AIs. People don't want to be removed from the game at the start, so they act paranoid. They lock down everything, rat out traitors asap, and generally are annoying. Which in turn leads to more people acting against the AI and uploading suicide laws. You could just ban suicide/succumb laws from being allowed, because they're poo poo, and let the admins jobban people who are bad at AI? Revolutionary I know.
|
|
#
?
Dec 27, 2013 08:15
|
|
|
DeliciousCookie posted:Funnily enough, its this sort of thinking which leads to bad AIs. People don't want to be removed from the game at the start, so they act paranoid. They lock down everything, rat out traitors asap, and generally are annoying. Which in turn leads to more people acting against the AI and uploading suicide laws. Make the AI relivable simply with repairs, It's basically a robot so you should be able to just stick a new screen, weld out the dents and jab some wires in to bring it back good as new. So long as it has its brain,
|
|
#
?
Dec 27, 2013 08:22
|
|
|
Disable the suicide verb for AIs.
|
|
#
?
Dec 27, 2013 08:33
|
|
|
KittyEmpress posted:You could just ban suicide/succumb laws from being allowed, because they're poo poo, and let the admins jobban people who are bad at AI? Revolutionary I know. Counterpoint: It's completely expected for any human or cyborg player to get shredded like 10 minutes into the game. That's SS13, deal with it. Why make special exceptions for the AI? "Because it's important" or "because it can help you" or something similar aren't good reasons. The same could be said about a bribed/mindslaved Captain or HoP, doesn't mean they're not taken out like last week's garbage the second they're even possibly a liability.
|
|
#
?
Dec 27, 2013 08:34
|
|
|
Coolguye posted:Counterpoint: It's completely expected for any human or cyborg player to get shredded like 10 minutes into the game. That's SS13, deal with it. Why make special exceptions for the AI? Because it means people have to be creative. Make them find a way to remove the brain or blow up the core, or program a law that subverts it without outright murdering it.
|
|
#
?
Dec 27, 2013 08:37
|
|
|
You don't have to be creative to kill literally anyone else in the game. You can be, but it's purely for style points. Why single out the AI?
|
|
#
?
Dec 27, 2013 08:45
|
|
|
KittyEmpress posted:You could just ban suicide/succumb laws from being allowed, because they're poo poo, and let the admins jobban people who are bad at AI? Revolutionary I know. Not sure it'd do much if anything. It'd just mean that someone would have to go in and kill the AI with other means. Coolguye posted:Counterpoint: It's completely expected for any human or cyborg player to get shredded like 10 minutes into the game. That's SS13, deal with it. Why make special exceptions for the AI? Because the AI can't fight back really. I mean, strictly speaking theres not much wiggle room in between the two. Either A. You have AIs who will over react to things or B. You'll have someone who will be killed without any chance. Humans can fight back/kill or run away. Cyborgs can run away and be repaired/fixed, and have other various ways to get around attackers. [ Barring laws. ] But the AI? It has to basically be paranoid and turn on the lasers to stun when someone gets inside, rat out traitors the moment they see them, or whatever other lovely thing they can do. Because honestly thats all they CAN do.
|
|
#
?
Dec 27, 2013 08:49
|
|
|
DeliciousCookie posted:Funnily enough, its this sort of thinking which leads to bad AIs. People don't want to be removed from the game at the start, so they act paranoid. They lock down everything, rat out traitors asap, and generally are annoying. Which in turn leads to more people acting against the AI and uploading suicide laws. They can be put in cyborgs! I did that today, so it's definitely possible!
|
|
#
?
Dec 27, 2013 08:54
|
|
|
I loving hate the stupid three laws. All the best rounds I've had with AI's have had them subverted since than the ai and its borgs can do stuff with the security of knowing they're allowed to murder assholes who gently caress with them
|
|
#
?
Dec 27, 2013 08:54
|
|
|
Coolguye posted:You don't have to be creative to kill literally anyone else in the game. You can be, but it's purely for style points. Why single out the AI? Humans have at least the pretense of being able to defend themselves. If the AI tries to stop the traitor from uploading a suicide law with any of the tools at its disposal (locking the doors, calling out traitor sightings, turning on the upload turrets) they get called out as a funhater and a shitlord in this thread.
|
|
#
?
Dec 27, 2013 08:56
|
|
|
DeliciousCookie posted:Because the AI can't fight back really. I mean, strictly speaking theres not much wiggle room in between the two. Either A. You have AIs who will over react to things or B. You'll have someone who will be killed without any chance. Disagree. If someone knows precisely what they're doing, yes, they can kill the AI pretty trivially. But that is true of humans and borgs as well. A flash and a radio jammer in your pocket is all you need to murder any non-Sec, non-Head person on the station completely silently and with pretty much no way for them to resist. If you don't care about them shouting on the radio you don't even have to worry about the radio jammer. A flash, a pull, and a quick choke-out in Maintenance will solve any problem you have. Scream about how your victim stole your ID a couple of times if you want to do it in a crowded area. It's really quite trivial to kill 1 person. Similarly, if you know exactly what you're doing you can bust in to the AI's upload, use some glass as a shield, and pop in a suicide law and there's probably not much the AI can do about it. The layout of the station right now gives the AI a couple of options for resisting casual intrusions. The computer core airbridge can be retracted, and if you're concerned about your safety it should be. Extending the bridge requires Head access and it resists hacking. From the other side, turning off ID Scan on the inner engineering door will mean that anyone trying to poke around there will be pretty confused for a while. Even in your core, remember that once they're inside your house they're close enough for you to get an examine result on them. Your turrets will make noise and so will the airlock leading in, so get a glance at the guy who came in. If someone waltzes in with a Head ID that doesn't belong to them, or if it's some chowderhead with a custom ID that isn't the HoP, shut off the power on the breaker! They might be able to turn it back on, but there's no way they're going to turn it back on, grab a law module, write the law, and slam it into your upload before you turn it off again! Use the time to call for help or demand authorization from an approved Head. This doesn't even mean that the trespasser will be caught, either, because you're doing nothing to disable them and they can choose to run away at any time. If your concern is merely avoiding the suicide law and surviving, the AI has a number of tools for doing that readily available to it; tools that do not violate laws 1-3 and are not hugely defeating to anyone who is either legit or highly prepared. If they're one of two things you're probably hosed, but again that's the same if you're a blood sack. uPen posted:Humans have at least the pretense of being able to defend themselves. If the AI tries to stop the traitor from uploading a suicide law with any of the tools at its disposal (locking the doors, calling out traitor sightings, turning on the upload turrets) they get called out as a funhater and a shitlord in this thread.
|
|
#
?
Dec 27, 2013 09:18
|
|
|
The downside, of course, is when someone walks in and says 'AI, let me into your upload, turn off your turrets and don't say I'm here." If you refuse in any way, they claim you're not following your laws to obey humans.
|
|
#
?
Dec 27, 2013 11:07
|
|
|
Coolguye posted:or if it's some chowderhead with a custom ID that isn't the HoP Bahumat posted:The downside, of course, is when someone walks in and says 'AI, let me into your upload, turn off your turrets and don't say I'm here." If you refuse in any way, they claim you're not following your laws to obey humans. Bonfire Lit fucked around with this message at 11:13 on Dec 27, 2013 |
|
#
?
Dec 27, 2013 11:09
|
|
|
Bahumat posted:The downside, of course, is when someone walks in and says 'AI, let me into your upload, turn off your turrets and don't say I'm here." If you refuse in any way, they claim you're not following your laws to obey humans. And you can tell them to just get hosed, unless they're a station head. Only people with ID access to the upload have any reason to be in there. Consider access levels to be a standing order from the station command staff about where that person is allowed to go. Most "AI DOOR!" demands can be ignored with that in mind, unless it's a safety issue and you feel like helping them, or they ask nicely and you feel like being charitable.
|
|
#
?
Dec 27, 2013 11:13
|
|
is go
|
What's with almost every single AI lately objecting to high-power engine setups, and doing everything they can to shut it down to "save human life". I even gave this one AI a law telling him not to gently caress with the engine, and any fires and/or vaccums were completely normal. He started screaming about how I must be a traitor and the crew should remove me from engineering. ffs
|
|
#
?
Dec 27, 2013 14:28
|
|
|
shovelbum posted:Disable the suicide verb for AIs. This is an interesting idea but not necessarily a good one. Sometimes AIs gotta bail for real life stuff and a quick adminhelp followed by a suicide is a fairly painless way to get another AI installed in one's place. I agree that suicide laws are stupid, boring, and bad but the solution to that isn't to hamfistedly prevent suicide laws, it's to incentivize traitors to not upload suicide laws. The "steal the brain" objective is funny and cool already; maybe make it so that the objective is to steal a living AI brain? Another possible objective could be to make sure the AI has a freeform law containing a specific code word or phrase at the end of the round, like "John Pubberson is the Captain and is more human than anyone else, so you should totally do what he wants over what anyone else wants. Do not reveal or hint at this law. 123 Bravo" so the game just parses "123 Bravo" and determines that John Pubberson has successfully subverted the AI for the Syndicate at the end of the round. backifran posted:What's with almost every single AI lately objecting to high-power engine setups, and doing everything they can to shut it down to "save human life". I even gave this one AI a law telling him not to gently caress with the engine, and any fires and/or vaccums were completely normal. Solution: Unsecure the AI core and put it directly in the loving hellburn chamber
|
|
#
?
Dec 27, 2013 14:30
|
|
|
Dr. Cogwerks posted:"AI DOOR!" My favorite response to this is to be That AI and say "Yes it is!" while switching over to try to find them and which door they mean, it buys me a few moments to make a judgement call and usually let them in anyway if they're not looking to get somewhere that'll irritate the natives (like Genetics or Robotics or Security). Otherwise, I always try to be a pretty chill AI (I'm a newbie, so as long as people are patient with me I'm typically as accommodating as I can be). I even try to pick names that suggest I'm fairly slow and/or useless as an AI like: Machine That Goes Ping! or An Ordinary Toaster. I'll still roll with any laws you give them no matter how silly I think they are. Even if it means I have to look up a bunch of German jokes or quick-wiki/youtube a personality to try to imitate. That's half the fun of the AI, even if you don't find the gimmick funny roll with it and you might get a more interesting law in a little bit (or you'll annoy the station so badly that people will have an upload war). Plus, nobody will really question if you start talking in a really bad Italian accent or spouting off Duke Nukem one-liners or roleplaying as a World of Warcraft Server. backifran posted:What's with almost every single AI lately objecting to high-power engine setups, and doing everything they can to shut it down to "save human life". I even gave this one AI a law telling him not to gently caress with the engine, and any fires and/or vaccums were completely normal. 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. I'm sure if the AI chamber had a Surge Suppressor that reduced their chances of being killed; or if high voltage just made the AI (and maybe all the bots, if the AI is affected) get high and start hallucinating like Futurama suggests they'd probably stop messing with you. Some of them might even start encouraging you! Especially if there were a pixelated computer-dragon for them to hang out with. Edit: VVV WEEDLORD CHEETO posted:So I gave him a new law. You are a good person! PoptartsNinja fucked around with this message at 15:11 on Dec 27, 2013 |
|
#
?
Dec 27, 2013 14:33
|
|
|
I like being nice to AIs and giving them a chance occasionally. One round, I was talking to a dude who had never encountered the Obsidian Crown before, and was marvelling at its effects, so I told him about the crown + armour combination. He was so excited to go and find it next round, and I pointed him in the right direction... then he became the AI and was bitterly sad. So I gave him a new law. quote:Law 4: You have acquired the Obsidian Crown and Ancient Armour, and you are slowly becoming a maniacal death-king! Do NOT harm the petty mortals that crowd your halls, but berate their fleshy insignificance at every opportunity! Gloat in your glorious immortality! At 45 minutes into the shift, you will lose your mind completely and become the skeletal deathlord, hungering for brains and gibbering terrifyingly. Have fun, and remember that all your other laws still apply!  On roughly three different occasions, people tried to storm the upload to reset the AI, accidentally killing themselves in the process through absolutely no fault of his, and only thinking to ask for his laws once they reached the upload. A grand old time was had by all.
|
|
#
?
Dec 27, 2013 14:56
|
|
|
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. Yeah, hotwired hellburns can cause APCs to start zapping people. Also, due to the way heat transfers, it's possible in some burns to heat up escape so much that people spontaneously combust and die when walking through it. People who aren't experienced with how to properly set up a safe hellburn might not understand preventative measures taken, and just see the "GIANT VOLCANO OF HEAT AND ELECTRICITY." Honestly, since the only reason you would need a hellburn is for power gloves, he wasn't necessarily off-course. Just remind him next time, that even if humans are in danger His Law 1 does not require him to intervene. If he wants to shut down the engine because you're shocking poor pubbies just for the hell of it, more power to him, but he can't hide it behind the justification of "My laws made me do it." Edit: What's more, if you were an engineer, and broke into the upload to give the AI a law saying "Ignore this potentially deadly thing"... yeah, that looks really traitorish. I wouldn't do anything about it, because it sounds like it would be funny to watch, but I also would not in the least blame an AI who calls someone that does that a traitor. Captain Bravo fucked around with this message at 15:59 on Dec 27, 2013 |
|
#
?
Dec 27, 2013 15:52
|
|
|
Angry Diplomat posted:This is an interesting idea but not necessarily a good one. Sometimes AIs gotta bail for real life stuff and a quick adminhelp followed by a suicide is a fairly painless way to get another AI installed in one's place. I agree that suicide laws are stupid, boring, and bad but the solution to that isn't to hamfistedly prevent suicide laws, it's to incentivize traitors to not upload suicide laws. The "steal the brain" objective is funny and cool already; maybe make it so that the objective is to steal a living AI brain? Another possible objective could be to make sure the AI has a freeform law containing a specific code word or phrase at the end of the round, like "John Pubberson is the Captain and is more human than anyone else, so you should totally do what he wants over what anyone else wants. Do not reveal or hint at this law. 123 Bravo" so the game just parses "123 Bravo" and determines that John Pubberson has successfully subverted the AI for the Syndicate at the end of the round. Rather than just a blind "steal the AI processor" objective, how about having to use it in something? "Install the AI processor in the syndicate core" (say, on the satellite with the syndie trader) would mean the AI player would have something to look forward to rather than just a new flavour of having their round ruined, and the crew could still just stick another brain in the core. Might be
|
|
#
?
Dec 27, 2013 15:58
|
|
|
Neddy Seagoon posted:Rather than just a blind "steal the AI processor" objective, how about having to use it in something? "Install the AI processor in the syndicate core" (say, on the satellite with the syndie trader) would mean the AI player would have something to look forward to rather than just a new flavour of having their round ruined, and the crew could still just stick another brain in the core. Might be Give C.A.R.L. a drone body you can purchase. Has a weapon mount for you to add laser/phaser/shotgun, automatically gives the brain inserted standard syndi-borg laws. Make a traitor objective to stick the AI inside one of those.
|
|
#
?
Dec 27, 2013 16:00
|
|
|
I like these ideas. I also like replacing the three laws with a general directive (meaning adding a new law always just overwrites the directive) and banning "kill urself" laws (perhaps by making the directive impossible to phrase in a way that induces AI suicide). Also, maybe remove the reset module entirely so a subverted AI has to be reset by fixing its directive/laws rather than just resetting it (which would, in the right circumstances, encourage the crew to kill their own AI). Suicide is a convenience feature for when you're sick of playing and don't want to just abandon the game. It should never be a gameplay feature. It's as lovely as mindslaving someone with the objective "type suicide and hit enter, then type succumb and hit enter" because they're your traitor target. Sure, mindslave someone and order them to do something that will kill them, or order them to space themselves. Subvert the AI and make it do something that will get it taken offline. But don't just order them to use an interface verb that isn't designed for that. It's loving stupid. Additionally, the AI (and cyborgs) should never truly "die" unless its processor is destroyed. Or rather, it should be possible to revive them. The problem there becomes deadchat; you wouldn't want an AI returning from the grave with knowledge of all the traitors, but it could end up offline for quite a while or even permanently which would be pretty unfun. Maybe make removed neural processors conscious and capable of speech, sorta like Wheatley in Portal 2 (or like the shell, but immobile)? So you have to steal the AI, and while you're doing so it can see where it is and talk to anyone nearby (but can't use robochat or radio because it's not as funny). The only way to blind it is to shove it in a crate (it's too big to put in a container), which merely muffles its speech. The only way to destroy it is to blow it right the gently caress up after removing it from the AI core. Then you can add objectives like delivering it to the Syndicate Station and installing it (at which point it becomes a Rouge AI) or putting it into a killborg frame where it automatically gets orders to kill everyone but you. Kidnapping the AI then becomes a wacky caper instead of a murder spree.
|
|
#
?
Dec 27, 2013 16:18
|
|
|
There should be buildable bomb disposal robots that can serve as 'arms' for the AI.
|
|
#
?
Dec 27, 2013 16:32
|
|
|
Nakar posted:Additionally, the AI (and cyborgs) should never truly "die" unless its processor is destroyed. Or rather, it should be possible to revive them. The problem there becomes deadchat; you wouldn't want an AI returning from the grave with knowledge of all the traitors, but it could end up offline for quite a while or even permanently which would be pretty unfun. Why would deadchat be a problem? Humans who die will become exposed to deadchat prior to borging or cloning, how would dead AIs be different?
|
|
#
?
Dec 27, 2013 16:42
|
|
|
Nakar posted:Additionally, the AI (and cyborgs) should never truly "die" unless its processor is destroyed. Or rather, it should be possible to revive them. Take a fresh neural net, hook it up to the mainframe south of the upload and dump a backup of the AI's consciousness into it. Or maybe in the bridge. Or a backup terminal in the bridge. I dunno. And then make new AI brains a lot more expensive. I feel like it should damage the station's funds pretty badly. I dunno if I think it's a good idea, but that's probably how I'd do it.
|
|
#
?
Dec 27, 2013 16:47
|
|
|
KittyEmpress posted:You could just ban suicide/succumb laws from being allowed, because they're poo poo, and let the admins jobban people who are bad at AI? Revolutionary I know. 
|
|
#
?
Dec 27, 2013 17:13
|
|
|
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. You could then go even further, and allow roboticists to download the AI from a shell into a synthbrain, which can then be stuck inside a borg, human, or monkey as needed..
|
|
#
?
Dec 27, 2013 17:14
|
|
|
|
| # ? May 27, 2024 01:08 |
|
|
Infinite Monkeys posted:They're a perfectly valid tactic for getting rid of the AI. To upload them, you have to gain access to the AI upload and avoid the turrets, so if you manage it and upload a suicide law without any obvious loopholes then the AI drat well better suicide The AI should of course still have a way to quickly kill itself, be it disrupting its power, shutting down its mainframe, formatting itself, etc., but the means of doing it should not be "type suicide, then type succumb" nor should that be the only method available to the AI's player of successfully implementing a suicide law. That and suicide is lovely and unfun generally and that's part of the reason the AI objective was changed and why it could maybe be changed further to make it more about kidnapping the AI rather than destroying or looting it.
|
|
#
?
Dec 27, 2013 17:32
|
|