/tg/Station 13

tldr: The AI is piss-easy to kill when it's designed to be difficult to get to, as a role that cannot run away from threats and by default can't kill most people. Fixing this requires code solutions. It's an antag objective you can kill with 2 TC (or 5 minutes of work) with zero counterplay.

So, everyone knows what the AI is. It's a player role designed to observe and aid the crew, at its core. Its job at its core consists of nothing but opening doors for people, alerting the crew when someone is dying through non-suicidal means, and doing mundane tasks with machines. Like a real, actual robot. The AI has a lot of control over the station and by default can see and access the vast majority of stuff. People might say this is powerful, and it is under the right circumstances, but at best it can lock people in rooms (most rooms have windows however so getting out of this isn't hard usually) and at worst it's only a minor annoyance to get around a non-compliant AI. Basic tools thwart any AI.

The tradeoff for this is that the AI can't move, but it is placed a room which is supposed to be designed to make the AI hard to access and kill.

In reality the AI sits in what basically amounts to an eggshell. It must be constantly afraid of any number of things which can kill it. The vast majority of these things have no counterplay for the AI to defend against. The AI can't run away from these, it has to rely on flimsy static defenses to, against a decent player, inconvenience the intruder. You can kill the AI without even entering the chamber, facing any static defenses, or going through any walls.

This is unacceptable. AIs have more counters then any other role in the game. No other role has counters that the player can absolutely not defend against and has no chance to combat or defend against without outside help. Even the Blob, with a huge counter in the xray gun, has a chance of winning before they're even researched, and can actively take actions to prevent this from happening (by eating research or cargo).

The only ''solution'' to preventing dying in your AI chamber requires being out of it, which requires outside help. If you're a traitor AI, you can shunt, but that''s the AI equivalent to giving yourself N2O internals. There's nothing else for you to do in the round but wait for end text, or try to get someone who isn't a validboner to card you (ha ha ha, no).

The following (which isn't an exhaustive list) are ways you can get to the AI with next to no effort:
-EMP. Really the most glaring and easiest way to kill the AI. A traitor with 2 TC can buy an EMP box which can kill the AI on most maps without ever entering the AI core sat. Drain the SMES and the APC and the AI is done. No amount of core defense can prevent this, and with an RCD you can just walk into the AI core and card it or whatever you want to do to make sure it doesn't revive. You don't even need the TC if you're a chemist or scientist or miner or any other job with basic access to a grinder, the ORM, and a few beakers. Xenobio can make infinity EMP grenades as well with yellow cores.
The only downside to EMP is that it's not instant and the AI will bitch for help on vox, though nobody ever saves the AI. Spend a few TC on c4 and you have easy access anyway. An EMP grenade is always spawned on Meta Station, so you don't even need the effort to make or buy some to kill the AI!!
-Cyborgs. AI cannot defend against rogue cyborgs. While this is kind of crucial for ops and shit to quickly kill the AI, it's lame and the AI has no counterplay. the AI core has one flash which only delays a few seconds. People can ride on cyborgs now which makes killing the AI even easier. a rogue AI can be countered just by building a new cyborg or flashing one it owns and unslaving it. Law 1, the AI is harmful, kill it. The AI has no counterplay against this, short of closing the shutters and depowering the doors and hoping it's not an engineer borg (hint: it is).
-Bombs, etc. Not really an issue, the AI will probably see it coming and one bomb usually isn't enough to kill the AI. It's one of the more acceptable ways to kill the AI since it's a static mob, it's loud and obvious, and requires actual effort on the killer's part to finish the job.
-Other AIs. Also not really an issue, most AIs will see this coming in time to deal with it.
-Mechs. Traitor AIs can counter this.
-Random Event (Meteors / Rod). The least fun way to die as an AI. Meteors and Rods treat the core walls like paper. While rod kills are very rare, it's not fun to die to a random event you had no chance to dodge, and probably won't be revived from, since very, very few people go out of their way to save a dead AI. Most people won't even realize it's happened (even the borgs who might never investigate when you stop responding)

-Turret Cheese. Let's face it, actually barging into the core and killing the AI is not actually hard. In fact, a decent player laughs at turrets. The slip foam is countered with magboots, let's ignore that. Killing the AI has become more of a 'how long can the AI stall' rather then 'can the AI actually defend itself' because core defenses are so shitty that they really are nothing but a timesink for intruders. If you die, that's because you were trying to invade the AI on Box Station, the only station with actually good turret design. Or you're just bad, I guess.
"How do I cheese the turrets" you ask? Oh, let me list the ways:
-Just shoot them. Turrets have a long windup time of like 2 seconds. You can fire at them a few times before they even have a chance to fire back. Turrets need LOS to even activate, and when they lose LOS they take another 2 seconds to unwind. Shoot then move behind a wall. If you die to turrets like this git gud.
-EMP them. Because, of course, who needs effort when you can just use the No Counterplay solution to win??
--Shoot them while they're EMP'd, or even better, just wrench them. Unwrenched turrets can't activate again. Good as dead.
-Use smoke. This blocks LOS, which, again, means you can just kill the turrets without them ever able to shoot at you. Or just walk around them, whatever. A chaplain in the AI chamber means the AI is dead.
-Just deactivate them. A two man squad can make their way to the core with little chance the AI can do jack shit about it.The rogue AI could blow the controls after activating them or turn off the power, but that leaves them stuck in one position (or even just off) and you often need both stun and kill to actually hurt anyone.
-Go around them. The most time-intensive solution, you can skip the majority of core defenses by just going around the back and through the walls. C4 them, thermite, abuse the hand tele, or just slowly decon them, eventually you're going to get to the AI and only have to face a couple turrets. Usually intentional design but you can quickly get the walls down with only minor preparation or a few TC.
-Build girders / lockers / any other solid object and block them. This doesn't stop upgraded laser turrets, but does block the stunbolts, which, again, you usually need as AI to kill anyone invading.
-Cut the power. EMPs will do this immediately (unsurprisingly). You can also hack through the wall if it's not the core chamber to manually deactivate / destroy the APC.

Almost all of these methods leave the AI helpless to defend itself. Even careful use of turrets isn't much to stop most people from just cheesing them and going in.



There's multiple code solutions we can use to make the AI actually have some chance to defend itself against the cheesiest ways of killing the AI.

Make AI Sat Rwalls and Doors block EMP blasts
EMP is cancer and the fact you can kill the AI without ever going into its core is cancer. Traitor or not, all AI housing should have basic protection against EMPs. This can either just be a special alloy wall/door that gives obvious protection against EMPs, or just make the default rwalls around the core have a special rule that blocks EMPs. In the latter solution Normal Rwalls wouldn't block EMPs, so if you had to rebuild a sat wall the AI would still be vulnerable.

Significantly Reduce the Turret Windup animation
The main problem with turrets is that they take like 2 seconds before they can even fire, so they often can't fight back. An easy solution would just be to make the turrets pop up instantly or extremely quickly, so they force combatants to actually have to dodge quickly and make them harder to actually kill.

Give Traitor AIs More Defensive Buyables
Traitor AI usually goes one of maybe 4 (or a combo therof) ways (Gas flood / siphon, autoborger, do nothing, or hope the round lasts two hours dominate a PHAZON and sax away while nuking) on top of the usual 'shock all the doors and overload machines'. Giving the AI a reason to actually want to defend its sat and not just teleport away to the mining base would make defending the core actually fun.
--Give it a module that EMP-proofs the whole sat. This powerful, probably 50 Cycle option completely neutralizes the one threat every AI hates. Combined with EMP-proof walls, ensures that even if a traitor goes in the sat or hacks through the walls, they'll actually have to deal with the turrets and other things. Includes the APCs, turrets, SMES, doors, and cameras.
--Give it a module that lets the AI player-control turrets. Something I've always wanted, especially since pixel projectiles made it so the turrets couldn't do their weird wall-bending shots, was the ability to actually dictate where the turrets fired. Either giving them a tile to concentrate fire on, or actually putting yourself in the turret (like a borg shell) would make defending the core amazingly more robust since you don't have to worry about requiring Line of Sight or someone abusing the AI by hiding behind a wall and firing at blind spots. Defending the sat should feel like a game of tower defense or an RTS, not just hoping someone will make a mistake and get killed by turrets. This also solves the core problem of dealing with rogue borgs, since you can now force the turrets to shoot at them. It might not actually be enough to kill the borg before it kills you, but it will definitely be an actual deterrent or stall them.

I actually have a lot of ideas for additional modules a traitor AI could buy. Auto-installing borg upgrades, the ability to place rechargers (or upgrade them since science never does), ability to hack borgs without the Robotics console, but that's ideas for a different thread.

AIs should be more robust and we need to cut down on AI sat cheese. Enough is enough.

Nice

Make reinforced walls block EMP's. Problem solved.

How about making a wall with special plating that results in a reinforced played wall that blocks EMPs.

D&B wrote:How about making a wall with special plating that results in a reinforced played wall that blocks EMPs.

This seems like the most reasonable solution to me. Buffing r-walls all around is a bad idea.

Make AI Sat Rwalls and Doors block EMP blasts
EMP is cancer and the fact you can kill the AI without ever going into its core is cancer. Traitor or not, all AI housing should have basic protection against EMPs. This can either just be a special alloy wall/door that gives obvious protection against EMPs, or just make the default rwalls around the core have a special rule that blocks EMPs. In the latter solution Normal Rwalls wouldn't block EMPs, so if you had to rebuild a sat wall the AI would still be vulnerable.

I had to edit this to be a little more clear. The latter solution was to make the sat rwalls snowflakey, like the turrets are.

The point being that the AI sat / chamber needs basic protection from outside EMPs kicking the AI's ass with nothing it can do. As soon as the APC and SMES runs out of power you'd better hope you didn't bolt down your core because nobody's going to get into that shit (but if you did, the traitor can just crowbar into the sat at that point anyway and you're fucked, so it's a lose-lose)

The AI Sat doors should block EMPs as well, so that you have to actually individually EMP every room to get through the sat.

I like the former solution a little more though (maybe plastitanium should block EMP and be treated like an rwall construction wise? this would have the double benefit of making the Operative Ship more robust) so that you can use the alloy for non-AI sat purposes.

At the end of the day the AI is still by far the most powerful player on station and harder to kill than anyone else.

We've already ingrained a philosophy for YEARS that when the AI goes rogue you should just call the shuttle because the sheer annoyance of trying to break in when there's still a dozen other antags on board is rarely worth it. It's not impossible but it's extremely tedious.

Yes its easier for traitors to do it but considering traitors can kill anyone on station by clicking them twice with a revolver, its not crazy to have the AI beaten down by a kit dedicated to fucking up electronics.

No more coddling for AI players.

Pingsky could have a mode to teleport people away like swarmers do if you want to go super hard mode

Rogue AIs can already win, they just have to actually surprise the crew, and take out the important people quickly. If you slowly fill the hallways with plasma as the CE and engineers arm up for war, you've already lost.

give them flamethrower turrets, rule of cool.

UnicornMan wrote:I just wonder why the AI needs to be more guarded than other players

other players can move...

APC jaunt removes you from the round in a shitty way. By the time you've jaunted you might as well be dead or ghost. You can't do anything in an APC but hope nobody finds you or hope someone's willing to card you and not wipe you immediately. The same goes for being in an intellicard. Better hope you're not an antagonist or you're just going to get validwiped anyway. The point of being in the AI core is that you're protected from EVERYTHING ELSE until someone enters. However, the AI core provides almost zero protection against a reasonably decent player with a gun and an RCD. God forbid if they have EMP, you're dead no matter what.

Every other antag can move to avoid threats. As a carbon you can always avoid getting shot, not getting near suspicious players, playing safe. AI doesn't enjoy that luxury. An EMP outside your core is a death knell, there is absolutely no way to counter it. A few EMPs means you have no power, antag or not you're fucked. Borgs aren't going to be able to get in to get you unless you left the door unlocked or it's an engi borg with an RCD, much less anyone else.

the point is if you can't feel at least slightly safe in a room literally designed to protect you, why should you feel safe anywhere else? If I have to play hide and seek every round so I don't get killed by a 2 TC trick, there's no point in having the core chamber in the first place.

You say "oh, what about the borgs" but AI doesn't always get cyborgs. On top of that, cyborgs are 1) other players, 2) not always on your side, and 3) not always competent. Not only do I get rounds with no borgs, I often get rounds where my starting borg doesn't know where the AI sat is, much less how to play cyborg. Telling me I should "suck it up" and use resources that I don't have isn't a fair response, and doesn't help me prevent a death I have zero self-control to avoid. The same goes for any other 'resource' you can use to get your core moved. If you happen to be an antag, good fucking luck getting a non-borg to move you. If you aren't, you're often putting them at risk, which is bad for your laws. It's a lose-lose. As a player I shouldn't be forced to absolutely rely on another player just to protect me from the absolute most basic threats. There shouldn't be a reason someone can spent 2 TC, stand outside my AI sat, and kill me, with no counter and no way to avoid it.

Or better yet, you do get a cyborg, except that the robo cut the AI wire, and then told the borg to kill you. What am I gonna do, run away?

Yes, there is counterplay to playing offensively... for everyone else. Yes, stuns suck, but you can see someone pull out the gun, or cult paper, or baton, or whatever. It's always a visible sprite you can recognize and react to. You can move. Non-AI players can move anywhere. They can avoid people in space, in maint, in lockers.

The AI just sits there, and takes it. Sometimes it has borgs to help, other times you're fucked buddy, looks like someone spent 2 TC to kill you, ripperini oh well.

I mean, shit like wizards I can understand. I'm not suggesting we anti-jaunt the AI core but anything else requires you actually go in the AI core and bash it till it's dead. There's no challenge to just tossing a 2 TC grenade and bam, dead, no entry required.

It's just extremely inconsistent how hard it is for people to breach the core.

Sometimes I kill dozens of people who try to storm you, other times a single hulk just smashes in and kills you. Or doesn't even smash in and just uses TK to bash you to death from outside....

Also why is jaunt even in the game? It adds nothing to the game but a way to cheese survival objectives or doomsday timers as an AI.

congratulations! You survived in a form where you can do absolutely nothing!

You can have your cake and eat it too. Just make the upgrades researchable in science and have them be manually applied in the core. It's not an instant increase in power, probably won't be done most of the time, but can still be used in medium to long games.

I do actually like the idea of cutting the painfully slow "I AM VULNERABLE SHOOT ME" windup animation on the AI turrets by, like, half. Currently its "stick head out, kill a turret with your shotgun while it winds up, step back into cover, wait five seconds for turrets to go back to shielded mode, rinse infinitely"

Not-Dorsidarf wrote:I do actually like the idea of cutting the painfully slow "I AM VULNERABLE SHOOT ME" windup animation on the AI turrets by, like, half. Currently its "stick head out, kill a turret with your shotgun while it winds up, step back into cover, wait five seconds for turrets to go back to shielded mode, rinse infinitely"

You can run out of range just long enough for the turret to run its pop-down animation, requiring it to finish that AND the pop-up animations before it can fire again. If the airlocks are not bolted or you have breached the AI's R-walls, in most chambers, that gives you just enough time to get inside the AI's windoor if you have access. The windows will block taser shots, but you are putting yourself at risk if the AI is not Asimov or you are not human.

If you are breaking into a chamber from a point that lets you reach the APC easily, and you have a screwdriver + multitool, you can use this trick to rush and short out the APC, not only shutting down the turrets but disabling the AI as well! Using a multitool instead of wirecutters means you can do this without insulated gloves!

Often times the player who kills you will be someone you didn't expect to whack out an SMG or an e-gun or a taser or a stun baton or an explosive lance or a revolver or anything that removes you in one click because let's not pretend the next click is what kills you, it just finishes you off.

This is a player problem, though. Taking that shit out gives an obvious sign, and you can move out of the way or even pull out your own weapon (if you have one) and stun them first. This is counterplay and requires them putting themselves at risk to get close to you to kill you. It's nowhere near the same as cooking up / buying a few EMPs, sitting outside of the AI core, where the AI can do nothing to stop you, catch you, run away, or anything else short of alerting the crew and hoping somebody else comes to save you, and then killing you with just EMP. There's nothing you can do. No time spent breaking down walls, fighting turrets, or any risky and loud and time-consuming behavior. EMPs require next to no work, and unlike bombs (which have to be made or have to be waited out with a 2 minute timer) can't be detected and stopped or warned of beforehand. EMPs are silent and deadly and there's no sort of player action you can do to counter or prevent it.

Honestly explain to me how sitting outside the AI sat and getting a free AI kill from a 2TC box of EMPs is fair

Do you play AI a lot?

I haven't played AI much recently, no, but that's unrelated to the change.

Making r-walls block x-ray vision/lasers and speeding up the turret animations would increase the AI's life span substantially.
I see hulks literally break down the R-walls, run past the turrets before they finish their wind-up animation, punch the APc once or twice which de-powers the room and kill the AI at their leisure.

Traitor AIs should be able to hack cyborgs not synced to them and effectively emag them and have the AI as it's Syndicate Agent law 0, maybe make it a 30 cost, 2-use module.


I'd still prefer a way of traitor AIs being able to reinforce the APC they're hooked up to and make it EMP-resistant so you have to do more than knock down two walls and toss an EMP in

Mate they spent 2TCs, learned the meta spot, took time to get there, bought a spacesuit if their role has access to none or spent time stealing one, may or may not have hit your weakspots with their limited number of EMPs and even if they succeeded the rest of the station is clawing for their blood and will revive the AI if they don't destroy the body.

>spent 2 TC, went into EVA with some kind of access (nobody is punished for hacking into EVA now), found the 'meta spot' for easy GG NO RE AI (AI probably won't notice someone came by beforehand unless it's always watching its core for some reason), spammed EMP, AI is now mute due to lack of power, blind so they can't identify who did it, can only use Vox to scream for help (can't give detailed reports of the perp, should they ever regain vision, all power is out and AI is slowly dying behind three sets of bolt-depowered doors, traitor flies to arrivals and takes their cheap EVA suit off, nobody ever suspects it was them. AI is dead and probably won't be revived.

Still sounds like fun to you?

The fact that you basically admit that players are going to be using metagame 'weak points' to kill the AI without any of the effort of killing the AI is pretty retarded.

I haven't read the thread but I like the idea of rwalls blocking EMPs

could create some interesting scenarios, or at least require someone to put more effort into EMPing secure areas than throwing a grenade outside and winning

This is an old topic but honestly it's piss easy for most of the departments or antagonist types to kill or disable a (malf or non-malf) AI.

* Science can outright build another AI or non-synced borg and get it to kill or otherwise disable them, build weapons out of its ass, build maxcap bombs in under 5 minutes, etc. Also free access to any kinds of law boards with enough research. Even great malf AIs are helpless against even a semi-competent science team.
* Security gets loads of ion rifles, flashbangs, and other weapons which will facefuck borgs, turrets, and AIs alike.
* Medical is probably the most helpless department outside of civilian regarding killing an AI, but if you are a crafty geneticist you could hulk in and TK kill an AI or if you are a competent chemist I believe you can still make EMP mixtures pretty easily.
* Engineering gets easy access to hardsuits, can easily weld through the AI sat and kill it while effectively bypassing all of its defenses, can unlock APCs, etc.
* Command (HoP, captain) can get easy access to the AI upload or effectively waltz into science, print themselves an AI upload, and make the AI kill themselves or easily subvert them. If it's malf, they're one of the biggest threats to the AI due to their easy access to hardsuits and high levels of access. I've also seen games where the captain or RD outright metagames and uses their cameras to check the AI core for hacked APCs preemptively.

If you're an antag?
* For syndicates, people have talked about this, but 2 tc, meta spot, etc. It's easy as fuck for syndicates to kill an AI.
* Changelings waltz over to the AI core and EMP EMP EMP and the AI is dead.
* Wizards can teleport in and zap the AI with zero effort.

AIs are one of the most helpless role in the game, anyone with minor knowledge of how the game works can fuck you over with little effort. I wish we'd see some changes regarding this and honestly a malf/traitor AI rework in general. The latter could warrant its own topic on the issue.

i dont know about you but killing the ai without even entering its core by standing outside and mixing iron and uranium is cancerous.

this is an explicit tradeoff given how much visibility and interaction power the AI has over the rest of the station tbh, it should be a glass cannon

oranges wrote:this is an explicit tradeoff given how much visibility and interaction power the AI has over the rest of the station tbh, it should be a glass cannon

Lets be honest all you need to do to defeat an AI is insulated gloves and a toolbelt and maybe a flash if it has borgs.

basically I feel like as of now it is a glass cannon, but it's too fragile of a glass cannon.

oranges wrote:this is an explicit tradeoff given how much visibility and interaction power the AI has over the rest of the station tbh, it should be a glass cannon

it's a fucking cracker cannon that shoots BB bullets.

Make rwalls EMP proof. At least then you have to be inside the core, either by breaking through the walls or hacking through the doors, to kill the AI.

you've failed miserably

EMP is the main concern. EMP is vastly overpowered and ridiculously deadly to silicons for how easy and cheap it is to get. Not only does it stun you for a vast amount of time (for the AI it's like 2 minutes to access your APC to turn it back on, if you had any power left to do so in the first place), but for borgs it can even stun you permanently since it pulses wires randomly now, including the lockdown wire. It goes through walls and relys on denying counterplay to silicons by just completely fucking them over.

I don't have a problem with one competent guy being able to slowly go through every door, fight the turrets and kill them one by one, and smash into the AI core to kill it. Or go through the rwalls and game the AI turrets to kill it. Or hell, yes, make a mech, smash through everything like paper, and destroy the AI. Whatever. A team of people could do it, and even faster. That's not just fair, that's how it should be.
Thwarting a nuking AI who spent 20 minutes hacking all the APCs to get the right to end the round on its terms should require as much time alone as it takes to get the timer down if you're completely (or not very) prepared and you're alone. Having to play hide and seek with my AI core should not be a requirement to nuking or even just surviving AI ROGUE radio calls for more then five minutes.

Even another AI, while a pain in the ass, isn't a complete death sentence. It can't easily drain your APC of power or anything, so you have to find and battle it out to wrest control of the station with it. However making other AIs is sort of discouraged in both IC and meta, and I've occasionally made posibrains only for salty ghosts to take them and ask to be AI to ''realize'' the other AI is rogue and kill it.

What I do have a problem with is one guy being able to breeze past the entirety of the AI core defenses with little to no effort. Hulks smashing into the core (let's be honest, genetics ruins every game type). Unsync'd borgs being unstoppable by an AI. And especially, ESPECIALLY someone who just sits outside the sat, spams EMP, and calls it a day.

I'll try and be a reasonable human being for a change and say that it feels like the AI has to put in little effort to have big impact on a round

I will absolutely agree with you on this point here. A malf AI should have to do a little hacking to get shit done. However, I think a malf AI should have more options at its disposal, to deal with threats that are normally an instant counter against malf AI play.
-Completely EMP proofing the AI sat as a module should be a high cost, high value defensive item that will completely wipe off an immediate and annoying counter to killing the AI. Force crew to go through the sat a hard way, or at least a creative way.
-Malf AIs should be able to buy off borgs. Let's be honest, unsyncing borgs is an incredibly deadly counter to malf and an AI shouldn't be ruined by something that is absolutely not inconvenienced by turrets or doors or often the shutter. Being able to buy off a module that forces a borg synced (it should ruin the wire too to prevent another desync) is something I would consider fair and the AI has to work for it, and gives the AI another method of counterplay against obnoxious GG NO RE tactics.
-I should go further and say borgs should be hackable remotely in Malf. The robotics console is an annoying asset to protect in a borg-heavy malf play since you simultaneously need it running to hack borgs and need it destroyed to not have borgs blown. Borgs can't move the console so I am usually forced to autoborg in science to prevent the highest chances of someone blowing borgs with any of the machines there. Even if it needs a cost to hack borgs without the console I would be okay with that (as long as it's reasonably cheap), since I still have to do some work to make borgs better weapons.

Like I said, I have a lot more ideas for malf modules that are both offensive (manual turret control) and defensive/supportive (AI-deployable borg rechargers, AI-grantable borg upgrades) but it requires people to not bemoan every instance a silicon gets anything good or new. Hell, I wouldn't mind porting the Malf research tree or abilities from bay (ability to hack other AIs under your control would be fucking sweet) as long as you get more toys for malf to play with. People come up with new traitor toys all the time, why can't malf have a few more options it can buy?

Procedurally generated AI core so people can't meta the weak spots is probably the best solution

The only defence a malf AI truly has is offence.

Burn down everything that is a threat before it can deploy the hard counters against you.

If I have a borg that is somewhat capable with atmos this will mean every part of the station connected to a vent is burning 10 seconds after initiation.

Without capable borg the flood will be rather meh. But still good enough to burn EVA, genetics, science, the teleporter and main hallways.

A very little known trick about AI core defence is that you can unbolt and move turrets around after turning them off from turret control. Unfortunately, this fact is so obscure that it blows all of my borg's minds every time I tell them how to do it

RCDs you can counter with the destroy RCD power but it's quite expensive and it doesn't destroy cyborg RCDs (which can deconstruct R-walls for some reason), and an unsynced cyborg is public enemy number 1 against an AI.

Malf powers I would like: Disguise hacked APCs, Remote emag, Purchase AI shell and Place Turret.
I'd also like r-walls to block EMPs because they're extremely cheap for traitors and changelings to buy. A traitor can buy an entire box of grenades w/ implant for only 2 TCs, all he has to do then is stand in the right spots and EMP the SMES and the APC (To turn off the SMES output and drain the APC and basically put the AI in crit).

Also the robo factory is really easy to destroy for how expensive it is, it was never designed to be destroyable so it has generic machine health and someone with a fire extinguisher can smack it to death.

CosmicScientist wrote: Also do you mean mindless as in vacant or mindless as in catatonic/suicide?

That is a good question. I have yet to decide. Humans have to be alive in order to be borged by the machine, so there is a good chance that captured humans will have players. Should the autoborger forcibly grab a player's ghost if he tries to escape force-borging? I am leaning towards "No." on this one.

So to answer you, the machine should create AI shells instead of borgs if the player has ghosted, or is otherwise catatonic. If he is disconnected, he will simply be a borg upon reconnecting as expected. Since the autoborger requires living humans, suicides are not a concern.

EDIT: Oh, I noticed that Xhuis goofed his refactor and made the RCD power one-use. I am going to bug him to fix that.

@timer 24h code EMP block by copying explosion block calculation code
Also someone escort cosmic from all ai threads he's clearly too unrobust to do anything to ais when the rest.of us have no issue.

CPTANT wrote:Also why is jaunt even in the game? It adds nothing to the game but a way to cheese survival objectives or doomsday timers as an AI.

congratulations! You survived in a form where you can do absolutely nothing!

It auto cancels doomsday if you leave the core

>i promise to code emp protection
>i still haven't done it yet
so much for the ""admin"" """""coder""""" who will never burn out.

ExcessiveCobblestone wrote:

CPTANT wrote:Also why is jaunt even in the game? It adds nothing to the game but a way to cheese survival objectives or doomsday timers as an AI.

congratulations! You survived in a form where you can do absolutely nothing!

It auto cancels doomsday if you leave the core

Well, technically shunting won't cancel the doomsday - the core being killed will cancel it but living as an APC copy isn't a very fulfilling existence

fun fact: "shunt" is a portmanteau of "shit" and "cunt"

DemonFiren wrote:fun fact: "shunt" is a portmanteau of "shit" and "cunt"

WHAT HAVE YOU DONE, YOU FILTHY LIZARD!? Now that is stuck in my head...

It would be nice if malf AIs could convert unsynced borgs by using the override machine module.

Gun Hog wrote:

DemonFiren wrote:fun fact: "shunt" is a portmanteau of "shit" and "cunt"

WHAT HAVE YOU DONE, YOU FILTHY LIZARD!? Now that is stuck in my head...

hhhhehehehhe

On the subject of EMP plating, could that be implemented the same way thin windows are? so you build the R wall and then stick the plating on top?

give AI build mode in their core for 60 points

can place up to eight turrets, four rwalls, one shield, and four pieces of glass

(this was my 1000th post)

John_Oxford wrote:give AI build mode in their core for 60 points

can place up to eight turrets, four rwalls, one shield, and four pieces of glass

(this was my 1000th post)

this is actually a great idea

another idea for malf modes: electronic devices tied to a hacked APC cannot be used by anything other than the AI who hacked it and are automatically hostile to everything. remove borg RCDs being able to deconstruct reinforced walls (why is this still a thing)

I have a PR up to fix the cyborg RCDs. It's literally changing a 1 to a 0