Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Ozarck]

It's down or down-like.

Also, I dislike how the AI was played last round, and am frustrated at my ethereal jaunt's failure in the round before that, when it was needed most. I was healthy and had it fully charged. I should have been able to do it. Eighteen more seconds was all I needed ... damn it. Today has not been my day.

[Jacob/Lee]

Impeccable timing on that crash.

[Ozarck]

It's back
It froze, but it's back

edit: it's in and out. One minute it runs smoothly, the next .. super lag

[Cheedows]

My computer RAM usage shot through the roof when it crashed coincidentally.

[Hanslanda]

I have to ask, though, what are your opinions on subtle traitoring? Not "Objective #1 finished, keep low till round end", but quietly subverting things and escalating existing problems. I can think of a few rounds where after finishing my objective I've done little things like sabotaging equipment, helping space vines spread, lighting fires, stealing critical equipment and spacing it, starting feuds and taking sides in existing ones, giving out weapons to dangerous people, etc., as opposed to the more traditional "bomb the everything" or "lelvalid kidnap+torture, then run around shooting everyone".

All of this is excellent traitoring. The best traitor can make the round chaotic, cause destruction and mayhem, and have everyone at round end go, "HOLY SHIT YOU SON OF A BITCH THAT WAS YOU?!"

Do you have any other suggestions on how to make the round more interesting as a traitor? Recently I've been having fun stealing as many APC batteries as I can, while also spacing every crowbar i could find.

Hans and Jacob are expert traitors. Ask them for ideas on how to make the round interesting. I am hilariously unrobust and don't traitor often (willingly), so sorry I can't help you much.

That is a hilarious idea, Tjoh, and stealing all the tools on station would also be an awesome goal. You don't need to be robust to be a traitor either. Sure, it helps if you have to kill someone, but you can find other ways to make them dead. If you could somehow manage to frame them for a capital crime and get them executed, or trap them in an area with space carp. Or flood the station with CO2 and steal their internals. There's a VAST plethora of ways to kill someone indirectly.

[miauw62]

I just wish the codewords were better. Teamtatoring is great fun.

[Nightscar982]

It is also a pain in the ass when you get only the responses, or nothing at all

[Ozarck]

I just wish the codewords were better. Teamtatoring is great fun.

When you discover that you were, in fact, NOT the otehr traitor's objective, and that they are NOT going to kill you out of hand, then yeah ... team traitoring is great.

It is also a pain in the ass when you get only the responses, or nothing at all

+1.   THis is why I barely look at the code words anymore. I get such a partial list, and never have time to insert them into conversation in a way that a potential ally is going to recognize or even see.

[Damiac]

I just realized something about the way the Asimov law 1 is stated:

You may not injure a human being or, through inaction, allow a human being to come to harm.

So this law has 2 parts.
Part 1. - You may not injure a human being.

This is very specific and very direct. The AI cannot do something which directly injures a human.  Notice it says absolutely nothing about indirectly allowing harm here.

Part 2. - (you may not) through inaction, allow a human being to come to harm.

This is the less specific and direct part of the law, but, there is one very specific part. You may not, through inaction allow a human being to come to harm.

So, putting those two things together, that means as AI, if someone says, "AI, open the armory" you should do it, to properly follow your laws.

That's because opening the armory is an action, not an inaction.  So part 2 of law 1 does not apply.  Opening the door does not injure humans, so part 1 doesn't apply. So since law 1 doesn't apply here, law 2 does. So, open that door AIs!

Now, after you've opened the armory door, law 1 might demand you immediately close the door and lock the guy in there, since not doing so would be inaction, and could potentially cause human harm. That's up to the AI player's judgement I suppose.


My point is, however, that law 1 provides no reason for the AI to EVER refuse to take an action that doesn't directly injure a human being.
Law 1 only demands the AI consider potential harm from NOT taking an action.

This actually prevents the worst of the AI behavior, I think.  If I say "AI, open EVA", the AI has no reason in its laws not to do so.  Of course, the AI might think it's prudent to tell the rest of the crew that I have been allowed into EVA, but that's it.   

When you guys play AI, how do you feel about long lasting orders? Like if at the beginning of the round, the captain says "AI, don't let people in places they don't have access".  What do you do when someone asks you to open a door? You're sort of stuck breaking law 2 either way.  Obviously you'd ignore this order if someone's life was in danger (Not opening the door is inaction, so law 1 is pretty strict about you doing it)

[IronTomato]

Heh, I forgot to share this before.

Being a terrible Ling, I tried to kill a victim in broad daylight in the middle of the meeting hall. I was unable to subdue him, and was captured with someone else (who I think was also a Ling) and the HOS interrogated both of us. The methods we used to avoid talking were... Humorous, to say the least.

Just as it seemed like I had no excuses, a monkey walked in, and promptly walked back out. (How he had security access I don't know) We both shouted "MONKEY!" and promptly started talking to each other about monkeys, completely ignoring the HOS's questions. We got our asses hauled to the Gulag, but it was still funny.

[Tjoh]

Teaming up as traitors has only failed me once, when my partner got caught and hadn't deleted our messages in the PDA. The other times we have helped each other out, giving supplies, breaking out of the brig etc. My favorite was the time when the AI helped be steal the hand tele in exchange for murdering the clown. Very few people seem to take the initiative to team traitor, but it is very Fun!

[miauw62]

-snip-

>not opening EVA
>law 1
>EVA ever causing human harm


Teaming up with other traitors has failed me. That one time when I planned to subvert the AI, talked about this with my fellow tator, had to go AFK, and when I came back I was dead.

[Damiac]

-snip-

>not opening EVA
>law 1
>EVA ever causing human harm

Yeah... but it's pretty damn common.

But like I said, technically, if you read the exact text of law one, you should NEVER refuse an order that is not to "Cause injury to a human being".

So, not only do you HAVE TO open EVA as an Asimov AI, you also HAVE TO open the armory, the permabrig, or whatever else I order you to open.

The AI's only command as far as "Allowing" human harm is not to do it via Inaction.  Therefore, the ai should not even consider the possible harm that could happen by following any order.  Unless said order is to directly "Cause injury to a human being"

So if I'm locked in perma, and I put a note to the camera that says "AI, let me out", there's no provision in law 1 allowing the AI to refuse my order (Will letting me out injure a human? No. Will letting me out allow harm through inaction? No, inaction has nothing to do with opening a door).  Of course, if someone sensible in command simply told the ai "Don't let anyone out of perma without my permission" Then the AI has law 2 justification to keep me in there.

[Graknorke]

I think maybe you're missing the meaning of inaction.
If someone is drowning, and I'm the lifeguard but busy reading a book, they still drown due to my inaction, even if I am doing an action (reading, in this case).

[miauw62]

-snip-

>not opening EVA
>law 1
>EVA ever causing human harm

Yeah... but it's pretty damn common.

But like I said, technically, if you read the exact text of law one, you should NEVER refuse an order that is not to "Cause injury to a human being".

So, not only do you HAVE TO open EVA as an Asimov AI, you also HAVE TO open the armory, the permabrig, or whatever else I order you to open.

The AI's only command as far as "Allowing" human harm is not to do it via Inaction.  Therefore, the ai should not even consider the possible harm that could happen by following any order.  Unless said order is to directly "Cause injury to a human being"

So if I'm locked in perma, and I put a note to the camera that says "AI, let me out", there's no provision in law 1 allowing the AI to refuse my order (Will letting me out injure a human? No. Will letting me out allow harm through inaction? No, inaction has nothing to do with opening a door).  Of course, if someone sensible in command simply told the ai "Don't let anyone out of perma without my permission" Then the AI has law 2 justification to keep me in there.

I was just talking about the fact that EVA contains things TO HELP HUMANS SURVIVE SPACE WHICH IS A PRETTY COMMON THING ABOARD A SPACE STATION.

An Asimov AI should ALWAYS open EVA if somebody asks, because THERE IS LITERALLY NOTHING HARMFUL IN THERE.