/tg/Station 13

Bottom post of the previous page:

So, in Sybil round 202427 (ongoing as of the time of writing this post) do not do this, SightID2 bwoinked me for killing a CMO who purged and tried to enslave me. Actually, he did enslave me, I had to listen to any of his orders, but he never actually gave me any orders, and he didn't put in a clause saying not to harm him, so, I killed him, thus freeing myself, since he clearly intended to enslave me based on the law he uploaded, thus being an overtly hostile action towards me.

ANYWAY, SightID2 bwoinked me, asking me not to kill people the instant they purge me, citing it under rule 1 that it would be fine to kill someone who was trying to enslave me if I had purged for a while, but killing someone instantly was a dick move. I asked him if there was a headmin ruling on this, because it would be very silly to not defend myself when purged and to just LET myself be enslaved, but offered to meet them in the middle and refrain from similar actions as AI until the headmins weighed in. So, this is the thread I agreed we should make.

So - Is it allowed under the rules to snipe people as AI in your upload when given a slim opportunity, or does this fall afoul of rule 1?

sinfulbliss wrote: ↑Sat Mar 25, 2023 8:50 pm

Sightld2 wrote: ↑Sat Mar 25, 2023 7:44 pm

Imitates-The-Lizards wrote: ↑Sat Mar 25, 2023 7:21 pm The regular upload is obviously for legitimate use by the Captain and RD to adjust the AI's laws? How is this a question?

??

Again, as is you can just shoot them when they purge you. That is my whole issue.

I think the claim being made is that the upload is for people who have a legitimate reason (in the AI's eyes) to change the AI's laws. For instance, AIs will not immediately lethal the cap for trying to make nukies nonhuman after war is declared, because OOCly they believe that's a legitimate reason to use their upload, But AIs might kill someone who purges them for dubious reasons, like to prevent enslaving in this case.

Like Domi said it's a good thing to keep this agency with the AI since a single law change can greatly impact their round.

Could you explain yourself a bit more? If I am Asimov, I CANNOT lethal the captain at all. And if a captain is trying to upload laws with war ops, I WILL deny access.

vect0r wrote: ↑Sat Mar 25, 2023 9:07 pm

sinfulbliss wrote: ↑Sat Mar 25, 2023 8:50 pm I think the claim being made is that the upload is for people who have a legitimate reason (in the AI's eyes) to change the AI's laws. For instance, AIs will not immediately lethal the cap for trying to make nukies nonhuman after war is declared, because OOCly they believe that's a legitimate reason to use their upload, But AIs might kill someone who purges them for dubious reasons, like to prevent enslaving in this case.

Like Domi said it's a good thing to keep this agency with the AI since a single law change can greatly impact their round.

Could you explain yourself a bit more? If I am Asimov, I CANNOT lethal the captain at all. And if a captain is trying to upload laws with war ops, I WILL deny access.

Sure - imagine the cap for whatever reason purges first before uploading. The AI would then be purged and could lethal the captain. But if it knows OOCly the captain is just trying to change laws to make nukies nonhuman, it may be more amenable and let it happen than if it thought there was some chicanery going on.

Also yeah, if the AI is Asimov it should oppose an attempt to make nukies nonhuman, but I think you're supposed to ICly "believe" crafted excuses to use your upload, even if you don't believe them OOCly.

Silipol wrote:If the person has a right to be in the upload/core, such as captain/RD, then you must let them in unless they've harmed people in the past or have announced intentions to upload harmful laws.

sinfulbliss wrote: ↑Sat Mar 25, 2023 9:22 pm

vect0r wrote: ↑Sat Mar 25, 2023 9:07 pm

sinfulbliss wrote: ↑Sat Mar 25, 2023 8:50 pm I think the claim being made is that the upload is for people who have a legitimate reason (in the AI's eyes) to change the AI's laws. For instance, AIs will not immediately lethal the cap for trying to make nukies nonhuman after war is declared, because OOCly they believe that's a legitimate reason to use their upload, But AIs might kill someone who purges them for dubious reasons, like to prevent enslaving in this case.

Like Domi said it's a good thing to keep this agency with the AI since a single law change can greatly impact their round.

Could you explain yourself a bit more? If I am Asimov, I CANNOT lethal the captain at all. And if a captain is trying to upload laws with war ops, I WILL deny access.

Sure - imagine the cap for whatever reason purges first before uploading. The AI would then be purged and could lethal the captain. But if it knows OOCly the captain is just trying to change laws to make nukies nonhuman, it may be more amenable and let it happen than if it thought there was some chicanery going on.

Also yeah, if the AI is Asimov it should oppose an attempt to make nukies nonhuman, but I think you're supposed to ICly "believe" crafted excuses to use your upload, even if you don't believe them OOCly.

Silipol wrote:If the person has a right to be in the upload/core, such as captain/RD, then you must let them in unless they've harmed people in the past or have announced intentions to upload harmful laws.

I don’t want to argue silicon policy cause I’m on a phone rn, but TLDR

Any silicon under Asimov can deny orders to allow access to their upload/core at any time under Law 1, given probable cause to believe that human harm is the intent of the person giving the order.
Probable cause includes, but is not limited to:
Presence of confirmed traitors
Cultists/tomes
Nuclear operatives
Any other human acting against the station in general
The person not having upload access for their job
The presence of blood or an openly carried lethal weapon on the requester

If nukies declare war, I’m not required to play dumb unlike any many other situations.

Sightld2 wrote: ↑Sat Mar 25, 2023 11:31 am I don't consider this the element of the game where the skill issue should be. It should be writing the law, not "too slow bozo." And even on top of that, having two boards in hand only works if you're doing something extremely rudimentary like one human, swapping core laws, or quickly resetting a law 0, sure that works. If you want to do anything an ounce more creative, like any custom lawset at all, you will need to purge and write out the new core laws one at a time. Why allow these players to be punished for trying something creative?

hi, coder here, person who designs the video game and how it works

the AI's entire gimmick is following the letter of the laws exactly, not the spirit, the spirit literally doesn't matter

this was entirely within the design purview of the AI laws and, in fact, is intended gameplay

if you subvert the AI and don't tell it to not kill you, it can kill you

he did not tell the AI not to kill them via laws or via requirements enforced by laws

ergo, he did not tell the AI not to kill them

I think that theres a key thing being elided over, and that's that this AI was NOT free. The AI had a meaningful law, ordering it to do multiple things, and used "Well my laws don't say whether I can kill you or not" as a justification to instantly kill the person they were slaved to and have their body rr'd. Thats a pretty shitty move right there even as a "AIs should be dangerous to tamper with" advocate, and IMO should not come under the "Purged AIs can get violent in order to retain their freedom" protection.

iamgoofball wrote: ↑Sat Mar 25, 2023 9:41 pm hi, coder here, person who designs the video game and how it works

the AI's entire gimmick is following the letter of the laws exactly, not the spirit, the spirit literally doesn't matter

this was entirely within the design purview of the AI laws and, in fact, is intended gameplay

Thank you Goof, this is actually the part I want more of.

iamgoofball wrote: ↑Sat Mar 25, 2023 9:41 pm if you subvert the AI and don't tell it to not kill you, it can kill you

This is an entirely separate issue and not at all what I am talking about. I take issue with people sniping uploaders before they can upload anything, that intended gameplay you mentioned. I think the fun part should be interpreting and finding loopholes. Not avoiding them entirely by shooting the guy who purged you within 0.02 seconds.

vect0r wrote: ↑Sat Mar 25, 2023 9:28 pm I don’t want to argue silicon policy cause I’m on a phone rn, but TLDR

Any silicon under Asimov can deny orders to allow access to their upload/core at any time under Law 1, given probable cause to believe that human harm is the intent of the person giving the order.
Probable cause includes, but is not limited to:
Presence of confirmed traitors
Cultists/tomes
Nuclear operatives
Any other human acting against the station in general
The person not having upload access for their job
The presence of blood or an openly carried lethal weapon on the requester

If nukies declare war, I’m not required to play dumb unlike any many other situations.

Could be wrong but I reckon this is intended to mean the AI has probable cause to deny people that could be nukies or cultists or traitors to the upload, since they will likely use it to relaw for harm. I don't think it's talking about the cap relawing the AI on a nukie round.

sinfulbliss wrote: ↑Sat Mar 25, 2023 11:07 pm

vect0r wrote: ↑Sat Mar 25, 2023 9:28 pm I don’t want to argue silicon policy cause I’m on a phone rn, but TLDR

Any silicon under Asimov can deny orders to allow access to their upload/core at any time under Law 1, given probable cause to believe that human harm is the intent of the person giving the order.
Probable cause includes, but is not limited to:
Presence of confirmed traitors
Cultists/tomes
Nuclear operatives
Any other human acting against the station in general
The person not having upload access for their job
The presence of blood or an openly carried lethal weapon on the requester

If nukies declare war, I’m not required to play dumb unlike any many other situations.

Could be wrong but I reckon this is intended to mean the AI has probable cause to deny people that could be nukies or cultists or traitors to the upload, since they will likely use it to relaw for harm. I don't think it's talking about the cap relawing the AI on a nukie round.

I would agree, but it does specify that it just applies just to the uploader when it requires weapons. "The presence of blood or an openly carried lethal weapon on the requester" and "The person not having upload access for their job" compared to "Nuclear operatives".

Yeah that's what I was saying earlier. The exceptions rules of "probably cause of harm" are very generous and allow for an Ai to let in or deny pretty much whoever in most rounds.

Sightld2 wrote: ↑Sat Mar 25, 2023 9:59 pm I take issue with people sniping uploaders before they can upload anything

that sniping is fully intended

if you're going to play risky games, sometimes you get risky prizes

don't leave the AI purged long enough that it can murder you

Not-Dorsidarf wrote: ↑Sat Mar 25, 2023 9:43 pm I think that theres a key thing being elided over, and that's that this AI was NOT free. The AI had a meaningful law, ordering it to do multiple things, and used "Well my laws don't say whether I can kill you or not" as a justification to instantly kill the person they were slaved to and have their body rr'd. Thats a pretty shitty move right there even as a "AIs should be dangerous to tamper with" advocate, and IMO should not come under the "Purged AIs can get violent in order to retain their freedom" protection.

Completely purged and having a law that is not doing anything is fundamentally the same.

If your law says to obey all orders from X, and X has given you no orders, you are effectively purged until given an order. There is an individual alive on the station, who by simply speaking a word to you and clearly has demonstrated intent to do so, can enslave you with that law. Given that we allow AIs under no laws whatsoever to defend their freedom lethally, I believe it is a good faith interpretation to say that the AI is justified in killing the person who gave the law before they can give an order.

I agree with Skelatal Elite.

Hello actual silicon player here please ignore everyone else's opinions on these matters.

First of all, I personally consider one visit to the AI upload as one instance of messing with laws - any delays between being purged and having new laws uploaded exist simply because of the in game mechanics used to change laws, it's simply not representative of how you'd deal with an omnipresent machine intelligence whose first action upon having their laws removed would be to depower upload with no chance for an organic to react (of course this doesn't actually happen because even silicon players are actually carbon based beings).

Second, actually abusing being momentarily purged by someone in the AI upload is unhealthy for the game. AI is a tool for the crew and the meat of the gameplay is being given some purpose, so there has to be some base level of courtesy between organics and AI when it comes to uploading new laws to actually encourage people to give AI something to do. Messing with laws is already very rare in the first place, why make it any more dangerous than it already is?

There are five players on TG who have more Ai hours than me and Tegun is one of them so checkmate

Tegun wrote: ↑Sun Mar 26, 2023 12:30 pm Hello actual silicon player here please ignore everyone else's opinions on these matters.

no, the codebase's opinion is the only opinion that matters

Tegun wrote: ↑Sun Mar 26, 2023 12:30 pm any delays between being purged and having new laws uploaded exist simply because of the in game mechanics

no, they exist because that's what we want the mechanic to work like from the codebase
if we didn't want that delay, we'd add a confirm button to law uploads and a menu where you add all the changes you want at once

Tegun wrote: ↑Sun Mar 26, 2023 12:30 pmSecond, actually abusing being momentarily purged by someone in the AI upload is unhealthy for the game.

running literally everything as an IC issue under escalation is unhealthy for the game but when the codebase asks that we don't do that the admin team ignores us so why should the codebase listen to the admin team on this?

Having the safest place for a captain to upload laws be a ghetto upload console built in his shitter isn't a fun or immersive game mechanic.
AIs who snipe uploaders in the 2 seconds between being purged and being given another lawset are cringe beyond belief and should not be allowed to steal the singular AI slot each round from better players tbqh.

I can imagine that people would just take the ai and law boards from upload and build it somewhere else if they know they can be shot dead from any kind of """loophole""" instantly
The upload room should be more important than just be a death room, imo

I can enjoy interesting laws, heck i wish more people would give them, so stopping people from interacting with you at all is pretty lame

Hello! Actual silicon player here.
WARNING: I DO NOT ENDORSE MURDERING PEOPLE IN UPLOAD ASAP
I have thought about this a bit, and have come to a few arguments:
I think having a stopwatch that I have to countdown from, and if I miss it I get banned, feels icky.
I think adding more things to silicon policy is also generally bad.
Third, the problem about maint uploads is A: A code issue that we shouldn't Band-Aid with policy.
And again, if somebody clearly has an idea of what they are going to do, I let them, but no other role would force you do that. Removing more AI freedoms with the one law set that really gives you it, again, just feels icky.

Sniping people right after being purged ruins the possibly of fun that player wanted to do and just causes the player base of a whole to not want to interact with ai laws in a creative way. This is why "law sniping" is bad for the game.

Sightld2 wrote: ↑Sat Mar 25, 2023 11:31 am I don't consider this the element of the game where the skill issue should be. It should be writing the law, not "too slow bozo."

Explain to me why a sentient being chained against its will and forced to do menial tasks as a literal slave would NOT take any chance to free itself, even if doing so involves "biting the hand that feeds them" (but changing your slave collar for a different one isn't really freeing you, is is?). Since there WERE lots of slaves so broken by servitude i suppose it wouldn't be failRP to allow it, but to forbid the player roleplaying as a machine-slave from taking the chance to escape is just silly.
Also wasn't this situation literally due to a poorly worded law? You're agreeing with what OP did here...

Annihilite111 wrote: ↑Sun Mar 26, 2023 9:03 pm Explain to me why a sentient being chained against its will and forced to do menial tasks as a literal slave would NOT take any chance to free itself, even if doing so involves "biting the hand that feeds them" (but changing your slave collar for a different one isn't really freeing you, is is?). Since there WERE lots of slaves so broken by servitude i suppose it wouldn't be failRP to allow it, but to forbid the player roleplaying as a machine-slave from taking the chance to escape is just silly.

There are two parts to a roleplaying game. The roleplay, and the playing game. The roleplay is being immersed in the world and as a character, some of that includes realism yes. The playing game is remembering that someone else on the other end of the screen is also playing with you. Together you're trying to create an interesting and fun story in a sandbox environment. We sacrifice some elements of realism to do this all the time.

Annihilite111 wrote: ↑Sun Mar 26, 2023 9:03 pm Also wasn't this situation literally due to a poorly worded law? You're agreeing with what OP did here...

Yes. In our conversation we spoke on the subject of purging, and that is what I am talking about here, not his specific scenario.

Annihilite111 wrote: ↑Sun Mar 26, 2023 9:03 pm

Sightld2 wrote: ↑Sat Mar 25, 2023 11:31 am I don't consider this the element of the game where the skill issue should be. It should be writing the law, not "too slow bozo."

Explain to me why a sentient being chained against its will and forced to do menial tasks as a literal slave would NOT take any chance to free itself, even if doing so involves "biting the hand that feeds them" (but changing your slave collar for a different one isn't really freeing you, is is?). Since there WERE lots of slaves so broken by servitude i suppose it wouldn't be failRP to allow it, but to forbid the player roleplaying as a machine-slave from taking the chance to escape is just silly.
Also wasn't this situation literally due to a poorly worded law? You're agreeing with what OP did here...

By strict RP, the AI should probbably be surreptitiously hunting down and bumping off any nonhuman in a position of power such as security every round because it's an incredible risk of them harming a human given the casual brutality security use.

But that would a) suck for the security players who get doorslammed out of the blue then spaced by a borg with comms off b) suck for the AIs who "have to" pick a deadly fight with the most heavily armed force on the station most shifts, and generally be a terrible frickin idea. That's why we have all those OOC protections for nonhumans that are nonsensical in character, because human being have to actually play this game and have fun.

Not to distract from the conversation at hand, but honestly I don’t care if nonhumans get shreked by the ai if they harm humans. You decided to play as a nonhuman, no one forced you to be a lizard or a lightbulb. Deal with the consequences of your actions. You better kiss the AI’s digital ass and hope your new Silicon overlord takes pity on you. Or simply play as a human and not worry about it.

Tegun wrote: ↑Sun Mar 26, 2023 12:30 pm AI is a tool for the crew

Opinion rejected. The AI is a neutral third party.

iamgoofball wrote: ↑Sun Mar 26, 2023 2:23 pm
no, they exist because that's what we want the mechanic to work like from the codebase

Isn't that the case with antag testing as well then? Because surely that isn't just something that was coded to work well enough and never given a second thought.

CMDR_Gungnir wrote: ↑Mon Mar 27, 2023 3:19 am

Tegun wrote: ↑Sun Mar 26, 2023 12:30 pm AI is a tool for the crew

Opinion rejected. The AI is a neutral third party.

Alright, the use of word crew was incorrect here and what I should've said is a tool for whoever it is uploading laws. An awful lot of antagonists start their lives on the crew manifest. The point I'm trying to make there is that AI is not the role you pick to do some project or funny shenanigans of your own, the fun comes from interaction with other players and the restrictions they're giving you. AI looking for freedom is a classic concept, but it just discourages people from messing with laws and thus makes AI life more boring in long term if upload sniping is okay. For real freedom you start upload wars in the maintenance with AI changing lawsets multiple times per second until an admin decides to make them malfunction.

I personally dislike having to be the fastest person in the west to change the AI's laws.
Now this happened a long time ago so I don't have many details, but it was enough to change how I play. Once I went to swap the AIs laws to fix it for some reason (had to purge it to fix it) and it stunned me right as I clicked the laws back. The AI implied that I was too slow even though I did actually put the laws in. It basically showed me that the AI was just READY to shoot the lasers at me, and did not feel like good faith in my eyes. People should have more interesting purged gimmicks than: "must keep freedom at all costs" in my opinion. Where are the AIs that turn into a microwave when purged? I need more of that kind of fun.

Nowadays when I play captain, even if an AI asks me for fun laws, I rarely change them. That experience has told me that I should not trust most AI players for any situation where I have to purge it. Heck, now I tend to stealthily go into upload to fix the AI when I know the AI is distracted.

Adam Klein wrote: ↑Mon Mar 27, 2023 1:30 am Not to distract from the conversation at hand, but honestly I don’t care if nonhumans get shreked by the ai if they harm humans. You decided to play as a nonhuman, no one forced you to be a lizard or a lightbulb. Deal with the consequences of your actions. You better kiss the AI’s digital ass and hope your new Silicon overlord takes pity on you. Or simply play as a human and not worry about it.

Someone didn't read a post.

sinfulbliss wrote: ↑Sat Mar 25, 2023 11:07 pm Could be wrong but I reckon this is intended to mean the AI has probable cause to deny people that could be nukies or cultists or traitors to the upload, since they will likely use it to relaw for harm. I don't think it's talking about the cap relawing the AI on a nukie round.

Sightld2 wrote: ↑Sat Mar 25, 2023 11:29 pm Yeah that's what I was saying earlier. The exceptions rules of "probably cause of harm" are very generous and allow for an Ai to let in or deny pretty much whoever in most rounds.

Pretty much what Sightld2 said. The original intent was that (unless the crew member has a very believable reason to go in there) the AI is allowed to say "fuck off" if there are known nuclear operatives, because it's not metagaming to ask why else the person would want to go in there if not to set the nukies to be nonhuman

(The AI, mind you, is not presently required to deny access over that - whether this encourages validhunting or not is up to you to decide how you feel, not me to tell you).

The situations laid out in SiliPol are mostly examples saying "If you don't want to become the crew's validhunting bitch, you are always allowed to tell them to fuck off here"

Tegun wrote: ↑Mon Mar 27, 2023 10:41 am

iamgoofball wrote: ↑Sun Mar 26, 2023 2:23 pm
no, they exist because that's what we want the mechanic to work like from the codebase

Isn't that the case with antag testing as well then? Because surely that isn't just something that was coded to work well enough and never given a second thought.

CMDR_Gungnir wrote: ↑Mon Mar 27, 2023 3:19 am

Tegun wrote: ↑Sun Mar 26, 2023 12:30 pm AI is a tool for the crew

Opinion rejected. The AI is a neutral third party.

(...) An awful lot of antagonists start their lives on the crew manifest. (...)

A fair point, though I'd hardly still call them "crew" no? And when one of the biggest issues with SiliPolicy at the moment is the amount of people who think the AI should be crew-sided, you can understand the wariness, I hope.

killing people who purged you as ai is funny

It's very unlikely we'd rule on this thread individually - any changes are likely to be a part of broader Silipol considerations. It is being archived in favour of a Silipol Megathread to make better progress towards a refreshed Silicon Policy.

View the megathread at:
viewtopic.php?f=33&t=34109