Thread Rating:
  • 2 Vote(s) - 5 Average
  • 1
  • 2
  • 3
  • 4
  • 5
Suggestion: AI/DWAINE Interaction Micronerfs
#9
(02-08-2024, 09:29 AM)Karakoran Wrote: I have spent a lot of time recently learning DWAINE, doing everything from teleman scripts to a DWAINE game to little prank scripts, and I am eager to continue. Most of DWAINE is quite innocuous, but I have seen it abused for evil a few times (confession: by myself). As a rogue AI, you can quite easily use teleman to obliterate the crew. This can be done maliciously with teletraps, but even more standard rogue AI behavior can get quite effective. For instance, you can clandestinely teleport all of the plasma cans from toxins to each air hookup on the map. However, it's worth remembering that this is really only doable as the AI. Silicons and humans find it difficult to get coordinates rapidly enough, even if they resort to using 3rd party programs (yuck) like online maps with coordinates. Thus, I think the appropriate solution is to try and nerf the AI's use of the mainframe, albeit only modestly.

Suggestion: Require the AI to be screwed down above a data terminal in order to access the Mainframe. The data terminal must itself also be connected to the mainframe.

These changes offers several new ways to shut down a tele-rampage AI:
1) Cutting power to the AI core will depower the data terminal (I think) and stop its mainframe access, and power can be cut in a number of ways.
2) Cutting off the wiring between AI core and Mainframe
3) Targeting the data terminal itself and using packets to remotely shut it off (highly theoretical)
4) Unscrewing the AI core

There are a number of other ways to fight the tele-rampage, but most of them aren't particularly effective or are very time-consuming. This would give the quick and easy strategy of just cutting mainframe access. Antagonists would also now have the option of sabotaging the AI this way, but it would be easy for another player (or AI shell) to repair once the antag left. Conversely, rogue AIs will have to negotiate for assistance from the antags and/or repair in spite of crew efforts to stop it.

-

The situation also presents a second problem. The game design surrounding the AI core is meant to be a sort of lair for the AI to defend, giving a fixed position for crew to rally at and fix. When the robotics control and law rack are gone/can't be replaced, the solution is meant to be killing the AI itself. Silicons were recently-ish reworked such that they can't interact with mainframe elements on a different Z-level. While this solves the nightmare scenario of an AI teleporting itself into the Debris Field and wrecking havoc with impunity and answered some interaction questions across Z-levels for things like the law rack, it doesn't solve the fundamental problem of a clandestine AI running amuck.



Suggestion2: Add some mechanics to the Mainframe_AI, allowing DWAINE users to ping it for location data and/or which data terminal it is using.

This would allow anyone with a working DWAINE terminal to easily locate the AI. I'm especially interested in how this would give Security a new tool in combating rogue AIs, as well as incentive to learn DWAINE - a game mechanic that Sec nerds might otherwise not touch. I think it would also be nice to add functionality for Mainframe_AI beyond logging in and calling the AI a nerd (>90% of current usage).

This second change will force the AI to either move rapidly to many different locations (dangerous + time-consuming, especially if it must screw in as in Sug 1) or remain in a single, fortified location. It would be preferable if this location data was only accessible when the AI is wired into the mainframe, but being accessible at all times shouldn't be a serious problem. 

-

Sadly, I don't think these changes can really be directly implemented at the moment. A handful of maps (Donut 2, Mushroom) have AI Satellites that are disconnected from the mainframe. Still, I feel it's the best way to address tele-rampage AIs without disrupting DWAINE. Hopefully it will get you guys thinking of better solutions!

Oh, and lastly, I urge people to avoid the temptation to call for nerfing the 'teleman' program itself (or worse, nerfing DWAINE, God forbid) and instead nerf things peripheral to it. 
Teleman remains one of the most engaging aspects of DWAINE and is a great learning step for would-be DWAINE nerds. It's how I really learned it. LMK if any of you want help learning, btw.
I believe we should also adopt a culture that frowns upon silicons using lethal teletraps. I've shifted towards kidnapping just a handful of victims of torturing them through various means (all much more interesting and engaging, I hope).
And teleman offers many, many, many benefits to the AI and station as a whole. You can use it to quickly scrape for artifacts, can gather dead bodies for cloning instantly, can give new players lots of expedient help (taking them where they want to go + delivering special requests), and so on. It would be quite sad to lose it entirely.

As a person who got screwed up by AI tele rampage I can tell that this one is good micronerf but the true problem is there is like 10-20% of all station quite familiar with DWAINE and only few of them are expert which mean only expert player can use this tool well.So I agree with you but this will only effect few of expert player who master in Teleman and Dwaine
Reply


Messages In This Thread
RE: Suggestion: AI/DWAINE Interaction Micronerfs - by meaow589 - 02-27-2024, 04:58 AM

Forum Jump:


Users browsing this thread: 1 Guest(s)