 Agent Smith Problem, The problem and how to solve it |
  |
 Feb 5 2007, 06:16 PM
Post
#76
|
|
|
Prime Runner  Group: Banned Posts: 3,732 Joined: 1-September 05 From: Prague, Czech Republic Member No.: 7,665  |
It wouldn't even castrate them, or prevent them from breaking the universe. If each node can have an agent on it, and each agent can run a search or medic program, you still have node proliferation where Hackers will have copies of their icon present on several friendly nodes that will heal them and issue batch commands to hundreds of nodes that all go out and run searches.
The basic Agent Smith problem exists as long as there is a 1:1 correspondance between having a new node and having a new agent, whether the agent can leave the node or not. -Frank |
 |
 
|
|
Feb 5 2007, 06:52 PM
Post
#77
|
|||||
|
Moving Target Group: Members Posts: 745 Joined: 2-January 07 From: Los Angeles, CA Member No.: 10,510 |
How is having a thousand drones scouring and searching the Matrix different from using a Browse program? The only reason the agent problem exists is because agents have been singled out to "work differently" than other operations. I see no problem in getting rid of them.
I see no reason why one can't treat IC as a general exception if you want IC in the game. In older editions of Shadowrun (Virtual Realities 2.0) there were IC, but there weren't hacker-agents and deckers couldn't put IC in their deck. This didn't unbalance the game in anyway. And in general I think having different rules for "security systems" and "comlinks" is a good idea. |
||||
|
|
|
||||
|
Feb 5 2007, 07:04 PM
Post
#78
|
|
|
Hoppelhäschen 5000 Group: Members Posts: 5,807 Joined: 3-January 04 Member No.: 5,951 |
That's because decks were completly unhackable.
BTW, Frank - read the Errata, there's a little catch on the Medic trick. |
|
|
|
|
Feb 5 2007, 07:35 PM
Post
#79
|
|||||
|
Immortal Elf Group: Members Posts: 11,410 Joined: 1-October 03 From: Pittsburgh Member No.: 5,670 |
the whole point of agents is to work independently of the decker. game balance is not the only factor, here. if you have computer programs that are able to independently direct vehicles in combat, it logically follows that someone could put together a computer program that is able to interact with the Matrix independently. simply getting rid of agents is not a solution, because agents are neat. there are problems with magic, too--maybe that whole part of the game should be cut out?
i don't. i think that rules which make sense--which basically necessitates a large degree of similarity between security systems and commlinks--are very important. (and before anybody makes any jabs about SR4, let me point out that the rules being discussed are basically mechanic-independent. they could be applied to SR3, d20, or BESM with equal ease.) |
||||
|
|
|
||||
|
Feb 5 2007, 08:10 PM
Post
#80
|
|||||
|
Moving Target Group: Members Posts: 745 Joined: 2-January 07 From: Los Angeles, CA Member No.: 10,510 |
Suit yourself. The idea that a corporation's entire IT infrastructure is all run off some exec's comlink is utterly ridiculous to me. |
||||
|
|
|
||||
|
Feb 5 2007, 09:03 PM
Post
#81
|
|
|
Horror Group: Members Posts: 5,322 Joined: 15-June 05 From: BumFuck, New Jersey Member No.: 7,445 |
Is it now?
What if said comlink is about the size of his desk and has logarythmically processing power than modern-day NASA, and the handset he uses is, while powerful, more-or-less merely a node. His desk is capable of running a crap-ton of Agents, but his handset can only coordinate so many. And if you want his desk to be more powerful, then do it! The rules provide very handily for expansion for numbers higher than 6 - just use numbers higher than 6. |
|
|
|
|
Feb 5 2007, 09:41 PM
Post
#82
|
|||||||||||
|
Moving Target Group: Members Posts: 745 Joined: 2-January 07 From: Los Angeles, CA Member No.: 10,510 |
You just reinforced my point. The desk has different capabilities than the handset. We shouldn't be using the exact same rules to hack into the desk as to the handset; the desk should get perks. In short, the desk should get IC and other agents and numbers higher than 6 and the handset shouldn't. These perks are the advantages of being a powerful corporation, government, policlub, criminal underground, etc. |
||||||||||
|
|
|
||||||||||
|
Feb 5 2007, 10:06 PM
Post
#83
|
|
|
Immortal Elf Group: Members Posts: 11,410 Joined: 1-October 03 From: Pittsburgh Member No.: 5,670 |
yes, you should be using the same rules, because they are both computers which interact with the Matrix in the same manner. should the desk-sized server have bigger numbers? yes, absolutely. but saying it should use different rules because it's bigger makes no sense at all. do you use different rules for shooting at a dragon than for shooting at a dwarf?
|
|
|
|
|
Feb 5 2007, 10:10 PM
Post
#84
|
|||
|
Moving Target Group: Members Posts: 745 Joined: 2-January 07 From: Los Angeles, CA Member No.: 10,510 |
I use an upgraded sniper rifle to shoot at the dragon. Tell me, is the dragon going to use an upgraded sniper rifle against me? Just because they are both entities wanting to do damage to eachother doesn't mean they should have the same weapons and tactics. An interlinked network of powerful desk-sized corporate servers could convievably have IC programs (or other Agents) that can't be installed on a comlink. |
||
|
|
|
||
|
Feb 5 2007, 10:18 PM
Post
#85
|
|
|
Immortal Elf Group: Members Posts: 11,410 Joined: 1-October 03 From: Pittsburgh Member No.: 5,670 |
no. but if he did, he'd use the same rules as you. and if he decides to eat you, he uses the same melee rules. more/better agents and IC != different rules. if that's what you mean by "different" rules, then we basically agree. if not--if you mean "get rid of agents except on servers", as indicated by your previous statements, then we disagree.
there's no reason to get rid of agents. they're cool, and a few patches can make them non-broken. getting rid of them will cause more problems than it will solve, because agents are a logical extension of technology that has existed in the game since SR1. getting rid of them requires ignoring the fact that any character with sufficient programming skills would invent them if they didn't already exist. for what it's worth, i believe that Unwired will have security options that commlinks will have difficulty running. i could be mistaken; i don't have any insider information. but it seems logical. then again, variable TNs seem logical... |
|
|
|
|
Feb 6 2007, 03:31 AM
Post
#86
|
|
|
Runner Group: Members Posts: 3,009 Joined: 25-September 06 From: Paris, France Member No.: 9,466 |
Is it logical that "big" systems should be more powerful than small ones ? Not necessarily.
2070s commlinks are totally different from today's computer, and size may not be a limiting factor anymore... That's why I feel comfortable with the idea that rating 6 is as high as you can get (generally speaking) and that the only difference between a commlink and a "host" (which I'll consider as a "node cluster") is that the host can handle much more connections at the same time (no subscription limit) and that each user connected to it can run as many programs as if he was all alone on a same rating commlink. |
|
|
|
|
Feb 6 2007, 04:46 AM
Post
#87
|
|
|
Moving Target Group: Members Posts: 870 Joined: 2-October 06 From: Athens Ga Member No.: 9,517 |
mfb WTF? How do you keep on going from a drone can drive itself around to an agent moving independantly in the matrix? They are not remotely (pun intented) related actions.
A drone has a pilot program which is staying in its OS (or functioning as it OS) and directing the vehicle it is in to move based on commands you are giving it or its own programming. At no point is the pilot program LEAVING the drone. It doesn't cause a problem because there a logistical and monitary issues that prevent someone from generating a thousand drones. Agents are a completely different issue. The problem is that you can make as many copies of them as you want to and they can be sent into the matrix like a free army. This is a serious problem. In game there doesn't seem to be a need to allow this kind of thing for hacking to work. It can work just fine without it and even with the messed up RAW it ISNT necessary. Out of game it is even worse of a problem because a player that figures this out could totally destroy the game balence. They could take down the city grids. They could blackmail large megacorps. THey could do a lot of hacking terrorism. They could bog down the game with things that are not the basis of the shadowrun setting. The thing that you are missing is that these are normal access programs that everyone can have. If everyone could do this then it would shut down the matrix. If just those who could break the copy right coudl do this it would shut down the matrix. You can't just think about what your character wants to do with this. You have to think about the consequences that a thousand people will do with this. If you don't think that this is an issue then let a magic user in your game create no limit high force rating spirits that don't go away every few minutes for a day. Once they have a thousand then let them go about their buisness for a while and come back and tell me there is no problem. I hear people on this forum talk about how to get around magical security because it isnt fair for a mundane shadowrunner to have to deal with. Yet allowing such a high number of agents could result in a thousand IC standing around holding VR boards pretending to be the walls. These wall can be see to be transparent from their point of view. They all can be linked so they all know the same things. They will always see you but you only see the walls. I'm not saying get rid of agents or pilots or IC. I just saying that if they were a function of your OS, like a combination between your browser, your help center, and a VR character then you have no problems. They should'nt be useful outside of your computer and they shouldn't be used as proxy hackers. |
|
|
|
|
Feb 6 2007, 06:19 AM
Post
#88
|
|||
|
Immortal Elf Group: Members Posts: 11,410 Joined: 1-October 03 From: Pittsburgh Member No.: 5,670 |
they are very closely related. basically, they are both making decisions based on external stimuli. if you have a program that can do that, it can do it anywhere. so if you have programs that can drive cars, it logically follows that you can create programs that surf the Matrix. if you can pass orders to the program that controls a drone, you can pass orders to an agent in a remote commlink. if you can exploit a security loophole that allows you to illegally logon to a host, you can create program that does it automatically. i agree that allowing effectively unlimited agents can be problematic. however, the nature of electronic data and networking makes it nearly impossible to logically limit the maximum number of autonomous programs a user can have working for him. so instead, i think any attempt at solving the problems inherent in unlimited agents should focus on solutions which limit or alter the effects of having that many agents, so that it doesn't break the game or suspension of disbelief. FrankTrollman's solution, for instance, neatly solves almost every problem i've seen mentioned. |
||
|
|
|
||
|
Feb 6 2007, 06:56 AM
Post
#89
|
|||||
|
Moving Target Group: Members Posts: 745 Joined: 2-January 07 From: Los Angeles, CA Member No.: 10,510 |
I would suggest this be a good time to start exploring the difference between Artificial Intelligence Matrix entities (which are as capable as any metahuman, if not better) and Drones (which, at their best, are always described as dog-brain stupid). I'm not sure on the netiquette of pulling information from other threads... but in a previous discussion the IC Agents of SR4 were compared to Norton Utilities scanning every individual file in the system, as opposed to N.I.C.O.L.E., the cheapest date ever. |
||||
|
|
|
||||
|
Feb 6 2007, 07:23 AM
Post
#90
|
|
|
Moving Target Group: Members Posts: 870 Joined: 2-October 06 From: Athens Ga Member No.: 9,517 |
mfb I'm not sure if you are missing the crux of the problem or not. It isn't whether they can surf the internet. It's that they don't need to leave the commlink to do this anymore then you have to climb into the internet to find something. The matrix is not a real space. It is the effect of visualisations. It doesn't need to climb around or move from server to server to find things. It needs to access a series of data mining systems. These data mining systems have a lot of information. Your browse program sorts through these based on your preferences and wishs. At no point would you want your agent actually going into thousands of servers, ignoring it's database, and creating one of it's own for each place.
The agent doesn't need to move anywhere to do it's job. It's the idea that it does move around the matrix that is causing the problem. If that actually were the case then there would be no problem in stopping them. All you would have to do is have the server not run that kind of file type. It will accept a file request from a browser for a web page it has or even one with a VR dialog but it won't accept anything self executable or having the file types relating to agents. Problem solved. Since there seems to be a lot of connection betwen agents and IC then the last thing that a server would allow was something that can do all the damage of a virus but can act semi intelligently in your system, especially if it can create more versions of itself. If they can be usefully copied and actually act independantly you are talking about super viruses, not data miners. |
|
|
|
|
Feb 6 2007, 07:54 AM
Post
#91
|
|||
|
Immortal Elf Group: Members Posts: 11,410 Joined: 1-October 03 From: Pittsburgh Member No.: 5,670 |
you're assuming that the filetypes are a) detected by the server for what they are, and b) can be automatically differentiated from the filetypes used by legitimate icons. the second point, at least, is provably untrue, because one of the pieces of information you need to roll a Matrix perception test to learn is the type of object you're looking at. as for the first point, well, what do you think the agent is doing when it rolls Exploit to get access to the node? it's disguising itself as something else. the node isn't even aware the agent is there, if the agent does things correctly. re: leaving its node, well, a decker doesn't really go anywhere, either. their icon represents the far end of their connection to the remote host. to the extent that a decker 'goes' anywhere when he surfs the Matrix, an agent can also 'go' places. |
||
|
|
|
||
|
Feb 6 2007, 08:24 AM
Post
#92
|
|
|
Moving Target Group: Members Posts: 870 Joined: 2-October 06 From: Athens Ga Member No.: 9,517 |
see I think I see the problem. You are assuming that the RAW makes sense which it doesn't. You can't have a file on a system that the system can't see. There is a difference between not setting off alarms from the security programs and actually not showing up on the system. Interacting with the system is how things run. You keep on assuming that programs can have any kind of reality outside of the systems they run on. If the system can't run it then that is the end of it.
The basis for making any kind of logical arguement about a real world concept is that it is based on a real world concept. You can model all of the ideas behind the Matix in shadowrun as long as you understand that the details the game gives you are non sense. Most of it is a legacy from earlier non sense. It sometimes gives you the outward appearence of logic but then says that it is powered by turtles. If you don't want to worry about any reality to these computer systems then there is no problem. Agents don't do this because they have a union and it is not in their contract. Pilots can fly because they were in the Air Force. IC is not a problem because the IC mutant powers don't work away from their radioactive rocks. |
|
|
|
|
Feb 6 2007, 08:45 AM
Post
#93
|
|||
|
Hoppelhäschen 5000 Group: Members Posts: 5,807 Joined: 3-January 04 Member No.: 5,951 |
Of course you can. Just play an 'old' Sony-BMG 'CD'... rootkit included. |
||
|
|
|
||
|
Feb 6 2007, 09:03 AM
Post
#94
|
|||||||
|
Moving Target Group: Members Posts: 745 Joined: 2-January 07 From: Los Angeles, CA Member No.: 10,510 |
I put up the stuff on rats so I could contribute them to this discussion. Basically, most AIs used in my custom hacking rules fall into the category of ICE or Viruses, and "rats" represent the lowest wrung of viruses. They are commonly available for purchase and used for all sorts of stuff. Here's the Info for rats: The Rat Infestation ...and for everything else: Cetiah's Custom Hacking Rules (under construction) |
||||||
|
|
|
||||||
|
Feb 6 2007, 09:43 AM
Post
#95
|
|||
|
Target Group: Members Posts: 87 Joined: 16-December 06 From: Kabul, Afg. Member No.: 10,381 |
Wondering if there is some way to simply use these three options with a single roll? Perhaps by giving a different threshold for each level: 2 hits Compromised, 4 hits-Alert, 6 hits-shutdown, modded as to how big/powerful the system is. You could still use Influence (as proposed by Frank) as more of a die pool modifier. The hacker/decker uses a round to "turn on" his dispersed network, next round giving it instructions, makes a die roll Hacking+?? (whatever fits best) to see how well he coded/instructed his dispersed Army of Agents, count the hits and you know if your DoS attack worked or not. He can then concentrate on whatever kind of other hacker nastiness is more pressing to the rest of the team. No new rules, just a little creative "reapplication" :D maybe I missed something? |
||
|
|
|
||
|
Feb 6 2007, 10:28 AM
Post
#96
|
|
|
Running Target Group: Members Posts: 1,498 Joined: 4-August 05 From: ADL Member No.: 7,534 |
Wouldnt the agent smith problem be solved, if there was a way to limit the number hostile agents in (or accessing) a node?
Adiddionally: Could somebody post a list of matrix actions that benefit from agent smith? |
|
|
|
|
Feb 6 2007, 02:14 PM
Post
#97
|
|||
|
Moving Target Group: Members Posts: 745 Joined: 2-January 07 From: Los Angeles, CA Member No.: 10,510 |
I think it might be a lot simpler just to say Agents have no Hacking skill, only Computer. |
||
|
|
|
||
|
Feb 6 2007, 03:27 PM
Post
#98
|
|
|
Immortal Elf Group: Members Posts: 11,410 Joined: 1-October 03 From: Pittsburgh Member No.: 5,670 |
Garrowolf, for pete's sake, read what i say. i don't think the RAW makes sense. if i did, i wouldn't be looking for ways to fix it. i didn't say the system can't see the files that the agent uses to run on that system, i said it doesn't necessarily see them for what they are. yes, the system can see the files that comprise the agent's presence on that system. no, that doesn't mean the system recognizes those files as comprising an agent. i'm not assuming that "programs can have any kind of reality outside of the systems they run on", or anything even remotely similar. i'm assuming--correctly, because it's the entire basis of hacking, both in real life and in SR--that systems can be fooled into running things that they wouldn't otherwise run. this includes, in SR, hacker icons and agents.
you're assuming that there's some sort of difference between how a user interacts with systems and how agents interact with systems. there's no difference. a system can't differentiate between an instruction sent by a user and an agent. to the system, it's just an instruction--the only thing the system can detect is if the request is legitimate or not (and if the hacker/agent does his/it's job right, the system can't even detect that.) |
|
|
|
|
Feb 6 2007, 03:54 PM
Post
#99
|
|||
|
Runner Group: Members Posts: 3,009 Joined: 25-September 06 From: Paris, France Member No.: 9,466 |
You mean something like the limitation of running programs by RAW ? :| |
||
|
|
|
||
|
Feb 6 2007, 05:32 PM
Post
#100
|
|
|
Running Target Group: Members Posts: 1,498 Joined: 4-August 05 From: ADL Member No.: 7,534 |
No, because this is
a) too restrictive, matrix cafes and databases would never work and is thus nonsense b) not limiting enough, it still allows for enough agents to severely limit the impact of the hacker |
|
|
|
|
|
Lo-Fi Version | Time is now: 17th June 2026 - 07:09 AM |
Topps, Inc has sole ownership of the names, logo, artwork, marks, photographs, sounds, audio, video and/or any proprietary material used in connection with the game Shadowrun. Topps, Inc has granted permission to the Dumpshock Forums to use such names, logos, artwork, marks and/or any proprietary materials for promotional and informational purposes on its website but does not endorse, and is not affiliated with the Dumpshock Forums in any official capacity whatsoever.