Printable Version of Topic

Posted by: Aria Jun 8 2011, 12:42 PM

There have been a couple of AI threads recently that have sparked my interest but raised some questions...

1. How neccessary are the inate programs? I only ask as eGhost seems to fit my concept well but it's only a 15BP boost for potentially loosing out some of the key AI stuff...

2. If an AI's node is a commlink can they make use of response enhancers and simsense accelerators? Simsense booster cyber is obviously out but it strikes me as odd that an AI might be slower than a meat hacker in the matrix?!?

3. Can an AI recieve SIMsense info? In this case I'm thinking about 'vehicle touch sensors' although probably without any rules bonus as they can't 'jump in'

4. With the assumption that modded vehicles don't have to look like modded motorbikes etc but can look like a 'human' walker (I know people don't all subscribe to this idea!) would this be reasonable?:

Emerging Futures (NeoNet) Avatar drone [designed for full emmersion hackers/riggers and AI's to interact with the world]

• Suzuki Mirage (rather than Contrail so that I have to overmod - not above 10 slots though - to get a cost semi comparable with a full cybernetic body) 6500¥
• Walker 3000¥ 2 slots
• Full Mechanical Arm x2 8000¥ 4 slots
• Overmod: Mimic R:3 15000 to 30000¥ 2 slots
• Overmod: Rigger Adaption 2500 to 5000¥ 1 slot
• Overmod: Touch Sensors 3000 to 6000¥ 1 slot
• Total Cost (without any cyber enhancements to arms) 58500, round to 60000¥
• Custom Agility +3 to both arms 9000¥ (I assume they start at 3 as it's not specifically mentioned)
• Right arm commlink
• Left arm sat uplink
• End Cost 70,000¥ +any sensor upgrades +commlink cost

Is there anything obvious I've missed or made a mistake on? 6 Agi and Str seems reasonable although some work might need to be done to reduce speed to something closer to the other cyborgs (?). As a GM I'd be happy with this but would anyone else?

Thanks!

A

Posted by: Ghost_in_the_System Jun 8 2011, 04:46 PM

Alrighty, lets see...

1) Not really that important, they are essentially a number of free programs with the ergonomic and optimize options.

2) Response Enhnacers are basically the equivalent of adding a coprocessor from what I read, so I see those working just fine for an AI (and are able to work on any electronic device, not just a commlink). The simsense accelerator greatly relies on the answer to 3)

3) My understanding has been that they can receive simsense info, but don't experience it. They would get for example the information that they feel a light touch on a back fender, but wouldn't actually feel it as a metahuman would. This means simsense accelerator wouldn't really help wouldn't help out much

4) I think you're better off modding a manservant or simply designing your own drone outright (Something between a manservant and the 150k one)instead of trying to mod a motorcycle. If however you're stuck on a motorcycle, it seems like what you have works, and at 70k isn't cheap nor exceptionally powerful. With all the overmod it is likely going to end up looking like an ork or troll though.

Posted by: Rubic Jun 8 2011, 09:31 PM

I've been working on an AI build, myself, and mechanically speaking, the Piloting Origin seems to handle vehicle sensors as well as anything else would. As far as simply knowing the vehicle was being touched, and where, they should be able to do that anyways, so long as they are subscribed to it.

Posted by: PoliteMan Jun 9 2011, 04:46 AM

Huh, the way I've read it, innate programs are critical. AIs can't take ware like hackers and can't thread like technos, what lets them keep pace is a few awesome qualities and the ability to improve their innate programs with Karma up to twice their System rating. Without those innate programs, well, AI's just seem to hit a hard wall once they have R6 programs and max skills. They're just done.


Can't see any reason why not.

Posted by: Ghost_in_the_System Jun 9 2011, 04:56 AM

I suppose that depends on the availability of 7+ programs in the campaign. I'd forgotten about the advancement aspect of inherent programs.

I'd consider talking with the GM who might allow an AI to more easily code programs (restricted to itself only especially) of 7+ fairly easily and at 2-3x normal speeds due to a lack of need to sleep.

Edit: Just to clarify my new position is - inherent programs are very important unless your GM is willing to be very compensating.

Posted by: PoliteMan Jun 9 2011, 05:15 AM

Ehh, War! already has pretty clear rules on R7-10 programs (p. 144). They degrade by 1 point every 2 months (or sessions) and can't be patched. Don't get me wrong, AIs that code are awesome but they're gonna be constantly chasing that SOTA curve. Inherent programs get around that, plus they can get higher. They just eat karma.

Posted by: Udoshi Jun 9 2011, 05:44 AM

Emphasis mine.

Intuitive Hacking(unwired): Emulate fixes this bit real quick.

See here for some http://forums.dumpshock.com/index.php?showtopic=30389&st=25&p=907429&hl=spawn&#entry907429


Emulate and Piloting Origin is a nice combination together. If your GM okays it, see if he's willing to let AI's use Processor limit to sustain various Emulations - it brings AI's back to an more sinister 'must spread out consume resources to the limit of my ability' feel.

Posted by: Ghost_in_the_System Jun 9 2011, 05:56 AM

Why do you need intuitive hacking?

And emulate isn't bad, but you're unlikely to get more than 6 hits on a software + rating test. It doesn't say anything about being able to boost existing programs, though that is up for debate. If it can be used on existing programs then it is quite exceptional. Otherwise it just saves a bit of money/time.

Oh, wait, I see what it is. That's shaky at best. As much as some stuff might 'technically' (debatable) be allowed in the rules, that doesn't mean a GM won't turn it down faster than you can say "It says on page..."

Posted by: Udoshi Jun 9 2011, 06:19 AM

Because, despite it being an AI only quality, they didn't give it a cost in RC.

The point you're missing is that Emulate lets you make programs from scratch. including very expensive autosofts, which are capped at rating 4, and 4 hits is a lot easier to manage.


Like any character, its something to talk over with your GM. Just saying, if your GM lacks a creative spine, and shoots down anything that isn't Exactly By The Book without looking at it, well..... then that sucks for you.

That being said, i've seen it used in a game, and its actually not really imbalanced. It removes most of the power abuse of threading - without sprites to boost complex forms to ridiculous levels, being unable to improve inherent programs with it, and not being able to do a thing about the -2 sustaining, all the potential abuse is pretty much gone, and its more a nifty trick to have in your toolbox than anything gamebreaking.

Posted by: PoliteMan Jun 9 2011, 06:52 AM

Hmm....I think that one fails the RAW test.

Sharing inherent programs and spawning agents (p.169) require you to have these qualities ("An AI who has both this and the Emulate quality..."). However else you might be able to get the abilities, you still don't have those those qualities, so you can't share inherent programs or spawn the right agents.

Posted by: Aria Jun 9 2011, 08:14 AM

Hmmm...thanks for the input!

Inherant progs sound better than I thought...maybe I'll go with 'pilot origin' as a rule base to achieve something more like the Eclipse Phase AIs but IC be an eGhost (Can't see why a GM would have a problem with this as it's a roleplay tool...?!?)

Udoshi's emulate 'trick' is good too, and not overpowered... The malware and viruses stuff is a bit beyond my current understanding of the hacking rules though

The motorbike mod was an attempt to find middle ground between the manservant and the cyborgs (I suspect their availability is linked to that rather than their 'drone' stats). Is the manservant being crap at combat fluff or written into the rules (can't actually remember at the moment )...some mods to the cyber arms for Str and Agi might go some way to offsetting this?

Posted by: Irion Jun 9 2011, 10:28 AM

One question about the AI lifestyle:
Could you just live in a commlink? I mean it is much cheaper to actually buy a rating 6 system than to pay for the lifestyle.

Posted by: Aerospider Jun 9 2011, 12:00 PM

Living in a commlink would be no different to living in any other kind of node. Paying for the lifestyle represents the continual maintenance required to prevent the node from degrading under the strain of supporting an AI and not doing so will see the node's ratings drop over time.

Posted by: Irion Jun 9 2011, 12:16 PM

I do not know. It just strikes me as silly expensive.
And you would probably need to buy a second lifestyle too, if you want some place to stash your stuff.
Why not just rule it in one.
Why not just saying everything included in one lifestyle.
You get the normal benefits, but instead of food and drinks and other stuff you do not need your node is maintained?

Posted by: Fortinbras Jun 9 2011, 01:16 PM

As an AI is not limited by it's home node's attributes, it shouldn't gain benefits from it's home node's attributes either. AI's run autonomously of their home node.
Moreover, a simsense accelerator only increases the IPs of a persona attached to the commlink, whereas an AI is not a persona, but an icon.

I'd say talk to your GM about letting you have the Technovantage power which would allow you to get Overclocking and Advanced Overclocking as per the Echo.

You might also consider a bio-drone for your person you will be walking around as.

Posted by: Aerospider Jun 9 2011, 01:53 PM

Expensive maybe, but we're in the realm of science fiction here so it's pure arbitration for game balance on the writer's part. After all, AI PCs tend not to have as much need for money as the average runner (no augmentations, health care, foci, transportation, etc.).

Posted by: Rubic Jun 9 2011, 03:06 PM

although heavy-handed, there is some basis in reality. If you leave your computer on constantly, not only will you use electricity, but such components as the cooling fan, RAM, and processor will burn out faster. Granted, in our day today, that takes months. Flash over to 2072, where computers have much more power behind them, and the entire world is populated with products founded on the ideal of planned obsolescence. The corps want it to break as soon as they have some shiny new toys that just devalued what you bought, after all.

If you're really THAT concerned about cutting costs, there's another option, presented in all the core books (at least for 4th ed). Share a lifestyle with another character. That's a 10% cost for your server/drone's housing, upkeep, and routine upgrades.

Posted by: Ghost_in_the_System Jun 9 2011, 03:12 PM

Except that a metahuman's lifestyle is entirely different from an AIs 'lifestyle' and I really doubt they could be properly shared.

Posted by: Tymeaus Jalynsfein Jun 9 2011, 04:18 PM

Granularity... For some it matters, and for others it does not...

Posted by: Rubic Jun 10 2011, 07:39 AM

If you consider exactly what is being shared... the home-node. The added cost figures in what the AI needs or has access to (according to the AI lifestyle chart), and allows the AI to utilize any of the mundane drones or the home node as their home, covering the increased cost due to such high-volume processing.

I'm sure that more tenuous explanations have been pulled to cover other issues, and my explanation makes sense.

Posted by: PoliteMan Jun 10 2011, 07:48 AM

Huh? Hackers don't pay lifestyle costs, why would AI's? That's what agents spoofing lifestyles are for (or drones, if thinks get tetchy).

Posted by: Ghost_in_the_System Jun 10 2011, 12:28 PM

Except I doubt the 10k a month spent on the maintenance of a single device is really covered by the 10k a month spent on an entire lifestyle (Rent, food, cable, night clubs, eating out, clothing, transportation, water, etc). It's really a per game basis, but it isn't something I'd allow. Might allow the node being maintained at one level lower than the base lifestyle in this setup though.


Hehe, good point, though it is near impossible to get more than middle lifestyle. Requires a 20+ spoof DP and about two weeks to manage it.

Posted by: Tymeaus Jalynsfein Jun 10 2011, 01:23 PM

Funny, The Technomancer in our group could do it in 4-5 Days... My Cyberlogician could do it in just at 7. For a High Lifestyle (48 Successes). Admittedly, our Technomancer was throwing about 23 Dice, and I had about 16. Prior to the Change to the Optional Rules we now use. No Degradation. With Teamwork Tests. For a Middle (12 Successes), we could cover that in about 2-3 Rolls for me, and 1-2 rolls for himself. The Technomancer's Sprites could do it for him in 12 Days for a High Lifestyle, or Luxury (at 100 Successes Base) in 25 Days. With BOUGHT hits, rather than rolled hits.

Not hard to get Lifestyle Costs Covered as a Hacker/Technomancer. Which leaves all your earned Nuyen for other things.

Posted by: Ghost_in_the_System Jun 10 2011, 01:41 PM

Well, you are apparently very lucky on your rolls because even without diminishing DP, 23 dice should take 7 days and 23 dice on a spoof test is quite a bit.

But I'm talking about with diminishing DP which increases the time required quite a bit (especially at lower DPs), and makes sprites and agents very unlikely to be able to ever accomplish the task. Actually I suppose it isn't too bad, 9 days on average with a base 20 DP. Still, presumably a hacker is going to want to be programming and cracking and patching and so on. At some point their time is more valuable and limited of a resource than nuyen.

Posted by: Tymeaus Jalynsfein Jun 10 2011, 02:59 PM

The Technomancer is very lucky on his rolls... I only needed to do so a few times myself, but I can do it, even with Diminishing Dice Pools. Though it does extend that time a bit.

With the Change to the Optional Rule Set we use, it likely takes a bit longer, because the dice pools for the Technomancer suffered a bit (and Program/CF Caps Hits), but not too much.

As For Sprites... Yes, Diminishing Pools hurt them, though I think that a Sprite can do High, even with that rule... Resonance 6 (12 Dice) with 2 others to Aid. Base Pool for starting Teamwork is 18 Dice... Bought Hits run out at 3 Dice and 40 Successes... Rolled Averages allows success for High Lifestyle...

Posted by: Fortinbras Jun 10 2011, 08:29 PM

Then you need to calculate the time needed to register those other sprites.
Even still, a fourth of your time being spent to maintain a Lifestyle seems a little much for me. What's a technomancer need all that money for anyway? You hacker's and your highfalutin ways. You don't know what it's like on the streets, man!

As a GM I love the Roommate rule because it encourages all the character's to live together. It makes logistics in storytelling that much simpler and creates a stronger sense of camaraderie. If there were an AI on the team I'd let it be a part of the team's roommate pool and Spoof it's Lifestyle, but each of the qualities represent something differnt than, yet similar to, it's meat world counterpart. Neighborhood, for example, represents how dangerous the node is from roaming sprites, probing hackers, GOD searches and other AIs just wandering around cyberspace.
That is something just buying a commlink doesn't represent; the real dangers and aspects that come with living in such a complex entity as the Matrix full time.

Posted by: Tymeaus Jalynsfein Jun 10 2011, 08:52 PM

As much as I like this Idea, it has been like pulling teeth to get it to actually happen in our groups. The level of trust between characters is not nearly where it should be for stuff like that, even when several of the characters try to set this up.

Posted by: Ghost_in_the_System Jun 10 2011, 09:00 PM

Well, let the characters that set it up enjoy the benefits and the ones that don't want in go on paying for their own place.

Posted by: Tymeaus Jalynsfein Jun 10 2011, 09:07 PM

Which is ultimately what happens...

Posted by: Rubic Jun 10 2011, 09:57 PM

what's so bad about that? Sharing a lifestyle just means it'll only take one THOR shot to TPK.

Posted by: Tymeaus Jalynsfein Jun 10 2011, 10:01 PM

This is true, but if you have to resort to a Thor Shot to take care of a Shared Lifestyle, well, something is wrong.

Posted by: Fortinbras Jun 10 2011, 10:10 PM

Having all roommates home at the same time is a once in a blue moon situation. I shared a room with a guy in Brooklyn because when he came home from work, I was going to school.
The former military street sam is up at dawn for his morning run while the technomancer is just getting home from a rave. Shared space is almost never shared schedule.

Besides, a Thor shot is a GM tool. If I wanted TPK, I could make it happen regardless of the living situation. Ritual magic, snipers or car bombs could all do the same thing for less than 1.2 million. I'm not limited by any minor constraints such as shared location.
Not much of an issue as I'm not a jerk GM.

Posted by: Rubic Jun 11 2011, 02:47 AM

As a more-often player than GM, I appreciate that. I am, however, of the opinion that players are almost to a one, Jerks. We can smell our own (then again I'm nicer as a GM, except I just don't shy away from Darwin-Checks).

Posted by: Makki Jun 11 2011, 03:15 AM

I played an AI for a while. They are limited in many ways, but if you exploit their advantages, they will be awesome.

Innate programs:
-Stealth, double digits here after some karma, like TMs, but without the need of threading and sprites. Which both need skills and fading. An AI is at least invisible, that's an advantage.
-Command. It functions as your attribute when using remote control. Double digits attributes anyone?

Piloting Origin is obligatory: Don't waste karma and BP on skills, just get pirated Autosofts, you may now use them in place of your skills.

Posted by: Aria Jun 11 2011, 06:04 AM

Whilst I don't disagree on a mechanics front, I object a little bit to there being only 1 optimal configuration

Couple more questions:

1. If an AI is homed in a commlink when do you use those (boosted) stats like a hacker would, and when would you use the derived stats? I'm thinking of firewall in particular!

2. Can a jumped in rigger or AI use autosofts on that drone or only their own skills?

Thanks again

Posted by: Fortinbras Jun 11 2011, 06:38 AM

An AI's Response and Signal are always the same as the node it is in at the time, regardless of it's home node.
It's System and Firewall are independent of any node, including it's home node. It's System is (Logic + Intuition)/2. It's Firewall is (Charisma + Willpower)/2.
Near as I can tell an AI does not gain any benefits from it's home node other than the lack of degradation and Response & Signal; especially not a Simsense Accelerator as those only boost persona and not icons.


A rigger uses his own skills. Only an AI with the second level of Piloting Origin can use autosofts.

Posted by: Aerospider Jun 11 2011, 09:46 AM

Just to clarify - Response and Signal come from the node on which the AI is currently running, not the node in which his icon is being projected.

Posted by: Fortinbras Jun 11 2011, 10:01 AM

Sort of. If an AI is running in multiple nodes at once, it's Response and Signal are that of the node in which it is taking the action or in the node in which an action is being performed on it.
An AI does not project from any one node the way a persona does. If it is subscribed to a node, it is running on that node. It lives entirely within the Matrix and is not, therefore, confined to any one node. It can exist independently of it's home node for a week without degradation. If it's home node is destroyed, it can make a new one.
An AI affects it's home node far more than said node affects the AI.

Posted by: Aria Jun 11 2011, 12:10 PM

Think I follow all of that

It seems that until an AI has karma to burn then they aren't going to directly compete with a good hacker or TM (lower progs, less initiative), but as they can be more or less invisible with the Rootkit quality (-6 perception) then I guess toe to toe is not the way to go anyway...

It's just as well that I want to try an AI 'rigger' of sorts anyway

Posted by: Aria Jun 11 2011, 02:02 PM

Further to this, can something attack you in the matrix if you attack it but it can't see you (-6 to perception checks and stealth running for example)? ...best policy would be to skirt on by I know but still...?!?

Posted by: Teulisch Jun 11 2011, 02:18 PM

well, by RAW it seems it would still need to win an opposed test to see you if stealth is running. but unless you have a mute option on your attack or get a one shot kill, you did just trigger a general alert and a stronger IC response on that node. if multiple agents are looking for you then its harder to stay hidden.

it seems to me that the crash program option is often better than attack. as it gives you a simple opposed role, instead of opposed to hit followed by damage resistance with a condition monitor.

Posted by: Aerospider Jun 11 2011, 03:30 PM

Oh, see that's not how I've understood it at all. The way I read it the AI does only run on one node at once, but unlike regular personae it can move to run on a different node so long as it has admin privileges there (AFB but the text definitely mentions an admin requirement). Then, from whatever node it is running on it may have subscriptions with other nodes and thereby project its icon into them. If you say that the AI is running on all the nodes it has a subscription to, wouldn't that make it rather vulnerable?

Posted by: Fortinbras Jun 11 2011, 05:37 PM

I sort of get what you are saying here. You're referencing Runner's Companion p. 88.
An AI only exists in one node at a time, but creates icons in other nodes to which it has subscriptions. It does this the same way any regular Matrix user would.
When controlling it's icon, the AI is operating within the node it exists. It's Response and Signal are that of the node in which it is operating, not any node to which it has a subscription.
It can change nodes with a Complex action and admin access.

It's icon still vulnerable to Matrix damage the same way any icon is. I'm not sure what you mean by "projecting", but it is still vulnerable to detection, damage and tracking the same as any icon operating within that node would be.
An AI is more vulnerable than a hacker or a technomancer in that it is more easily traced, but it makes up for that vulnerability in spades by having no meat body and only being killable three ways.

Posted by: Aria Jun 12 2011, 05:08 PM

Following an AI gen thread someone said AIs don't get a hot sim bonus for jumped in so what is the point of lvl 1. of pilot origin as remote control would give better bonuses?

Posted by: Aerospider Jun 12 2011, 08:27 PM

The hot sim bonus applies to remote control as much as jumping in and an AI gets neither, so reasons for jumping in are the same as for fleshy riggers -

1 - The Command program slows you down because every action is a complex action, including free actions like changing a device mode. Jumping in allows you to run at full speed.

2 - Jumping in automatically blocks out all other commands to the drone/vehicle, legitimate or otherwise, including spoof attempts.

Plus, AIs don't have to worry about dumpshock.

Posted by: Ghost_in_the_System Jun 12 2011, 09:01 PM

It does? I can understand many commands being overridden quickly, but I don't remember anything ever saying that it blocked command and spoof attempts entirely. In fact, I'm fairly sure I remember seeing a command to kick off the rigger as being a good way to disrupt drones.

Posted by: Aerospider Jun 12 2011, 09:45 PM

SR4a p.246 under ELECTRONIC WARFARE - HACKING AND SPOOFING

"Such hijacking attempts against your drones can be foiled by jumping into a device. A jumped-in rigger overrides any other control of the drone, including by its Pilot."

Posted by: Ghost_in_the_System Jun 12 2011, 10:31 PM

Hmm, I'd still think you could try and disconnect the rigger, but I suppose not. Cool, thanks for that.

Posted by: Aerospider Jun 12 2011, 11:25 PM

Oh yeah, you can still mess with the rigger himself but control of the device is off-limits until you do.

Posted by: phlapjack77 Jun 13 2011, 10:06 AM

I think I know why you said this - DS occurs only if the user is in VR, and VR requires simsense, and AIs can't do simsense...can't find the reason AIs can't do simsense tho...


This makes sense - if you're jumped in, you ARE the vehicle, so of course you're going to ignore commands that don't originate from yourself (and you're likely to know if you just ordered the vehicle to power down)

Posted by: Fortinbras Jun 13 2011, 10:14 AM

They have no brains for the ASIST to interact with. They don't experience emotions chemically.

Posted by: Aria Jun 13 2011, 12:19 PM

Had another go at an 'Avatar' drone based on the Manservant. I'm trying for a little bit of robustness but looking human/humanoid

Manservant 2500¥
+Mimic [R:3] 15000¥
+Touch Sensors 1500¥
+Turbocharger [R:1] 900¥
+OM: Concealed Armour [R:6] 12000 (x2)¥

+Response Module [R:5] 4000¥ (AI Homenode bonus +response 5+weapon skill shouldn't be bad and makes some use of the pilot origin 1)

Total 36,000¥

Doesn't seem overpowered to me whilst still being useful. Additional extras might be cyberarm gyromounts etc for a bit more combat oumph!

Comments?

Posted by: phlapjack77 Jun 21 2011, 05:29 AM

I guess it seems that at the end of the day, ASIST is just a bunch of electrical impulses, right? Electrical "signals", so to speak? So instead of transmitting those signals to a brain or whatever, those signals are transmitted directly to the AI.

At some point there must be a hardware / software interface between the simsense signals and the brain, this interface translates the signals into something the brain understands. Shouldn't an AI be able to also do this translation, possibly much more intuitively?

Posted by: Aerospider Jun 21 2011, 11:30 AM

The key factor here is that AIs do not have a brain. Or at least, not a biological one. They understand computer language as it is, no need for translation. When you're in VR and you see the icon of an AI moving around and interracting with the sculpting, it isn't experiencing the environment as you do in that it doesn't see, feel, hear, taste or smell the iconography. All your sensory input when you go VR is just the result of your sim module translating computer code to make it intuitively understandable by drawing parrallels with the meat world (in accordance with the sculptor's design). ASIST is for people because they aren't AIs.

So in answer to your question, I doubt that an AI would be able to translate a language it understands (computer code) into a language it doesn't (sensory information) in real time more intuitively than a piece of hardware designed for the purpose (sim module) or that one would ever want to.

Posted by: Aria Jun 21 2011, 12:45 PM

Unless it's an eGhost? Presumably they experience the matrix as they would always have done and their code supports this?!?

Posted by: phlapjack77 Jun 21 2011, 01:00 PM

I see your point, and I'll concede it.

But it still doesn't sit well with me that a human receiving electronic impulses can benefit from the impulses more than an AI could, a being that lives in the world of pure electronic impulses. But I guess it's just a game and all that...

*edit*

I agree totally. People need assistance to translate these electrical impulses into something they understand. AI are able to understand it natively. So why do people get the simsense bonuses using the matrix, but AIs can't?

Sorry, this idea is now whirling around in my head. Feel free to ignore all this blather forthcoming...

So "simsense" is just simulated, or recorded, sensorium. Essentially all it's doing is bypassing the real-world creation of each sensation. In the real world, there is an apple. The photons from the apple strike your eyes, and your eyes transfer this signal to your brain, and you see an apple. In simsense, there are no photons, it's just the straight signal dumped right into your brain, no sensors (eyes) needed. In both cases, you see an apple ultimately because there is a signal sent to your brain that you see an apple.

If AIs can receive signals from various Sensors(routed through drones or vehicles or something), that means they're able to receive signals about the real-world just like a person does. Similarly, it shouldn't matter where these signals originate from. If a Sensor sees an apple, the AI in the drone will see an apple. Why then can't the AI receive the same signal from simsense, bypassing the need for a Sensor? There would need to be a simsense translation package bought for the AI maybe or something, but otherwise...

Thanks for listening..err, reading

Posted by: Fortinbras Jun 21 2011, 02:23 PM

As I understand simsense, the simrig is recording brainwaves and pulses and the sim module and ASIST are sending those signals to the sensory part of your brain. In the apple scenario, it's sending an electronic pulse that stimulates the visual part of your brain. ASIST turns those simsense signals into brainwaves.
It isn't recording information like a camera or a microphone, it's recording chemical and electronic reactions in the brain.

AIs don't have brains, so that signal doesn't affect them. They can get recorded information from something like a video camera, or cybereyes with an image link, but that pulse which stimulates the brain has nothing to stimulate in an AI.

What we need is someone to map the AI "brain" and create and interface for them so they can experience real world simsense the way metahumans experience the Matirx. That would be a cool adventure.

Posted by: phlapjack77 Jun 21 2011, 03:00 PM

Fortinbras, thanks - your interpretation has helped to quiet the rumblings in my head some.

And if I ever run a SR game again, I'm totally using your idea as a major plot hook

Posted by: Rubic Jun 21 2011, 03:08 PM

All an AI would need is the ability to interpret electrical signals for the data they represent. Thought has been codified enough that visual data has an electronic equivalent (visual simsense data). The AI does NOT need to have eyes to understand data, because at the point where the AI is dealing with it, this signal IS NOT chemical. The necessary translation to the digital has already been made, and at the digital level, the AI reads the data the way a brain reads data, with the exception that the brain needs to translate the data further. Corrupt data, IC, and feedback can and should cause damage to AIs comparable to any MEATBAGS, and each AI is inherently living in VR, as much as a full-con-borg.

Posted by: phlapjack77 Jun 21 2011, 03:15 PM

I'm thinking something like, it's like trying to read a Word document with Paint. Both are digital, but they just don't speak the same language.

As Fortinbras said, a translation package for humans has been created - the ASIST / DNi / whatever systems. Creating a translation package for AIs would be a great adventure plot.

Posted by: Rubic Jun 21 2011, 03:28 PM

You're right about that... but that also brings up the question of how a newly-exposed AI would react to the simulated bio-assault by Black IC or a computer virus.. Computer systems generally have precautions for it, but this could turn any given AI into a sort of 'Carrier,' similar to HMHVV for metahumans...

Posted by: phlapjack77 Jun 21 2011, 03:55 PM

VERY interesting plot point..."beta" version of an AI simsense translation package at large...has some bugs and other strange behaviors to it...PCs are contacted by a Johnson to retrieve / investigate, meetings are always over their commlinks or via VR, never face-to-face...

Posted by: Fortinbras Jun 21 2011, 04:25 PM

AIs are currently immune to Black IC, so a program that could translate simsense could be both a blessing and a curse.
Maybe Cerberus needs to make a reappearance...

Posted by: Aerospider Jun 21 2011, 04:49 PM

Simsense is artificial. If you have a real, physical apple then to see it a human brain needs eyes and an AI 'brain' needs cameras. Simsense will make a human brain think it sees an apple by translating man-made code into the relevant cranial stimuli. To do the same for an AI you already have code it will understand so you just give it that. The AI doesn't literally see through cameras any more than people do - the sensor turns photon-collisions into electronic data and whilst that's enough for the AI the human needs it turned into something else, even if it's back into a bunch of photons shooting out of a trideo for the eyes to then translate for the brain.

Posted by: Aerospider Jun 21 2011, 05:38 PM

How would Black Hammer/Out affect an AI any differently to Attack? Matrix damage is the only damage it can suffer. You'd have to go a long way to devise a more fearsome AI-killer than Nuke.

Posted by: Fortinbras Jun 21 2011, 06:06 PM

Black IC does Physical or Stun damage, not Matrix damage. AIs can only take Matix damage. AIs are immune to Black IC because they can't experience simsense.

A simsense file is a series of commands written by and sent to a metahuman brain. The ASIST allows it to be written. A Matrix entity can interpret it as a simsense file(it's not invisible) but can't experience it.

Think of it as a file that can only be read and written by a certain program, like MS Word(poor example, but I don't know computers)
The brain is MS Word and the sim module and sim rig are the computer that allow it to be written. So even if you have a computer, if you don't have MS Word(the brain) you can only read the source code, but not the file.

Again, sorry for the poor metaphore, computer people, but I'm trying my best.

Posted by: Rubic Jun 21 2011, 07:22 PM

I'd say it's more like the brain is Windows, while the AI is Linux. Linux has programs that can translate/process the information (sufficiently to read the data), but command codes that work in one will not work in the other. Because of this, Black Hammer/Out function like viruses for the metahuman brain (Windows), while not affecting an AI (Linux). "I understand what that attack was saying to me, but it didn't affect my code."

Posted by: Aerospider Jun 21 2011, 08:30 PM

Dude, I do know this stuff and hope that perhaps my posts in this thread have helped explain it for others. But I must have misunderstood you because your previous post reads as though you were speculating with others about creating an adventure campaign around some hypothetical AI simsense system that would somehow make Black programs a problem for them.

Posted by: Fortinbras Jun 21 2011, 08:39 PM

Yeah. Doing so would totally change the current rules as written. Sort of like what Dreamchipper did with simsense, Dragon Hunt did with dracoforms using VR or Tempo did with mundanes viewing astral.
It would be a total game changer, which is why I think it would be interesting.

I think it would be more interesting if the PCs get to decide whether to destroy the thing or put it on the Matrix for all to see, but that's just me.