Remove ads

Shepard Quest Mk V, Base of Operations (ME/MCU)

TheEyes said:
Not really, no. Mobile Frame research is about linking an AI to a robot body closely enough that they can take direct control of its actions, essentially building a cerebellum into the AI core. Without it an AI could probably drive a robot, but much the same way a human would drive a robot: it would be able to order the robot around, but many of the minute, second-to-second adjustments would be done by the onboard hardware rather than the remote blue box (which at this point can still be the size of a server room).
Click to expand...
Click to shrink...
...what? What magic will the AI be using to pilot the body then? You don't need to build a "cerebellum", the AI will create a program for piloting a body on it's own.

Again, they're close, but not quite. Ride-along AIs still have relatively large blue boxes; this upgrade just lets them be kept elsewhere while the AI can project its consciousness directly into someone's powered armor.
Click to expand...
Click to shrink...
Remember the "Alliance Infiltration Units" in ME3? A bluebox isn't enormous - it's something a person can comfortably carry. Nor does this require it's own upgrade, EDI does that by herself.

Software AI could forego the blue box entirely and upload directly into the armor, or it could stick to a remote location and project like a Ride-along. Both of these are relatively subtle differences, but can be important to the narrative. We might want to restrict the development of Mobile Frames to start with, to reduce potential comparison to the Geth; in fact it may be required we make that restriction to start with. Pure Software AIs are similarly dangerous, as they make it relatively trivial for the AI to self-replicate if we get a bad Trait roll.
Click to expand...
Click to shrink...
Don't want Software AI? Then make Bluebox AI - congrats, they require a hardware component. These distinctions are redundant.

Yeah, genetic engineering should totally be covered by reproduction; otherwise there's no point in having kids. :)
Click to expand...
Click to shrink...
I...don't even know what to say. Just look at what the practical effect of those techs is. Is that something that requires it's own step in the chain? Or is it something you would expect to already be included in a 3200 point tech?

What you mean is there's no way to customize or custom-build an AI, and you'd be right. We'd be stuck with whatever natural/artificial evolution decides to throw at us over time.
Click to expand...
Click to shrink...
What makes you think the personality of an AI can't be molded during creation, or even afterwards? VI have dedicated modules for their "personality", I don't see why AI wouldn't too even if they would be much more complex and change over time.

Eventually the positive and negative Traits would rack up so much that we'd have to build our own Geth Consensus just to keep the AIs from becoming drooling basket cases, which I think is an interesting narrative consequence (and a good enough origin story for the Geth).

Well, when it comes to AIs there's serious precedent for Tony Stark coming with those abilities built-in. I mean, Jarvis is a thing in just about every Iron Man continuity there is; he's almost as integral to the stories as the suits are. I guess I can see an argument for everything from Advanced Logic Probes on through Advanced AI Simulation waiting for more Science Heroes; maybe we can recruit Parra Ventura or Dr. Light? :D Dr Light would be especially interesting as he could maybe have other tech that would reduce the Uncanny Valley effect for human-looking AIs, to say nothing of learning how to make a Copy Chip.

I'm a bit hesitant to say so, but most of the Mass Effect/eezo tech, on the other hand, even the stuff already part of the tech tree, really has no precedent in the Iron Man continuity, other than the fact that Tony must have inertial dampeners and kinetic barriers of some kind in those suits or he'd have pasted himself dozens of times in the movies alone.
Click to expand...
Click to shrink...
Yeah, a good approach would a Science Hero unlocking a couple techs that improve AI capabilities from what they are described as initially. Don't see why any of the ones you suggested would be involved though, those all boil down to "experience and specialized tools let you make AI more gooder" - when they should all really be a part of what allowed you to create an AI in the first place.

Unless you are proposing that we can make an AI without being intimately aware of how it works?
 
Van Ropen said:
...what? What magic will the AI be using to pilot the body then? You don't need to build a "cerebellum", the AI will create a program for piloting a body on it's own.
Click to expand...
Click to shrink...
How is it this that sticks out as being logically impossible? Half the sci-fi tech in this thread is at best "elegantly incorrect" with the rest being outright impossible. Humans need a cerebellum to operate a body with anything like normal human fidelity and range of motion in real time; I'm just saying an AI needs one too. It's like having dedicated hardware for graphics: sure, you can emulate all that on a CPU, but not fast enough to stream video.

Remember the "Alliance Infiltration Units" in ME3? A bluebox isn't enormous - it's something a person can comfortably carry. Nor does this require it's own upgrade, EDI does that by herself.
Click to expand...
Click to shrink...
EDI's bluebox is the size of a server room, and she connects remotely to the gyroid to operate it while her "consciousness" stays on the Normandy in her blue box. She can do ride-along inside the squad's HUDs because that's her job, and so obviously Cerberus figured out how to get her the capability to do it. Nobody explains what the "Alliance Infiltration Units" are in any of the tech pages; I'm assuming they function like the person their skin is based around, and that there's a bluebox flying around overhead in a ship somewhere like there is for EDI herself.

Yeah, a good approach would a Science Hero unlocking a couple techs that improve AI capabilities from what they are described as initially. Don't see why any of the ones you suggested would be involved though, those all boil down to "experience and specialized tools let you make AI more gooder" - when they should all really be a part of what allowed you to create an AI in the first place.

Unless you are proposing that we can make an AI without being intimately aware of how it works?
Click to expand...
Click to shrink...
Most AIs in fiction have some sort of "black box" area that can't be easily edited without destroying the AI. Mass Effect AIs have their quantum blue boxes which serve this function. As far as I can tell, the normal Mass Effect way of "fixing" an AI is to wipe it and start over. The notions of logic probes and growpamming, which I stole from Schlock Mercenary, are about being able to peel back the blue box without destroying it, which, yes, should be non-trivial when talking about blue box AI.
 
Last edited:
Said Gyroid was originally a AI that passed for a human researcher by herself, so presumably her bluebox had to fit within the body and whatever else the body needed to appear human while still being able to outrun Spectres Biotic charging her. (Seriously, Bioware: What the heck?)

And note that EDI is designed to be a Cyberwarfares suite, as well as running the ship by herself if she wants to. That server room? Enough processing power to combat the collector Station by herself, and a host of other things.
 
TheEyes said:
How is it this that sticks out as being logically impossible? Half the sci-fi tech in this thread is at best "elegantly incorrect" with the rest being outright impossible. Humans need a cerebellum to operate a body with anything like normal human fidelity and range of motion in real time; I'm just saying an AI needs one too.
Click to expand...
Click to shrink...
Except that doesn't make sense, because this is something that should be handled purely via software. If VI can control mechs without any problems, I don't see why an actual AI wouldn't just have programs dedicated to doing so. So yes, there is a cerebellum-analog: it is software that doesn't need to be developed separately. Esbilon has made clear that speed-bump techs are unwanted.

In fact...yeah:
She adapts the body's existing software for hand-held firearms rather than downloading one from a security firm.
Click to expand...
Click to shrink...
It is explicitly stated that she is modifying and creating the software for control of the body as she uses it, that is the fluff explanation for her leveling up.


EDI's bluebox is the size of a server room; she can do ride-along because that's her job, and so obviously Cerberus figured out how to get her the capability to do it. Nobody explains what the "Alliance Infiltration Units" are in any of the tech pages; I'm assuming they function like the person their skin is based around, and that there's a bluebox flying around overhead in a ship somewhere like there is for EDI herself.
Click to expand...
Click to shrink...
So the entire time that Dr. Eva Coré was on Mars, a Cerberus ship was hanging out in orbit remotely controlling the body? Same for every deployment of those Infiltration Units?

EDI's hardware in general is the size of a server room. Given her function is handling electronic warfare for the Normandy (her job is NOT doing ride-along), I imagine there is a lot more there than just her bluebox in that room.


Also, what benefit is their to a Ride-Along AI tech at all? Is there anything it allows that an onboard VI (which we already have) or miniaturized QEC's in general won't allow?


Edit - I am going to kill you Alanek. With a sneak attack, for maximum irony.
 
I just lurk and lurk, till my bugs pick up your typing. Then I post my response. :p

I'm just happy I managed to get 2 of your 3 points almost exactly the same a mere two minutes before you.
 
Just dropping in to clarify what Van Ropen said. The tech tree works on a narrrative logic rather than a scientific one. That means that techs that should work very differently under the hood are rolled into one. Speed bumps suck.
 
Do you think that since we have bionics we should add a biotic skills tech tree? perhaps developing our own unique biotic techniques.

Oh and I just looked up biotic charge again,

http://masseffect.wikia.com/wiki/Charge said:
This power also compensates for impeding obstacles by allowing the Vanguard to 'phase' through solid objects en route to their target. It can also be used on enemies who have been levitated by other biotic powers, and can multiply physics damage to send them off into the distance.
Click to expand...
Click to shrink...

They explicitly stated phase through that has practically infinite combat applications!
 
Why are you limiting it to combat? That's useful for so many other things too.

EDIT: Hey, if that works similarly to the massless corridors involved in FTL, what happens if you collapse the field while the two objects are superimposed?
 
Last edited:
Remember the comment on how techs here serve narrative functions? We have already moved beyond conventional biotics with our amp. Biotic Armor should give us the deep understanding and ability to replicate certain effects mechanically. These don't need individual techs.
 
Still things like the Assisted Biotic Charge would be useful not only in transport but deployment of forces, in combat manoeuvring and even for things like bunker/armour penetrating missiles. specialised devices would be smaller, much more so.

DamionKenley said:
Why are you limiting it to combat? That's useful for so many other things too.

EDIT: Hey, if that works similarly to the massless corridors involved in FTL, what happens if you collapse the field while the two objects are superimposed?
Click to expand...
Click to shrink...
I'm not sure, perhaps they will either explode due to nuclear fusion or the objects would separate - violently or the objects would simply forever merged.
 
If Biotic Armor is full on allowing non-biotic soldiers to use biotics attacks with a suit of specialized armor, then it should be no problem to turn the Biotic Charge to industrial use.


Can you collapse the field while things are within it? Because if whatever is sent through the relay travels literally instantaneously, then there is no moment that the field can be collapsed. Of course, the Charge seems to be a different albeit related phenomenon - given how the person creating the field is within it, not sure how it could be collapsed.
 
Hmm.

You know, I was going to try to respond to the last replies, but I'm starting to realize I'm just wasting my time, again. Once again I seem to have stuck myself with defending a proposal that is universally disliked: forget constructive criticism, I can't even get so much as a "Like" for the idea! Clearly there isn't a glimmer of interest for this idea either. Better luck next time I guess.

Instead of defending my own idea, I'll just articulate the problem I was trying to solve, which in hindsight is probably what I should have done anyway. So, previously we have @Esbilon telling us he wants to stay away from superintelligences:
Esbilon said:
I've mentioned before that AIs are probably going to be limited to human-level intelligence and rough personality. Not because there are no precedents I can lean on for writing inhuman ones, but rather because a true artificial superintelligence would make everyone else irrelevant.
Click to expand...
Click to shrink...
At the same time, it's also been heavily implied, now stated outright, that he doesn't want to deal with "speedbump" tech barriers:
Esbilon said:
Just dropping in to clarify what Van Ropen said. The tech tree works on a narrrative logic rather than a scientific one. That means that techs that should work very differently under the hood are rolled into one. Speed bumps suck.
Click to expand...
Click to shrink...

The problem is how to create daylight between those two statements, so that we can have AI tech featuring prominently in our story without instantly creating AI superintelligences. Statements like those from @Van Ropen would imply that is impossible:
Van Ropen said:
Unless you are proposing that we can make an AI without being intimately aware of how it works?
Click to expand...
Click to shrink...
Van Ropen said:
If VI can control mechs without any problems, I don't see why an actual AI wouldn't just have programs dedicated to doing so. So yes, there is a cerebellum-analog: it is software that doesn't need to be developed separately. Esbilon has made clear that speed-bump techs are unwanted.
Click to expand...
Click to shrink...
The above statements imply that we can't learn how to make AI without immediately knowing how to make an AI arbitrarily smart, and everything an AI would want to do must not require specialized hardware, because doing otherwise would violate the "no speedbump" rule. How, then, can we do any research into AI without instantly travelling into territory that Esbilon has said is verboten?
 
Last edited:
TheEyes said:
Hmm.

You know, I was going to try to respond to the last replies, but I'm starting to realize I'm just wasting my time, again. Once again I seem to have stuck myself with defending a proposal that is universally disliked: forget constructive criticism, I can't even get so much as a "Like" for the idea! Clearly there isn't a glimmer of interest for this idea either. Better luck next time I guess.

Instead of defending my own idea, I'll just articulate the problem I was trying to solve, which in hindsight is probably what I should have done anyway. So, previously we have @Esbilon telling us he wants to stay away from superintelligences:

At the same time, it's also been heavily implied, now stated outright, that he doesn't want to deal with "speedbump" tech barriers:


The problem is how to create daylight between those two statements, so that we can have AI tech featuring prominently in our story without instantly creating AI superintelligences. Statements like those from @Van Ropen would imply that is impossible:


The above statements imply that we can't learn how to make AI without immediately knowing how to make an AI arbitrarily smart, and everything an AI would want to do must not require specialized hardware, because doing otherwise would violate the "no speedbump" rule. How, then, can we do any research into AI without instantly travelling into territory that Esbilon has said is verboten?
Click to expand...
Click to shrink...
This is what the GM said on page 120:
Esbilon said:
Regarding limiting growth, I'm canning that idea. Go ahead, spread across the galaxy like a plague of high-tech locust. It might be educational :)

At some point, you'll probably want to look into constructing 40K-style industrial worlds in order to have room for your facilities, but I think you'll manage.
Click to expand...
Click to shrink...
Your point?
 
Quite simply. Quality vs. Quantity. Intelligence being something of a quality, not directly related to the speed of thought and the amount of data at easy access. An A.I. can have a thousand conversations in parallel, perhaps, but it won't get any more leaps of logic than a human would. Basically, intelligence in the sense of inspiration, ability to create qualitatively new knowledge and such can, in fact, be quite limited. A.I.s would be quite different from organics and their development wouldn't be a speedbump. They would play big roles, I think:
1) Managing large massives of data
2) Living in and exploring environments not fit for organics, including photospheres of stars, virtual reality, gas giants and death worlds, deep space even
3) Bringing different perspective

And other stuff. A.I. doesn't need to be a transcendent Mind to play an important role and be worth the investment.
 
TheEyes said:
You know, I was going to try to respond to the last replies, but I'm starting to realize I'm just wasting my time, again. Once again I seem to have stuck myself with defending a proposal that is universally disliked: forget constructive criticism, I can't even get so much as a "Like" for the idea! Clearly there isn't a glimmer of interest for this idea either. Better luck next time I guess.
Click to expand...
Click to shrink...
I am interested in your proposals, I just haven't yet taken the time to look at them in detail. They, along with all other suggested techs, are on my list.

And I fully expect AIs to be relevant and different if you decide to pursue them. Striking a good balance should be possible, for the reasons Yog mentioned if nothing else.
 
DamionKenley said:
Why are you limiting it to combat? That's useful for so many other things too.

EDIT: Hey, if that works similarly to the massless corridors involved in FTL, what happens if you collapse the field while the two objects are superimposed?
Click to expand...
Click to shrink...
Explosions.
BIG explosions.
 
Should we develop a house protocol like in Iron Man 3? That way if we're ever captured without armor, we can have our VI send some unmanned suits our way?
 
TheEyes said:
Hmm.

You know, I was going to try to respond to the last replies, but I'm starting to realize I'm just wasting my time, again. Once again I seem to have stuck myself with defending a proposal that is universally disliked: forget constructive criticism, I can't even get so much as a "Like" for the idea! Clearly there isn't a glimmer of interest for this idea either. Better luck next time I guess.

Instead of defending my own idea, I'll just articulate the problem I was trying to solve, which in hindsight is probably what I should have done anyway. So, previously we have @Esbilon telling us he wants to stay away from superintelligences:

At the same time, it's also been heavily implied, now stated outright, that he doesn't want to deal with "speedbump" tech barriers:


The problem is how to create daylight between those two statements, so that we can have AI tech featuring prominently in our story without instantly creating AI superintelligences. Statements like those from @Van Ropen would imply that is impossible:


The above statements imply that we can't learn how to make AI without immediately knowing how to make an AI arbitrarily smart, and everything an AI would want to do must not require specialized hardware, because doing otherwise would violate the "no speedbump" rule. How, then, can we do any research into AI without instantly travelling into territory that Esbilon has said is verboten?
Click to expand...
Click to shrink...
...fuck no. You've either entirely missed my point or you're deliberately skewing what I am saying.

Those statements were meant to say that the techs you have proposed are unnecessary, because our AI will do those things anyways. Not that there was no way to have AI feature predominantly without them being superintelligences. Not everything we do requires a separate research project.

Pretty much every single proposal you made? Something we can have our AI do, we just don't need a tech for it. I don't see why you think AI won't feature predominantly, that's a story thing - nothing to do with techs.


Edit - Oh, and as far as I can tell everything offered has been constructive criticism. They haven't been personal attacks or anything of the sort, just reasoned statements about why the proposal is redundant.
 
Last edited:
Ridtom said:
Should we develop a house protocol like in Iron Man 3? That way if we're ever captured without armor, we can have our VI send some unmanned suits our way?
Click to expand...
Click to shrink...
Unlike Tony, we actually trusted other people with our suit... instead of having a VI with some unmanned suits, we have a few dozen highly trusted bodyguards in suits.

It might be nice to have some kind of remote deploy system for our suits tho. While the suits in IM3 seemed incredibly fragile compared to previous versions, the remote deploy shown near the end of Avengers would actually be a highly practical system, assuming you don't need to be in freefall to use it.
 
Esbilon said:
In your time zone, you got it right, in mine you missed it by a few hours, but thanks!

I'm happy to say that I passed my defense and should be able to get back to our regularly scheduled questing shortly. Though not before my celebratory hangover passes.
Click to expand...
Click to shrink...
Just a question... are you still hungover? :p
 
Lagrange said:
It might be nice to have some kind of remote deploy system for our suits tho. While the suits in IM3 seemed incredibly fragile compared to previous versions, the remote deploy shown near the end of Avengers would actually be a highly practical system, assuming you don't need to be in freefall to use it.
Click to expand...
Click to shrink...

That fragility might be linked to them being mostly prototypes and therefor not particularly well armored.
 
Van Ropen said:
...fuck no. You've either entirely missed my point or you're deliberately skewing what I am saying.

Those statements were meant to say that the techs you have proposed are unnecessary, because our AI will do those things anyways. Not that there was no way to have AI feature predominantly without them being superintelligences. Not everything we do requires a separate research project.

Pretty much every single proposal you made? Something we can have our AI do, we just don't need a tech for it. I don't see why you think AI won't feature predominantly, that's a story thing - nothing to do with techs.
Click to expand...
Click to shrink...
The problem is that you're, perhaps accidentally, providing an argument by induction: for every amount X of intelligence that we can create n our AI, we must be able to impart X+C, because that little bit extra is not narratively important enough to justify a tech. Further, we cannot restrict our AI from interacting with the world in increasingly complex ways, because every way that can be thought of is "only software" and therefore is also not narratively important to justify a tech. Therefore, the moment we are able to make any AIs at all, we must be able to make arbitrarily smart, arbitrarily well-connected AIs. Even if we decide to reject that argument, our AI "tech tree" is going to remain hilariously short, with a grand total of two entries: Blue Box AI and Pure Software AI, although we'll probably have to remove that distinction because it, too, isn't really narratively important when blue boxes are so small that they can fit inside a humanoid mobile frame.

Lagrange said:
Unlike Tony, we actually trusted other people with our suit... instead of having a VI with some unmanned suits, we have a few dozen highly trusted bodyguards in suits.

It might be nice to have some kind of remote deploy system for our suits tho. While the suits in IM3 seemed incredibly fragile compared to previous versions, the remote deploy shown near the end of Avengers would actually be a highly practical system, assuming you don't need to be in freefall to use it.
Click to expand...
Click to shrink...
Well, to be fair, it's not so much that the suits are fragile as it is that even the unstable Extremis that Killian had was pure undiluted bullshit. Melting and tearing down a water tower after running your fingers along one edge of one support strut for a few seconds is pretty ridiculous. But yeah, a remote deployment system would be really useful, even if it's only a "football" that we carry around like in Iron Man 2.
 
Last edited:
You must login or register to reply here.
Remove ads
Back
Top