SDN Worlds 4 Commentary Thread III

[Steve]

In the grim dark future of humanity, the Imperium of Man would designate some xenos as pets, others as food, and others, to be purged in a great fire of atomics.

What about the ones to be left alone, because they are good Imperial subjects and the wrath of the Anglian Empire would be upon you for harming them?

What the Imperium cannot achieve by the sword, it achieves by religion.

Ha. That'll be the day. Damn Emperor-worshiping crazies, apostates and heretics the lot of you!

[Shroom Man 777]

Maybe the Space Catholic/Anglican/Anglian/Orthodox/whatever church decreed that Heraclius couldn't get divorced, then he gave them the finger, married and divorced and beheaded a whole bunch of wives, made his own church, and declared himself its God. I think the Anglians would understand this, of all people.

Also. I can't seem to edit the wiki with my account. Any reason why?

EDIT:

Sigma meets Unit 7!

[Ilya Muromets]

James Bondroid? Ethan Huntbot? Sydney Brisdrone?

Eh, whatever.

I just hope that's a somewhat plausible way of sneaking in undetected, even if a bit of a stretch.

EDIT: Man, Shloomy, you shouldv'e taken that opportunity at that SEMENARY to become a priest. If sermons were like that I'd still be Catholic.

[Shroom Man 777]

STAS BUSH I know you're reading. Man, how does it feel to be a suicide officer?

Man, if you pulled that stuff in Bragule, your drones would've been picked up by the passive-aggressive sensors of the Highway Police's SAM sites, Malachuschus.

[Ilya Muromets]

Man, if you pulled that stuff in Bragule, your drones would've been picked up by the passive-aggressive sensors of the Highway Police's SAM sites, Malachuschus.

That's because BRAGULANS are more paranoid than paranoia itself. They probably annihilated a huge portion of their avian (or equivalent) population through overzealous anti-air defenses and ended up, I dunno, selling them as meat or something.

[Shroom Man 777]

Please educate yourself before posting more.

[Ilya Muromets]

Domo arigato, Mister Shroomoto.

I needed the lulz.

Bragulan acid trip!

[Simon_Jester]

Why do I get the feeling that a Bragulan acid trip involves actually tripping and falling into acid?

[Fingolfin_Noldor]

Why do I get the feeling that a Bragulan acid trip involves actually tripping and falling into acid?

Because Bragulans are masters of the literal.

[Shroom Man 777]

*looks at checklist*

*checks off Shep's request for moar FIGHTER MAFIA*

*checks off 'post parody of SDN thread'*

*checks off Fin's special request*

Yeah, I'm good for today.

[Steve]

And you do all that and Siege and I come in with a reunion post that's gone incredibly, horribly dark. I believe TVTropes uses the term "Mood Whiplash".

Poor Clarice. You all only get one guess as to which group would be responsible for this crap.

[Siege]

Yeah, no more lambs for poor Clarice...

[RogueIce]

You all only get one guess as to which group would be responsible for this crap.

Evil Moogles?

[Steve]

You all only get one guess as to which group would be responsible for this crap.

Evil Moogles?

*inflicts a Gibbs smack on you*

[Shroom Man 777]

Oh yeah, I forgot to give credit. All screencaps in my FF7 post were generously donated by Fin.

[Simon_Jester]

Update to the Battle of Hawk's Nest is posted, introducing the Umerian frigate USS Farbanti!

[Simon_Jester]

In light of Darkevilme's latest... I can see why it's relevant. I suspect what screwed over Sneakily Does It was the AI's conflicting directives about achieving the mission, avoiding capture, and so forth and so on. Moreover, the AI had de facto tactical control of the ship (the captain could not order it to stand down), but had no sense of strategy. And therefore no ability to subordinate tactics (such as making a last-ditch attempt to achieve the mission while committing suicide to avoid capture) to strategy (such as realizing how bad a plan that was in the broader context).

But an AI with the discretionary ability to decide what to do in the strategic context is an AI that may well decide to act against the intentions of its makers, either "for their own good" or "for the greater good;" for examples of this we need look no farther than Schlock Mercenary, where AI-controlled warships do this on a regular basis.

Governments have reason to be alarmed by that prospect. Which leads to AIs with hardwired directives that they cannot bypass and don't see the need to... which in turn leads to buggy AIs that do bloody stupid things.

“So what was your plan? Come in, rescue the poor little children, and get out? Against people who you had to know would be ready to deal with gifted individuals like you’ve become?” Q shook his head and made a dramatic “tsk tsk tsk” sound with his tongue. “I suppose you really weren’t thinking at all, were you? You just knew your precious charge was in trouble and wanted to do something. Unfortunately heroics don’t always work that way. You can’t always save the innocent.... as you’ve now discovered.”

And this is where picking up dialectics comes in handy.

Thesis, antithesis...

Synthesis: Next time, have a plan.

[Fingolfin_Noldor]

This is precisely why the Imperium is deeply suspicious of AI/CIs and why they are heavily controlled. We have no way of knowing whether the said AI/CI is subverted in any manner, and complex AI/CIs require days of analysis to ensure they are free of "known bugs".

[Simon_Jester]

This is precisely why the Imperium is deeply suspicious of AI/CIs and why they are heavily controlled. We have no way of knowing whether the said AI/CI is subverted in any manner, and complex AI/CIs require days of analysis to ensure they are free of "known bugs".

Not surprising: it's easy to become a victim of a vicious cycle when programming AIs. Attempting to ensure loyalty of AIs by brute-force programming of imperatives into them makes them buggy, and debugging them is effectively impossible because it's the brute-force loyalty imperatives that creates the problem. If, by definition, any acceptable AI is one with powerful loyalty imperatives, no AI can be truly acceptable and useful at the same time.

[Ryan Thunder]

This is precisely why the Imperium is deeply suspicious of AI/CIs and why they are heavily controlled. We have no way of knowing whether the said AI/CI is subverted in any manner, and complex AI/CIs require days of analysis to ensure they are free of "known bugs".

Yeah, the Union doesn't use the sort of AIs most people here mean for anything important. They do have specialist 'intelligences' for things like war robots and managing factories and what not.

Most of the things they use them for are things that you wouldn't really want to inflict on an involuntary sapient mind, anyway.

[Darkevilme]

Simon is pretty much right on the money regarding this. The Hierarchy was at the same time envious of the extra potency granted to the Solarians by their shipminds and paranoid about AIs hence the result of the Blade experiment.
Arri of course had a code to immediately shutdown the AI and return the ship to normal control on her, it was no coincidence Callahan waited till she wasn't on the bridge to try and pull this.

And i suspect the honest answer to Tia's question will be 'there is no need, the Sovereignty is its AIs, the correct question is how does one keep the humans loyal'.

[Simon_Jester]

To which, in turn, the answer is "we don't."

The Umerians almost went down that road; it really hinged on Geppetto deciding in the historic past not to take over the country, either "for its own good" or (more temptingly) for experimental purposes. He could- he's good enough at social engineering that if he wanted to invest the time and effort into amassing more computing power he could be running the place.

But he lacks ambition or interest in anything other than "psychology in all its forms," and whoever programmed a sense of ethics into him some time in the late 32nd century did a good enough job to keep him from being willing to run controlled experiments on billions of people.

Or at least, the kind of controlled experiments he'd need to be running the country in order to perform.

More generally, the awareness that we could wind up going down that road is a huge discouraging factor in Umerian AI research, and a huge promoting factor in their development of nonsentient expert systems. They want computers smart enough to react to well-defined circumstances gracefully, but dumb enough to simply not try to react to the unexpected, rather than making a sophomore mistake like Sneakily Does It did.

[Siege]

Despite all signs to the contrary, CIs are not running the Sovereignty. They could if they wanted to. But as the good Brigadier points out, why would they want to?

[PeZook]

Hrmf...they control the infrastructure, make it possible for the government to be extraordinarily small, run space traffic, maintain the stock exchange, keep the armed forces operational.

They do run the sovereignty, they just don't control its policy. And yeah, if Olympic decided he's gonna be President, there'd be little the puny humans could do

Collectors got to this point a long time ago, of course, so we thankfully no longer have that problem.

[Simon_Jester]

My reply is a variant on PeZook's:

Despite all signs to the contrary, CIs are not running the Sovereignty. They could if they wanted to. But as the good Brigadier points out, why would they want to?

They are, however, administering the Sovereignty. They may not be the directing intelligences of the system (as Geppetto would put it), but since it is their behavior which defines the constraints of what can and cannot be done within the system.

At a bare minimum, Olympic "controls" the Sovereignty government in the same sense that the laws of physics "control" the motion of an airplane. Olympic is more flexible and easier to reason with than physics, but ultimately no more subject to human authority. And if the pilot thinks he is really, indisputably in charge, he has only to try to do something impossible within the constraints of the system to prove himself wrong.

That is a great deal more influence than some people would be comfortable extending to something whose behavior they are inherently incapable of modeling. Autocracy is easy to understand, democracy only somewhat less so, and while there's a huge amount of friction that makes it hard to get things done in either system, it's still within the bounds of what human beings can understand. Any given component of the system is something that a human can examine, analyze, and comprehend.

Rule by Olympic... is not.

That is not a criticism of Sovereignty policy, by the way; it's an observation on the culture. If you want critical examination...

[Geppetto]
"I still cannot understand why you consider it a point of pride to deem yourselves "computational." I am an artifact, a system designed by intelligent minds to accomplish tasks, and proud of it. The alternative is design by the blind idiot god of evolution- granted, most of the intelligences in this region of space are evolved, but it's nothing to be proud of..."
[/Geppetto]