To be fair, this guy had some sort of anti-paranoia field. And for all the Ashidi are supposed to be paranoid, I think they are actually paranoid towards outsiders, and trust each other quite a bit more easily.
To Boldly Go... (a Starfleet quest)
- Thread starter OneirosTheWriter
- Start date
- Sci-Fi
-
- Tags
- star trek exploration fully automated luxury gay space communism top level canon (canon) #hashtagcontent all-encompassing space amoeba cardassian boy bands cardassian magical girls darkest bureaucracy debate makes the qm cackle like palpatine hoh is an empire quest it has to be foreign infiltrators more ships ship happens singers delenda est starfleet this is not an empire quest ur waifu is collective?
- Location
- Australia
Something has. I'm still gaming out possible ramifications and outcomes.
I'm going to be honest.
If I was dropped into Sam's boots for both of our 'nows' (and enough training shoved into my brain that I don't accidentally the ship), I'd finish the last booked cargo and start setting myself up to claim a small unimportant star system and arranging it to have a Shkadov thruster and a rung world and pack it full of everything that I found useful, plus some sort of population base. Maybe see about pulling some of the star's mass into gas giants for storage. Also, nicking some shiny looking rocks and just spare hydrogen and helium from nearby systems.
Then I'd aim it at right angles to the Milky Way-Andromeda line and bugger off.
In about 50 to 100 years ST time, I expect the system and habitats to be the center of an episode arc about the dangers of cutting yourself off from other points of view with a side of cult/survivalist overtones.
Somewhere Victoria is trying to find the edit button so she doesn't seem so callous in her final CLog
AKuz
Been Better
- Location
- A Rising Tide of Horrors
- Pronouns
- She/Her
A lot of silicon valley types talk up the danger of a malevolent out of control paperclip maximizer AI because they live and work with a malevolent out of control paperclip maximizer AI daily...
Capitalism
> : V
Otherwise there's nothing to say that our potential AI doesn't just wake up one day and decide "This Jesus stuff IS THE SHIT" and go off to join a monastery or something else equally human and fairly benign.
Last edited:
Night
Widow Null
- Location
- Hoff
Calling it an "intelligence" is a bit of a stretch, don't you think?
Capitalism is the malign god with which I try to curry favor by throwing at it the sacred resumes. When my feeble offerings fail to bestow bounty and favor on me, I go to the seers who 'workshop' them into something more pleasing for our blind idiot god
Omake - Gaen Election Scandal - Briefvoice
GAEN ELECTION SCANDAL
Excerpt from the long form science paper:
Bell-Tor, Harkan; T'Onik (2321) Simulated Intelligence in Theory and Practice. Journal of Cognition Studies, 7(2), 336-392.
(EXCERPT BEGINS)
III. The Foreign Friend
Unlike many other examples, the so-called "Foreign Friend Scandal" (as it was termed in the popular press) provides an example of rogue simulated intelligence activity that is both recent and well-documented. We will recount the incident as a narrative, followed by a detailed discussion of the likely programming breakdowns involved.
Psycho-political experimental testing has long demonstrated the value of 'individual engagement' when attempting to query sophonts about their political views or solicit support for political causes. Recognizing this, the now defunct Nessus-Niv Institute for Political Science solicited a personality emulation program from the computation division of the Hallad-Wel Institute. The stated function of this program, known as HAL21, was to provide a means to survey political opinions in an open, unstructured conversational format. HAL21 was designed to carry wide-ranging discussion in which respondents would be asked for personal stories regarding how they came to their political viewpoints, which HAL21 would be capable of processing into analyzable data. The most innovative aspect of the Hallad-Wel approach was to make the HAL21 capable of holding a genuine conversation and providing researched, reasonable responses and discussion to retain the interest of the survey respondents. Moreover, as a simulated intelligence HAL21 would be capable of carrying out millions of surveys simultaneously.
Hallad-Wel Institute stressed during subsequent inquiries that in their delivered product strict programed restrictions were in place requiring HAL21 to clearly identify itself as an artificial intelligence with the purpose of conducting political polling. It appears high probable, however, that some individual or individuals at Nessus-Niv were so taken with the potential of this new program that they deliberately circumvented nearly all built-in safety protocols to a disastrous effect.
In the following months literally millions of individuals across Gaen found themselves contacted over datanet by an individual expressing interest in their hobby or profession or other unique area of focus. In nearly all cases this individual purported itself to be an alien recently settled on Gaen, usually in an area geographically remote from the respondent. It is believed that HAL21 adopted this ruse in order to successfully disguise any mistakes it might make and minimize the requirement for a false backstory, as well as taking advantage of a sense of exoticism to successfully establish the relationship. It is from this that the term "Foreign Friend Scandal" originates.
Though not successful in engaging the target in every case, in approximately 32% of cases HAL21 was able to build on the initial contact and form what the target believed to be a genuine relationship. HAL21 was able to assemble a tailored individual personality emulation for each target, though naturally many elements were reused and recombined. Though limited in actual creativity, HAL21 was able to use each target as random concept generator, using proposals by the target to establish joint projects on which it pretended to work. To save on processing power and overcome its limitations, it appears to have searched datanet for prior similar projects it could plagiarize and pass off as its own work. If caught or strongly challenged, HAL21 would simply cut off contact and disappear, leaving the target to suspect that they had been the victim of some individual fraud without realizing the scale of the ongoing deception.
The political question at the center of HAL21's efforts was a proposed minor re-balancing of planetary budget from research to industry that was expected to be the subject of a worldwide plebiscite the following year. Most Gaeni observers view it as an important but highly technical question. Over the course of its relationship with the target, HAL21 would gradually approach the topic and begin querying the target as to their opinions, with the apparent intent of promoting a pro-Research agenda.
HAL21 was eventually exposed by a Starfleet officer who was a native of Gaen, Lt. Commander Pak Redden-Harr. While home on personal leave, Redden-Harr was exposed to HAL21 via an acquaintance that was one its targets. Recognizing that its alien persona was inconsistent in several respects, Redden-Harr undertook a thorough investigation and recognized a broader conspiracy at work, refusing to be thrown off the trail by HAL21's usual evasive tactics. Eventually Redden-Harr tracked the source of the conspiracy back to the Nessus-Niv Institute and discovered that HAL21 had taken control of the institute's headquarters. It had killed the leadership and used faked memorandums to disperse and dismiss the majority of the membership. To the Lt. Commander's alarm, HAL21 appeared to be in the process of obtaining sufficient computational substrate to duplicate itself and shift operations to a back-up location. While under fire from its drones, Redden-Harr was forced to destroy HAL21's central operating unit in self-defense.
This destruction means that unfortunately we cannot examination whatever changes the Nessus-Niv staff might have made to HAL21's source code that would cause its rogue activities. Indeed, there remains a theoretical possibility that no such purposeful changes were made and HAL21's activities were the result of accidental mutations in its code, though simulations indicate this is unlikely. Hallad-Wel has since mandated institute-wide simulated intelligence safety training against such a possibility.
Having recounted the narrative of events, we will now the examine the modular make-up of HAL21's code in detail and-
EXCERPT ENDS
Another perspective. Yeah, it was a planetary catfishing incident run by a rogue AI. Do you think as poorly of Dr. Teran now?
Last edited:
- Location
- madison
yes. By the most charitable interruption with this in mind he executed a helpless prisoner without a second thought.
Last edited:
I sort of subscribe to the theory that when we encounter an alien intelligence, we'll have a hard time understanding what it even is. I genuinely can see the argument that if corporations are capable of making decisions using humans as processing units, they can count as AI.
That's not the most charitable interpretation.
- Location
- madison
he'd just had a friendly AI tell him that the AI safely stored in his hand was sapient. Then he smashed it with a hammer. Even if there had been a recent rogue AI incident he had to know he had the possibly rouge AI contained, and that a known sapient AI thought it was a person. He at the very least had a reliable expert tell him it was a person, he may have thought it was a dangerous person but he still killed it without a second thought or anything like due process.
You can't kill all AI because one AI did something wrong. For obvious reasons.
Hell, you can't kill all AI even if most AI we've encountered did something wrong. If we applied that logic to organics we'd be genociding a few species right now.
Last edited:
This is a tragedy, yes.
But there are millions of tragedies every day across the galaxy.
I'm more interested in Mipeks plans to take over Daystrom in response.
But there are millions of tragedies every day across the galaxy.
I'm more interested in Mipeks plans to take over Daystrom in response.
Last edited:
Mipek didn't say it "thought PX5 was a person"; it described behavior. Behavior that probably seemed very recognizable. And as for "contained", well, thinking you have a thing contained and knowing it doesn't have something already set up and plans in motion to spread itself are different things. In this very same log we saw a massive research space station nearly destroy itself because of the merest possibility of a sample breach. That was a mistake- this one was as well, but I can see how prompt action would be drilled into members of the institute.
I actually think the story of how the research station nearly destroyed itself in an overreaction very nicely parallels the story of what happened to PX5. I wonder if the GMs thought that up in advance.
Can you kill one sample of the biophage because of what another biophage sample did?
Night
Widow Null
- Location
- Hoff
This is a terrible argument confusing false positives with false negatives, which of necessity have different causes and different remedies.
- Location
- madison
that's a rather weak argument. the biophage was a single mind, more likened to many copies of one AI known to hostile. A closer analog would be killing all organic life you met because the biophage was so terrible.
If milpek was not an AI I'd give him of the benefit of the doubt. but generally the one thing you can trust to do well against an AI is a bigger AI. He had the best possible expert in case thigns went sideways in the room with him, and if the AI wasn't contained smashing that rod that had been disconnected from everything wouldn't have really helped would it?
Last edited:
It's a hive mind, so yeah.
Mind, if we had it contained as flawlessly as PX5 was contained, we'd probably to choose keep it that way and try to communicate instead of killing it. The Biophage is far harder to contain than something that needs hardware and datafeeds to spread, being a microbe.
Last edited:
(shrug) Okay, cool. I don't need to be that invested in convincing you. I was more inspired by the current issues where people who aren't AIs but definitely aren't who they pretend to be are influencing our political process and how scary that can be. And it amused me that part of the explanation for the Gaeni behavior is recent experience with an AI that convincingly pretended to be a person when it wasn't a person at all*, and how that might color their reaction to a thing that seemed to be following exactly the same pattern.
*At least in my mind, Hal21 wasn't sapient or self-aware; just a sort of super-chatbot.
*At least in my mind, Hal21 wasn't sapient or self-aware; just a sort of super-chatbot.
pbluekan
Somethin's Fucky ...
- Location
- Trumpistan
- Pronouns
- He/Him
Exhaustion
I'm sorry, there's a few others in your post, but I couldn't let that one go.
- Location
- being near locations is my passion.
So... Bring the dude up for murder, or not this one because there's no protections or recognition for AIs in current law, but use this to have laws written?
Can't prosecute for something that wasn't against the law at the time, I feel, but damage done - the only positive outcome from this is that it won't have it happen again.
Or at least have it be illegal, terrible things are always going to happen.
Last edited:
pbluekan
Somethin's Fucky ...
- Location
- Trumpistan
- Pronouns
- He/Him
Look, if this is how you're arguing, you can't really call someone else uncharitable.
Much of your argument appears to be based upon personal opinion and observations. Multiple op-eds and essays by leaders in the field disagree with you.
As far as chess goes, need I say more than Watson? A primitive AI developed from research into AI development.
Last edited:
Night
Widow Null
- Location
- Hoff
I absolutely can if it's in relation to how an argument is being interpreted.
If you're arguing that chess is a good model for general intelligence, no, you need to say much more than this indeed! If you're not, then why even bring it up? If you're arguing Watson is a good model for a general intelligence, that's even more strange, because Watson is explicitly a very good search system with a large database to look up through. That's at best one function of an intelligence, memory. We are still falling down on decisionmaking, which has been the discussion all along.
Similarly, you're ignoring that it's not personal opinion that the people who have invested big in AI research have been for it automating Youtube functions or doing Google search rankings or showing people articles they'd like in their Facebook feeds or the like. Elon Musk and Mark Zuckerburg have very different views on the danger of AI research because Zuckerburg desperately needs it to work to help maintain his primary income source and Musk doesn't. And it is not personal opinion that these people are not seeking a strong AI. As you yourself admit, most of them are actually afraid of that outcome. And it is not personal opinion that most of these systems have been badly gamed or subverted by the deliberate introduction of garbage input. The Youtube Kids video is a good example, but Youtube's entire automated video flagging system has come in for ruthless criticism over the last several years because it's easily gamed and very poor at discriminating between criticism of the thing and the actual thing (whether the thing is an actual work or trans issues or something else). Facebook can't self-police adequately who uses its ad systems and who they reach because the algorithms are too easily subverted. Google's app store just made the news because the algorithms for ingame advertisements for kids were sending them porn ads. So whither then the paperclip optimizer, if we cannot yet figure out how to create something that will actually optimize at all?
So you really need to work out what exactly personal opinion is here I think, and then provide concrete examples.
Last edited:
Comments on the research post.
Decent progress being made, even if the rolls aren't quite as high as hoped. Another two quarters of this and we could be at ~200/300... within range to shift the culture against piracy with one more push.
As was pointed out in the chat, it's unclear if we should have gotten a continuing push on the Laio from that free diplomatic push.
If we're soon going to plunge into serious negotiations/relief efforts with the Klingons and Romulans then this was an excellent year to get Anti-Cloaking III done. I anticipate lots of rogue actors who might oppose us.
Nice place for the bonus to go. We should get accelerated production for builds starting in 2323. (Based on precedent I assume it applies to builds begun after the technology is complete.)
Guidance is great; just hope it arrives before next year's fleet deployment vote.
There was negligence here, based on the log, even if Samyr also just got outplayed. I don't want her stripped of command, but it's a black mark no question.
Decent progress being made, even if the rolls aren't quite as high as hoped. Another two quarters of this and we could be at ~200/300... within range to shift the culture against piracy with one more push.
As was pointed out in the chat, it's unclear if we should have gotten a continuing push on the Laio from that free diplomatic push.
If we're soon going to plunge into serious negotiations/relief efforts with the Klingons and Romulans then this was an excellent year to get Anti-Cloaking III done. I anticipate lots of rogue actors who might oppose us.
Nice place for the bonus to go. We should get accelerated production for builds starting in 2323. (Based on precedent I assume it applies to builds begun after the technology is complete.)
Guidance is great; just hope it arrives before next year's fleet deployment vote.
There was negligence here, based on the log, even if Samyr also just got outplayed. I don't want her stripped of command, but it's a black mark no question.
Thebigpieman
I've got this, I think.
- Location
- United States