「ジャンクの墓はレアメタルの山, アラスカ戦場跡迎撃線 II」 (Jyanku no Haka wa Reametaru no Yama, Arasuka Senjyouato Geigekisen II)
“The Graveyard of Junk is a Mountain of Rare Metals, Interception at the Site of the Alaska Battle II”

Just when I thought I was out, they pull me back in.

I bet you didn’t expect to see me here again so soon, eh? Takkun is out of town at the moment, so I’m taking care of Heavy Object in his stead. Or revisiting my old show, I guess. He’ll be back next week, so you’ve only got to put up with me here one more time. And maybe for part of the final impressions on the finale post. Damn you, Heavy Object! I can’t escape you! *shakes fist*

That’s Not How AI [Will] Work

I handed this off to Takaii because I felt like I was saying the same things each episode, and the most oft-repeated sentiment was that Heavy Object’s sci-fi is shifty as hell. Before you throw your hands up in disgust, know that I’m only mentioning this again because this episode was on a new level. The Juliet system isn’t anything close to something that deserves to be called artificial intelligence. When we start loading AI into our war machines, they’re going be at or near general intelligence, which has another name: human intelligence. They’re going to be as smart as we are, basically. (And about three seconds later they’re going to become superintelligent and leave us in the dust, but we’ll deal with that when it happens.)

Juliet, on the other hand, is just a program. An AI of general intelligence could do what a human would have done—examine the information, figure out what happened, and react accordingly. Or at the very least, retreat to a safe distance while it analyzes, instead of locking up. And I wouldn’t have a problem with them calling a glorified computer macro a combat AI if not for the other damage it did.

Ohoho: A Singer, Not A Fighter

The Juliet AI was a transparent move to make Qwenthur’s cockpit infiltration fail. Or at least, not succeed as well as it should have. This is something that could have been accomplished much easier by giving Ohoho-chan a gun in the cockpit, but that would have made too much sense. That transparent nature made it all the more cringe-worthy.

But the bigger problem is that Ohoho-chan isn’t piloting anymore. It’s no longer a fight between Milinda and Ohoho-chan. It’s a fight between Milinda and Juliet. And we don’t care about Juliet.

Look—conflict works best when it’s between parties we care about. We don’t have to like them both—no one likes the Empire in Star Wars, or those Empire wannabes in the new one—but we do need to care what happens. There are no stakes in this Alaska Part Two battle, or at least none that we care about. What made it work last episode was that it was Milinda and Rush, blasting each other to the soundtrack of a verbal catfight—but even that was taken away. Whyyyyy!

Plus, I found the mid-battle song more puzzling than amusing, though please understand that it’s a thin line. Usually absurdity is my thing, but it just didn’t strike me this time. Everything else I said would still stand, but I’d probably be talking about how the song made up for much of it, if it had made me laugh. Alas, it didn’t.

Qwenthur & Havia Make This Show

Milinda’s adorable crush and Frolaytia’s heavy objects may be the obvious draw, but Qwenthur and Havia make this show. They carry it all on their shoulders, and even if G-cup isn’t as heavy as we thought, that’s a lot on them. Qwenthur’s plan, triggered with the “I’m at my wits’ end” and completed with Havia’s help, was so much fun for the “All according to keikaku” feeling at the end. (Note: Keikaku means plan.) And it makes sense—they were tricking a kid, albeit a pretty smart one, but intelligence isn’t the same thing as wisdom. The Water Strider targeting system was smart, but best was planting the bomb, because you get a real sense that shouldn’t have worked—if the Intelligence Union was going to blow her up, why would they make the bomb visible? Why the long timer? But Qwenthur pushed hard enough and fast enough, Ohoho-chan fell for it, and suddenly he’d captured a slightly used Object. Take that!

Though obviously, he should have taken Ohoho-chan prisoner instead of letting her go. Then she could have joined their team. Can you imagine how smug Milinda would be when she finds out Ohoho-chan is a DFC? Or how much Ohoho-chan would still get under her skin in the future? Though I hope the first part still happens. That’ll be amazing.

tl;dr: @StiltsOutLoud – Qwenthur and Havia are carrying this show, and stealing enemy Objects as they do it #heavyobject 16

Random thoughts:

My first novel, Wage Slave Rebellion, is available now. (More info—now in paperback!) Sign up for my email list for a FREE sequel novella. Over at stephenwgee.com, the last four posts: Star Wars: The Force Awakens – The Introduction, What Star Wars: The Force Awakens did right, What Star Wars: The Force Awakens did wrong, and Star Wars: The Force Awakens – The Conclusion.

 

Preview

39 Comments

  1. I don’t know, I actually thought this episode worked pretty well. There are things throughout lots of sci-fi that get called AIs that are really just programs capable of analyzing and responding to situations. In many cases they are quite limited, and possess no intelligence, and are as stupid as a regular computer program: meaning when they encounter something that’s completely outside their ability to respond, they cough up an error and wait for someone to make things “normal” again.

    Juliet really was not needed to make Qwenthur’s infiltration fail. All Ohoho had to do to deal with him was move Rush, and with or without Juliet she could have done that and thrown him around the cockpit (like she did) without him managing to cut her throat. Because let’s be honest: if he could help it he wasn’t going to kill a girl who was clearly no older than 12, and was probably more like 10. Not if he thought he didn’t have to. And she knew it, and knew that she could use that hesitation, and his inability to handle an object’s acceleration, to protect herself.

    Putting all that aside, however, what I liked best was how Qwenthur essentially won Ohoho over to his side. Or at least a lot closer to it than she was before. The seed of doubt was planted due to that bomb, and it grew during that bit at the end. That doubt is always going to be there now. That fear. Do her commanders actually value her, or is she just a convenient tool for them to use to develop combat computers? Next time she gets in a cockpit will she really be safe? If she ever does something that slightly varies from the way they think combat should be handled, might they try to kill her again? These doubts and fears are going to be haunting her now, and she’ll be looking with suspicion at the people she once trusted.

    Conversely, the one person she’s most likely to trust is currently an enemy, yet he dealt fairly with her (so far as she knows), saved her life, and just generally proved himself to be a good person (again in her eyes) and a capable warrior in his own way. He’s impressed her, and effectively pulled her into his harem, albeit unintentionally. Of course, I suppose they all ended up there unintentionally…

    Whatever else one might say about this episode, this girl has moved a lot closer to being an ally than she was before, and that was what I wanted from this encounter. 10 out of 10: mission success.

    Wanderer
    1. An abundance of bad sci-fi doesn’t make new sci-fi (that’s also bad) any less bad. It just means they’re all sucking together. Or in this case, not doing their research all that well.

      You’re right that Juliet wasn’t needed to make his infiltration fail, which is all the more annoying that it was used. It would have been interesting if she had to keep moving (and occasionally juking) to keep him from being able to crawl over to her and interrupt her while she’s trying to fight. There would have been more tension in that! Instead we got crappy quasi-AI. Though it did set up the end of the episode, which as I mentioned, was awesome.

      I do think Qwenthur earned Ohoho-chan’s respect and trust (to a point). However, becoming allies is a totally different thing. Soldiers on the battlefields of WWI had more in common with their brethren in the trenches across the way than they did with their own commanders, and they knew it. And when things got grim (which was practically all the time), they hated their own leaders FAR more than the men they were killing. They blamed them for having them live out in those damn trenches, and march into the steel and flame of the enemy’s weapons.

      But they still fought, they still killed, and they still died. Maybe our generation wouldn’t do what they did. Maybe we would rebel. But the world of Heavy Object is supposed to be regressed in certain goofy societal ways. You could certainly make an argument that Ohoho-chan is an elite (as well as an Elite), and thus more likely to turn traitor because she’s powerful, influential, and valuable. But will she abandon her country just because some mechanic dealt fairly with her?

      Maybe. I hope so, just for the hilarious bantery chaos it would sew. I’m not sure it’s any more reasonable than anything else about this show, though.

      1. Just to clarify two points:

        I do think Qwenthur earned Ohoho-chan’s respect and trust (to a point). However, becoming allies is a totally different thing.

        That’s why I qualified it with “at least a lot closer to it than she was before.” I actually almost included something about how they’ll still undoubtedly fight if they meet each other’s groups in opposition on the battlefield again, but I believe the dynamic will have changed. I suspect she may be less eager to kill next time. I didn’t include that comment because my previous post was long and rambly enough without going off on more tangents.

        But will she abandon her country just because some mechanic dealt fairly with her?

        No no. If she abandons her country it’ll be because she believed they think she is nothing more than a tool: worth less to them than the computers they put into her objects. She thinks they already tried to kill her as soon as she showed the slightest hint of having outlived her usefulness, and now that that suspicion has been planted in her mind it’s not going to go away. After all, if they were willing to do it once, what’s to stop them from doing it again, and using a bomb or other device she can’t escape this time? What else might they have been doing while she hasn’t been paying attention because she trusted them? Doubt and suspicion and fear have been planted in her mind, and they’re only going to grow.

        That doesn’t mean I think she’s simply going to defect next time we see her. I just think she’ll be… less antagonistic, shall we say. Although, of course, she’ll still be sure to piss Milinda off, just because I’m sure she finds that hilarious.

        Wanderer
      2. I think she notice the lie when her underling talked about the backup and fail-safes. Ohoho-chan wouldn’t hit her subordinate if thought her position is in jeopardy, she to hit her for treating her like a child. She’s a lot angry at Qwenthur. As for why Qwenthur didn’t kidnap Ohoho-chan, it’s very likely that they will torture her for information or imprisoned. She doesn’t have a real reason to betray for country and ask for asylum so he didn’t push her in that direction.

        aichan
  2. I mean Qwenthur has plot armor but I still would liked if Milinda showed some concern towards Qwenthur when a normal guy who was getting hit by the G-forces of the Object. There’s time to be jealous afterwards.

    xxvan
    1. My only big problem with this episode. She sees him as far deep into enemy territory as you can get and what does she do? Shoot at him. Not a second spent on thinking he might be there to help her. That’s beneath Milinda. Glad the point eventually got through to her, but it took too long with their history. Should’ve been her first thought. She’s adorably jealous, not stupid. That’s those lesser tsunderes.

      Also hope we get to see some of her rampage over “Angelina” and “Julia”. Both the engagement ring reactions were great!

      Aex
  3. I find Qwenthur’s uselessness while inside the cockpit to be utterly infuriating. There is absolutely no way that the G-forces were so absurdly strong that he would be completely pinned to a wall while Ohoho-chan can still freely move her arms and hands. Either she is Goku-level strong or Qwenthur is a piece of paper. Just the G-forces required means that the Object is CONSTANTly accelerating at ALL TIMES, which would make it a speed demon on par with a certain yellow tentacled teacher in another current anime. Sure, the Object was also zigzagging which kept Qwenthur off his toes, but since the Juliet AI was controlling everything, Ohoho-chan should not have fared any better from the random sharp turns that she did not consciously control.

    ptolemaios00
    1. You don’t have to pull 9G’s to be unable to freely move. The 2-3G’s that a racecar driver experiences is enough to keep them firmly pressed against their seat. Ever taken a fast corner in you car at 45-60mph and felt yourself getting pushed to the outside of the turn. Plus, she was seated, accustomed to the g-force and we don’t know what kind of physical/genetic modification the elites are subjected to as well.

    2. Those G-forces are all caused by acceleration not just by high speeds. You may get pushed into your seat on an airplane, but after a while you can freely walk around. Thus my point on the CONSTANT acceleration required to hold Qwenthur onto the wall. Thus my point on the zigzagging creating those changes in acceleration and G-forces.

      ptolemaios00
      1. Objects top speed?

        500 KM/H?

        So, if she constantly accelerate in the G-force to push him on the walls.. how long can she do that?

        🙂

        Also, if all fails…

        … It’s Anime Magic!!!

        Worldwidedepp
    3. I don’t recall it being mentioned in the episode, but in the LN from which this episode comes from it is stated H.O. Elites are engineered\enhanced humans. Both Milinda and Ohoho-chan are supposed to be stronger and faster than most people in order to pilot. Of course, this enhancement is never shown outside the cockpit. Go figure.

      rh75
  4. Havia will be envious again, not only did he missed the Boobs of her superior Officer, no now he also missed the Ohoho boobs also 🙂

    Havia is on the short end of the stick, where the large one, do not care about 🙂

    Worldwidedepp
  5. That’s Not How AI [Will] Work

    I handed this off to Takaii because I felt like I was saying the same things each episode, and the most oft-repeated sentiment was that Heavy Object’s sci-fi is shifty as hell. Before you throw your hands up in disgust, know that I’m only mentioning this again because this episode was on a new level. The Juliet system isn’t anything close to something that deserves to be called artificial intelligence. When we start loading AI into our war machines, they’re going be at or near general intelligence, which has another name: human intelligence. They’re going to be as smart as we are, basically. (And about three seconds later they’re going to become superintelligent and leave us in the dust, but we’ll deal with that when it happens.)

    Juliet, on the other hand, is just a program. An AI of general intelligence could do what a human would have done—examine the information, figure out what happened, and react accordingly. Or at the very least, retreat to a safe distance while it analyzes, instead of locking up. And I wouldn’t have a problem with them calling a glorified computer macro a combat AI if not for the other damage it did.

    Having of taken a AI course and wrote a simple AI in the course of it, this is AI. Everyone imagines some super futuristic AI like Cortana or Skynet or whatever. The only difference between those and a simple AI such as Juliet is a much larger database, more complex routines and decision trees and a lot more code to make that all work. The search engine that runs google is a simple AI. The AI’s that are currently be testing by the US military among others are no more advanced than Juliet and won’t be for quite some time. Even something like Watson is pretty stupid as far as human decision making and learning goes and requires a mini datacenter to host it.

    Even if you stuck a Cortana or Skynet in something, it will still take hundreds of hours of learning and refining its database, decision trees and routines before it actually gets to a point of human learning.

    1. also this AI is made for quick Melee combat and looking the surroundings. In one swoop of attack, the AI would fire all weapons that come into range, take some workload off the shoulders of the Pilot. The AI here was more an Weapon and best shooting angel Supporting Programm.

      Worldwidedepp
      1. https://randomc.net/image/Heavy%20Object/Heavy%20Object%20-%2016%20-%20Large%2020.jpg
        The reason behind this attack, to uncover this wreak from Snow, is to put the “Water strider” remains into the Sensor range of the Enemy Object. Faking the AI to think there is the Water Strider, and the Target Lockon System was the “ice cake” to put the AI Calculation program into chaos

        in some sort they speculate into an self DDOS attack, putting the AI into an Deadlock without exit

        Worldwidedepp
      2. in a nutshell

        The Action fight and Weapons Sci-fi creator is a very good one, he really do his Homework. He “saved” this Show into B-Movie

        Remember the Mass Driver weapon last arc? Very good background research

        Worldwidedepp
    2. I think the same. Riding a horse is indeed a good comparison, if Juliet is at horse level, then indeed it can go in it’s own and learn something, but still a supervisor is needed. But even with a one the horse can shied in battle or after stepping one snake, our main characters just prepared a really big snake for that ohohoho horse.

      Botan
    3. @Smokex365

      AI, as a term, is a bit of a moving target in general parlance—what seems like AI at one point becomes business as usual after we get used to it. That’s why I opined that we’d need to develop artificial general intelligence before we start putting them in charge of our war machines (though Botan’s comment about it being like riding a horse is good).

      But talking about databases, decision trees, and routines is going down the wrong path. That’s a sophisticated program, not intelligence. Or, well—it depends on how you mean those, but if you’re talking about “If X, then Y” decision trees, then that’s limiting. We can’t possible program all contingencies, so it’ll always trip up eventually and expose itself as not really that smart.

      Instead, revolutionary AI will need to take advantage of some of the same kinds of learning that humans utilize, which is happening in some places (they’re not programming decision trees into most self-driving car software (from what I’ve read), they’re teaching them by having them observe humans driving) already. Which might be what you were saying with the hundreds of hours of learning, in which case consider this clarification.

      Though I should note that hundreds of hours of learning for an AI wouldn’t be a difficult feat. They don’t need to sleep, and especially if they’re networked together. Kowai.

      (All that said, I was mainly saying that I didn’t think Juliet qualified as more than a glorified program, which some people would call AI, and that’s semantics. Mostly the point was that it detracted from the viewing experience in other ways.)

      1. I don’t know about you, Stilts, but I’d rather not put your definition of an AI into a warmachine capable of deciding how wars end. Artificial General Intelligences are harder to predict. I mean, just imagine Skynet piloting an Object– that’d really end well for the humans, right?

        Beedle
  6. https://randomc.net/image/Heavy%20Object/Heavy%20Object%20-%2016%20-%20Large%2011.jpg

    Our Ohoho Girl, loved to tease. She is provoking the enemy Pilot with this Live Concert of her, demonstrating her that she is no real enemy, when she can put her under pressure while she has time to sing with her hands free.. She is try to offend her skills.

    Lucky for all, Ohoh girl here underestimated the enemy Object, well in the end she nearly was about to finish her. Thats because there was to short time for repairs, she would had won the “Pitstop fight”

    Worldwidedepp
  7. I am not surprised about Hohoho-chan since I got spoiled already.

    Indeed Stilts. He should have captured her. I would love to see the Princess’ and Havia’s reaction just like when Havia’s reaction about the boob fondling. She would make a nice ally and doesn’t need to capture Qwen since he’ll be her engineer as well.

    vaans02
  8. If anything, I would argue that Heavy Object’s depiction of AI was very accurate of what the situation with AI development is like today. The military is already testing and using computer programs that do very similar tasks. If you don’t consider that AI, that is fine. But colloquially, the term has begun to be used in such a manner (such as in descriptions of the programs running robots or self-driving cars). I don’t think you are correct in thinking that the first AIs we will load into war machines will be at or near general intelligence. We are nowhere near the point of achieving general intelligence machines (much less having the processing power for them). And while they may not be perfect, the advantage of even a simplistic AI system over a human being can be large. And in war, any advantage is taken and exploited.

    dmanatunga
    1. I more meant that we’d need more advanced AI before we give it as much control, but as Botan noted above, it’d be a bit like riding a horse, so even that’s probably inaccurate. However, I have a suspicion that we’re much closer to general AI than most people expect. Technology begets technology, after all, and the pace is constantly accelerating. The world will change more in the next 10-20 years than it has in the previous 100—and even that’s probably too conservative.

  9. Okay, as far as this AI thing goes, it’s worth putting this out there ’cause I think we’ve gotten a little too… er, spoiled when it comes to such things. When we think of AI these days, we tend to think of it in a very broad, human-esque sense, hence the term “Artificial Intelligence.”

    I don’t think that’s the way it’s going to be, at least not completely. When it comes to machines of war; yes, we’ll have AI with all the technical knowledge and analytical skills of a human, but it’s not going to analyze things and react the way we would. That would be stupid. There would be way too many unpredictable variables in such a thing, and when you’re in a war, that’s the last thing you want.

    It’s better to call such things Selective Artificial Intelligence, in that it’s obviously selective to a very specific thing and nothing else. It’s easier to think of Juliet like that.

    Obviously, this still has its flaws and doesn’t completely account for Juliet shutting down and leaving her Elite open to getting slaughtered (mayhaps that song system took up too much space?), but hey, whatever. Oh, are those overpriced Ohoho-chan plushies I see over yonder?

    Ryan Ashfyre
    1. Ah, but we already have selective artificial intelligence, in many arenas. As much as I said Juliet isn’t AI, what I meant is that she’s not advanced, revolutionary AI, much less general AI—she is a selective artificial intelligence. Which, as I said, is sometimes nothing more than a glorified computer program, but if a computer can do a sophisticated task better and/or faster than a human, that’s selective AI.

      Yes, you’re correct that general AI wouldn’t analyze things and react the way we would. They’ll be better at it. More processing power, better memory, and learning gathered from all the AI across the world? They’ll be much better. Plus, they won’t have any of those squishy glands feeding them illogical feelings, for better or worse.

      They won’t be any more unpredictable than trusting human soldiers or generals, and likely more predictable, if we do it right (and don’t Skynet ourselves). Saying we’ll stick to only selective AI tools of war is a failure of imagination, unless you’re thinking we won’t fight each other so much in the future, in which case that’s a triumph of imagination and I hope you’re right, haha

  10. I think the fact that Ohoho-chan wasn’t carrying a gun inside the cockpit can be excused in this instance. In-universe, infiltrating an Object was likely inconceivable. If memory serves, even Millinda doesn’t carry a gun.

    Anyway, did a quick wikia search and lookit what I found on the Heavy Object Wikia:

    A special characteristic of the Gatling 033 is the fact that it’s being used as a testbed in the Information Alliance’s development of unmanned Objects. The Legitimacy Kingdom’s Angelina List showed that current AIs are still very limited and cannot pilot an Object fighting at high speeds better than an Elite. However, the Information Alliance reached a simple solution. Instead of being completely unmanned, the Gatling 033 has both a strategic AI called Juliet and an Elite. While both Juliet and the Elite can pilot the Object on their own, to use the Gatling 033 at full specs requires Juliet to pilot and the Elite to deal with any mistakes or bugs that appear. Juliet’s complete control mode is still in testing, so the odds of misfires and erroneous judgments are still high. – Volume 3, Chapter 1, Part 12

    Well, I retract my earlier statement. Skynet Objects, tally-ho!

    Also, lol at Ohoho-chan being less a pilot and more a debugger…which, given the layout of her cockpit, does make a lot more sense now. o_O

    Beedle
    1. Well, she is capable of Manual drive the Object. She was about to take over, but Qwenthur came up with the fake Bomb.

      Well, he improved with this little time he had this Bomb. i do not ask where he had storage these Containers and this Digital Countdown Clock. No, i do ask about the Anime Magic infinity Bag space 🙂 Like in Log Horizon 🙂

      Worldwidedepp
  11. A bit of public understanding of the term AI vs the computer worlds use of the term is making for confusion.

    To the public AI is only a thinks like a human or better level of programing. As trying to duplicate what a neuron does in a simplified way takes a entire mainframe computer so we are no where close to the base 4, wave multi directional processing a brain can do. And clearly some of humans thinking ability comes from this considerably different way of computing. Plus the brain also uses chemical computer processing along with the neurons to add yet more layers to it’s processing. Data retrieval and reacting with simplified controls are way better on computers but there are still gaps. In example for the yearly AI robot contest no one has figured out how to get a robot to totally straighten it’s legs without crashing everything. The programer joked to the reporter that the programs starts to try to divide by zero (which is impossible) if they straighten the leg.

    Ghost in the shell deals with what is life with the true Artificial Intelligence as the public understands the term becoming a new invention. It’s fun when the combat robots the hero’s use who without others knowing it reach full AI use that to crash a human looking female robot who is not quite fully AI by asking questions that are paradoxes. But seriously the Major the main character becoming in effect immortal by being able to function on the net without a body for a time is a case study of the issues faced by someone who is not really human anymore, a total artificial part cyborg normally, even though she was born one.

    RedRocket

Leave a Reply

Your email address will not be published. Required fields are marked *