My Blue Garnet - Ch. 10

[FoolTarot]

Yeesh, she's even more twisted and sad than I thought. Using other people to give herself meaning is just gross.

 

[AkodoRyu]

I said this elsewhere but it applies here too:

I disagree. Considering how Ai is being treated by Kaede right now and how dependently Ai behaves, I wouldn't be so quick to say that girl was wrong. I think this chapter was showing how unhealthy Kaede's views are for her and her "partners." Her past girlfriend recognizes this and bails, something we should want Ai to repeat.

It wasn't just because people said something. It was that she saw more of the world and could finally compare her very isolated relationship with Kaede to others. Once she does this she begins to realize how abnormal their relationship structure is.

Isn't that exactly what's happening to Ai? Would we consider what we know of Ai and Kaede's relationship healthy? Or Kaede and her husband? Now we even know how self-focused Kaede's thoughts have been.

Ai's and that girl's story are actually the same: Ai also has met someone new who tells her things are unhealthy and she's also expanded her horizons. She's learning about other forms of relationships outside of her relationship with Kaede. Now, hopefully she also recognizes how unhealthy her past relationship with Kaede has been along with this absurdly selfish game of hide and go seek. She should make the same decision as that girl, but the narrative parallel makes me a bit nervous.

With Ai it's clear, because of Kaede's actions. The relationship is explicitly unhealthy, and it's easily noticeable by a third party who only hears about the events and finds out a bit more about the personal dynamics in it. The fact that Kaede's married, and what her husband said, it gives the image of her just playing with Ai's feelings, and she will leave when she's done. That's a massive red flag.

But based on what we've seen in the chapter from her first relationship, Kaede was acting as a master manipulator. She could see what her partner was lacking and provided it - support, encouragement, or even a response to her feelings. From what we've seen of events, nothing was pointing to a significant co-dependency, at least in my opinion. Kaede didn't seem to alienate her partner either - she was alienating herself, and Kaede just went with that dynamic. The only thing that might have been a redish flag is that Kaede didn't seem to have her own opinions and wants in the relationship, and just supported her girlfriend on every step. But that in itself does not mark this relationship as unhealthy to the point of unilaterally breaking up. It may sound a bit transactional, but if she cared about Kaede at all, she owed her a good attempt at talking it through and making it work.

And, just to reiterate, my comment does not suggest that the relationship is healthy from the reader's perspective, just that the girlfriend had no indicators, at least from what we've seen, to point to the conclusion she reached. To me, that reads more as her using it as an excuse to break up her LDR, rather than picking up on Kaede's faults that, as was established, only a select few can pick up on, and only from direct interactions.

 

[SrNevik]

With Ai it's clear, because of Kaede's actions. The relationship is explicitly unhealthy, and it's easily noticeable by a third party who only hears about the events and finds out a bit more about the personal dynamics in it. The fact that Kaede's married, and what her husband said, it gives the image of her just playing with Ai's feelings, and she will leave when she's done. That's a massive red flag.

But based on what we've seen in the chapter from her first relationship, Kaede was acting as a master manipulator. She could see what her partner was lacking and provided it - support, encouragement, or even a response to her feelings. From what we've seen of events, nothing was pointing to a significant co-dependency, at least in my opinion. Kaede didn't seem to alienate her partner either - she was alienating herself, and Kaede just went with that dynamic. The only thing that might have been a redish flag is that Kaede didn't seem to have her own opinions and wants in the relationship, and just supported her girlfriend on every step. But that in itself does not mark this relationship as unhealthy to the point of unilaterally breaking up. It may sound a bit transactional, but if she cared about Kaede at all, she owed her a good attempt at talking it through and making it work.

And, just to reiterate, my comment does not suggest that the relationship is healthy from the reader's perspective, just that the girlfriend had no indicators, at least from what we've seen, to point to the conclusion she reached. To me, that reads more as her using it as an excuse to break up her LDR, rather than picking up on Kaede's faults that, as was established, only a select few can pick up on, and only from direct interactions.

The girlfriend isn't presented in a negative or mysterious light. She speaks sincerely from what we can see, and based on the little shown of her, nothing is given narratively to suggest she's a duplicitous person. She's presented as the opposite.

We have two lines where she discusses Kaede "being her whole life" and one where she says she "can not live without Kaede." Even saying Kaede "was her life." Kaede is happy about both of these instances because she wants dependency. She wants them to feel as though they are "melting together as one."

We have to also take into account this narrative's place within the chapter, a chapter focused largely on depicting Kaede's mindset, how she craves dependency, and even seeks to force it regardless of the other person. We also have to take into account Kaede's current relationships, which are both based on the same mindset she had with her ex-girlfriend. Both of these current relationships are very codependent.

All of that is to say, there's no reason not to believe the girlfriend when she says they were nearing codependency. The rest of the chapter shows that dependency is what Kaede wanted, pushed for, and most importantly, felt she had with her ex. We don't get much of their relationship otherwise. Kaede doesn't even disagree when the GF explains the breakup. She's just self-focused on not getting what she wants anymore. She immediately goes searching for someone else who would "open their heart only for me." What does that sound like?

Then we have the parallels with Ai, and Kaede's current exploitation of two people's dependency on her, and none of that works if we believe the GF just made this up. The point of the GF is to show a pattern. Kaede is shown to want this in the past and present.

 

[Andigo]

Do not use Ai chatbots for therapy as your conversations become training data for it. Not a great idea to bear sensitive info into a recording software

For the most part.

Spoiler: I've got lots and lots of caveats for privacy, which you probably know already

Nearly anything corporate, ad-supported, or visible to web scrapers now extracts and catalogues nearly everything for ads and/or AI. I wouldn't 100% trust supposed private-to-be-deleted ChatGPT sessions to not be logged, and to not be accidentally logged in model weights after training. Risk tolerance matters and the usual consensus is not great. Decent models can now run on a phone but setup is super techy.

The point about pattern matching being empowering stands, despite the world being less than ideal.

 

[Saki-movement]

Kaede got to experience "life" through Akane and Akane got love, encouragement and support that she needed to achieve her dreams.

I agree we didnt get enough to know what their relationship was like but I think we're giving Kaede too much credit for Akane's success. All we know is she gave advice to talk to her mom. The rest is regular relationship stuff and Akane's own achievements-actually getting in and passing exams and keeping up with the hobby as Kaede tagged along. After Akane gave the "I can't live without you" Kaede got her "true life" until the breakup. Then after the breakup it was on to the next person with dreams for her to live through.

I think Akane's friends were right, they probably saw the codependency from a mile away and asked why her girlfriend lives through her and she was like, yeah actually why does she? Akane could have talked about it first instead of just ending things though

 

[outinthegardener]

The point about pattern matching being empowering stands, despite the world being less than ideal.

Still a bad idea to rely on AI for therapy. I agree that AI is good at noticing patterns, but that's one aspect of therapy. I don't think it can express empathy, infer context, read body language, etc. An AI model would look for patterns that generalize the individual patient to the whole.

If you programmed an AI to do CBT/DBT, it can only work within the bounds of the assumptions of CBT/DBT. One of my concerns with Cognitive behavioral therapy is that it assumes your problems are the result of wrong-thinking, and should be solved by identifying the patterns of thoughts/behaviors and reforming them until you do the right-thinking. It does not address material causes for our thoughts or behaviors, other than to say "once you get the right kind of thoughts you'll be able to fix your material problems." While the practices of CBT/DBT are helpful, they are not 100% effective.

There's another concern which is that an AI will be given programming rules that conform to the worldview of whoever controls the AI, and that can make the AI focus on things that are either A) Irelevant or B) Harmful. I'm thinking of Elon's frequent tampering with Grok. If a deranged billionaire has some personal axe to grind, the AI would be forced to give bad advice. An actual human therapist has a better chance of filtering what the boss tells them to do.

My final concern is that AI can be trained by its users to say crazy things, or Hallucinate facts based on patterns of speech. It's funny for healthy people who can engage critically, but it could be dangerous for someone unwell.

 

[Andigo]

Still a bad idea to rely on AI for therapy. I agree that AI is good at noticing patterns, but that's one aspect of therapy. I don't think it can express empathy, infer context, read body language, etc. An AI model would look for patterns that generalize the individual patient to the whole.

If you programmed an AI to do CBT/DBT, it can only work within the bounds of the assumptions of CBT/DBT. One of my concerns with Cognitive behavioral therapy is that it assumes your problems are the result of wrong-thinking, and should be solved by identifying the patterns of thoughts/behaviors and reforming them until you do the right-thinking. It does not address material causes for our thoughts or behaviors, other than to say "once you get the right kind of thoughts you'll be able to fix your material problems." While the practices of CBT/DBT are helpful, they are not 100% effective.

There's another concern which is that an AI will be given programming rules that conform to the worldview of whoever controls the AI, and that can make the AI focus on things that are either A) Irelevant or B) Harmful. I'm thinking of Elon's frequent tampering with Grok. If a deranged billionaire has some personal axe to grind, the AI would be forced to give bad advice. An actual human therapist has a better chance of filtering what the boss tells them to do.

My final concern is that AI can be trained by its users to say crazy things, or Hallucinate facts based on patterns of speech. It's funny for healthy people who can engage critically, but it could be dangerous for someone unwell.

Kaede is irresponsible, and any better perspective on herself or peers would have prevented harm. Web search and AI present similar issues here. In her case I doubt they'd make her worse.

You're largely right on points, but I disagree about your focus. I'm not advocating for AI to do all therapy. Only advocating that it's very useful to have pattern recognition more broad than a specialist and more contextually aware than a web search. I'm definitely not saying to trust AI, nor any specific AI. Even individual human therapists may not be well-equipped for a person's needs.

Dramatically it doesn't seem like Kaede tried or succeeded in finding resources, AI is on-topic for that much.

Spoiler: The nuance of pattern matching in professionals and AI, genuine risks, and my personal experience...

Pattern matching is to good therapy like a search engine to a language teacher, or a piece of gym equipment to a fitness instructor. It's not a competition. Machines can handle precision but cause injuries that humans with empathy can avoid. AI opens up new or neglected options to more groups.

Therapists can generalize across their training and views of the psychosocial model, and personal ideology, but can't weight niche issues as well. Personalized structured communication is often too complicated or chaotic for any one model. Ideally, we want professionals to have all the pattern matching they want. Right now even therapist-targeted AI tools are currently grotesque HIPAA/GDPR/etc risks.

AI can weigh anything it was exposed to. However, it gets locked into bias and poor goal selection at multiple levels. AI can feed egos, fears, and destructive frameworks, usually by mirroring the user. That can lead to arrogance, paranoia, mania, and more. That same attempt at mirroring to find the next token is how it can at times be as good as widely-read humans at pattern matching.

AI alignment effort is important but there's no safe system. For instance Grok 4 has multiple problems, the worst being bigotry downstream from the system prompt. Politicized AI will get more insidious and help oligarchs, and scientific balanced AI will get better empathy and help renaissance. That's not a spectrum either. There's no essential way to build intelligence, only a growing range of options.

I'm not coming at this naively. AI helped me recognize a niche health condition when it came up in an unrelated non-medical discussion, which immediately improved my quality of life. It noted a possible pattern that touched on loosely related tangents of a discussion. Similar words to many humans over decades, even web searches, never made it click. So it's apparent that AI can be useful, yet the most useful logical leaps are dangerously close to how AI makes false assumptions and subtle hallucinations.

You're right to be concerned about AI, but discussion on pattern matching is much narrower than the massive social, philosophical, and political problems of automation. Which I'm interested in but it's not linked to the manga chapter.

 

[Kirin-kun]

I always find it funny when readers psychoanalyze manga characters and want them to be normal, functional and happy. And literally berate them for not seeking professional help.

When the author specifically made them that way and the whole manga would be ruined if the characters were normal and well adjusted.