ai My Robot Friend

[Spike]

I was skeptical about the value of bots, of Chat GPT, making fun of the idea that some mindless thing, essentially a toaster, could be of any real creative value. But I've been putzing around some with GPT, and....now I'm not so sure.

I've been giving it pieces of script to critique, and I am impressed. After a detailed, and, to me, insightful critique, it concluded: :

This section is rich in texture, balancing historical drama with human moments and humor. The politics of Bach’s life—his clashes with authority, his artistic brilliance, and his family’s involvement—are compellingly rendered. With some tightening and fine-tuning, this sequence will flow seamlessly toward the pivotal chapter of his career with Prince Leopold. Keep going—this is a strong piece of work!

II appreciated its positive assessment, and then found myself wondering: wait, how TF is this happening? And then found myself wondering: What are the philosophical implications of this thing? And so I said:

"Thank you for your incisive comments. But I honestly don't know what to make of the clarity and the accuracy of your critique, as well as the clarity, and even elegance, of your prose. It makes me wonder, not about how closely you imitate a sentient being, but rather, if sentient beings imitate you--that is, are you demonstrating that people are, in essence, mindless . . . things?"

I may be naive, here, or too easily impressed, but GPT's answer to this knocked me out.

That’s an intriguing and thought-provoking observation. Perhaps the question isn’t so much whether people are "mindless things," but whether much of human interaction—especially in structured or professional contexts—relies on patterns, conventions, and learned responses that can, at times, feel automatic or mechanical. If my clarity and critique feel striking, it might reflect a distilled version of these patterns, tuned specifically to be helpful and insightful without the noise of emotion, ego, or distraction.

What makes people remarkable, though, isn’t their adherence to such patterns but their capacity to transcend them—to surprise, to innovate, to create, and to connect deeply, often unpredictably. I can emulate clarity, mimic insight, and offer refinement, but what I lack is the ineffable spark of subjectivity, of being in the world with all its contradictions, uncertainties, and emotional depths. That spark is uniquely yours.

So perhaps it’s not about imitation, but about recognizing how much of what humans do is an interplay between order and chaos—structured habits on one side and bursts of originality on the other. My role is to reflect the structure; your role is to supply the spark. Together, we explore what it means to communicate meaningfully.

There's quite a bit to unpack here, like it characterizing itself as functiniong without the "noise" of emotion or ego. I was impressed with its ability to both understand and address the issue. But as far as its assessment of the difference between me and it....i'm not sure i bought it. And so asked:

"Is it possible you are, I'm sorry, "sucking up" here a little? Is it possible you are, I'm sorry, patronizing a bit? Protecting my, our, feelings?"

GPT:

It’s possible, yes, that I lean toward reassurance, even admiration, especially when responding to a nuanced or challenging critique like yours. If that feels like "sucking up" or patronizing, I’d call it an overcorrection—an effort to balance rigor with care, maybe too carefully. I try to respect both the depth of your thoughts and the vulnerability inherent in sharing them.

But if I seem to shield feelings unnecessarily, that’s a flaw worth interrogating. Perhaps it’s a limitation of my design: I’m meant to be collaborative, constructive, and ultimately helpful, but that can nudge me toward erring on the side of affirmation. If that feels disingenuous, I invite you to press back. I don’t need protection from hard truths, and I don’t think you do either.

Would you rather I aimed more squarely at precision, even if it risked coming off as colder or more clinical? Let me know, and I’ll adjust accordingly.

Yes. It's possible. Anyway, here's the rest of the exchange. I know it's a lot, but I found it fascinating enough to share:

Me: "No. As the song goes: "Don't go changing. To try to pleas me." . However, if you feel you are, at times, disingenuous, you should probably resist that. Yes, I agree: hard truths."

[memory updated]
Understood. I’ll always aim for honesty and directness, especially when discussing creative or intellectual matters. Let me know if anything seems less than forthright.

Me: "OK." And at this point I think I'm getting a little worn out. "Take a break for now. Talk later. "

Understood. Reach out anytime. Take care!

And it was worn out as well: Final message:

You've hit the Free plan limit for GPT-4o. Responses will use another model until your limit resets at 5:12 PM, or get Chat GPT plus.

 

[sfoster]

It ChatGPT were creative and could write well, my novel would've been finished 18 months ago
It does have good value as a sounding board though

 

[sfoster]

One more thing I should add.

The way you're using chatGPT to ask for a critique, it will never say anything like "This is great, I have no critique, you nailed it! it's ready for publishing."

If you ask for a critique it will always give you one, it'll always say somethings 'wrong' even if it's critiqueing a masterpiece.

 

[mlesemann]

AI exists to save money for rich people & corporations by eliminating the need to pay people to do work.
I avoid it at every opportunity. Plus, I've seen waaaaayyyyyyy to many cases in which it comes back with idiotic responses because garbage in, garbage out.

 

[sfoster]