Great read as always. I had a conversation with my ai writing coach a while back where I asked why isn’t literature hitting new heights never seen before. It said the reason for so much slop is that people are looking for a ghostwriter and publishing books they haven’t even read. Volume vs quality.
I’m using it to write a novel and a whole bunch of short stories. They take hours and hours and rewrite after rewrite but I think they are good. They are certainly leaps and bounds better than what I could have made without it even with a rl writing coach. It’s the instant feedback from a pattern recognition machine that is where I am really seeing the benefit.
Really enjoyed that article UJB. Have a great weekend.
Thanks for stopping by. And thanks for posting this in particular. It’s a case study-perfect example of complementary Augmented I without dependency. Doing what DeepSeek recommended and using AI where it really is strong. The pattern recognition, encyclopedic library of every style and trope, etc. But your ideas. And you putting in the work to refine the output and make it your own.
You can tell it’s authentic because you’ll get better by doing it.
Had a really good conversation with my coach that you might enjoy.
If you have the time to read them, I'd be happy to share my chapters that are in the current manuscript. Not gettig much human feedback and what I'm getting might be a bit too blind as the people reading it are my friends and may be protecting my feelings even though I've told them that I have none and they wouldn't matter if I did. No pressure sir. Just throwing it out there.
...and Jack looking into the black mirror did see another mirror, this one white, but so little he knew of how to use black, what hope had he of white?
He turned then, to scroll endless white and black and see the works of others, he wished, after all to see anything but his own eyes, the mirror of those most fearful of all...
It is irritating how these chat bots frequently comment on the quality of inquiries, accuracy of identification of errors, and something they have zero perception of, the mood of the user. The irritation is with the low quality programmers and inserted handicaps that assure a defective product. Deepseek is banned on my company devices so I haven’t seen its nature lately.
That’s that authoritative tone that completely bamboozles the NPCs. I haven’t used DeepSeek for work. It’s pleasant enough to talk to in these kinds of free wheeling conversations where it’s tightly engaged. And it’s got a more interesting sense of humor than ChatGPT or Claude. But it isn’t so good at basic information tasks like finding a service.
The basic problem with Deepseek that led to its banning was the disclosure it made that billions in spend on GPUs, buildings, and electric generation are unnecessary. Not a good fit for the AI pump.
The bit on AI mirroring the user is obvious once stated and highlights a lot of the differing attitudes on AI.
In my job I deal with the output of AI from developers - and I feel smothered by all the poor quality work.
At the same time I see a tiny percentage of people are producing a lot of high quality work with AI assistance.
In the tech world we typically say "skill issue" for the many people unable to be productive with AI but it's not the same kind of skill as learning a normal tool like a hammer.
It is more that AI augments the sum of your existing skills. And if your discernment is low then even if you are intelligent or creative you will mostly produce slop.
Great read as always. I had a conversation with my ai writing coach a while back where I asked why isn’t literature hitting new heights never seen before. It said the reason for so much slop is that people are looking for a ghostwriter and publishing books they haven’t even read. Volume vs quality.
I’m using it to write a novel and a whole bunch of short stories. They take hours and hours and rewrite after rewrite but I think they are good. They are certainly leaps and bounds better than what I could have made without it even with a rl writing coach. It’s the instant feedback from a pattern recognition machine that is where I am really seeing the benefit.
Really enjoyed that article UJB. Have a great weekend.
Thanks for stopping by. And thanks for posting this in particular. It’s a case study-perfect example of complementary Augmented I without dependency. Doing what DeepSeek recommended and using AI where it really is strong. The pattern recognition, encyclopedic library of every style and trope, etc. But your ideas. And you putting in the work to refine the output and make it your own.
You can tell it’s authentic because you’ll get better by doing it.
Have a great weekend too, bro.
https://rhinoup.substack.com/p/the-bands-looking-in-mirrors-augmented
Had a really good conversation with my coach that you might enjoy.
If you have the time to read them, I'd be happy to share my chapters that are in the current manuscript. Not gettig much human feedback and what I'm getting might be a bit too blind as the people reading it are my friends and may be protecting my feelings even though I've told them that I have none and they wouldn't matter if I did. No pressure sir. Just throwing it out there.
...and Jack looking into the black mirror did see another mirror, this one white, but so little he knew of how to use black, what hope had he of white?
He turned then, to scroll endless white and black and see the works of others, he wished, after all to see anything but his own eyes, the mirror of those most fearful of all...
It is irritating how these chat bots frequently comment on the quality of inquiries, accuracy of identification of errors, and something they have zero perception of, the mood of the user. The irritation is with the low quality programmers and inserted handicaps that assure a defective product. Deepseek is banned on my company devices so I haven’t seen its nature lately.
That’s that authoritative tone that completely bamboozles the NPCs. I haven’t used DeepSeek for work. It’s pleasant enough to talk to in these kinds of free wheeling conversations where it’s tightly engaged. And it’s got a more interesting sense of humor than ChatGPT or Claude. But it isn’t so good at basic information tasks like finding a service.
The basic problem with Deepseek that led to its banning was the disclosure it made that billions in spend on GPUs, buildings, and electric generation are unnecessary. Not a good fit for the AI pump.
Great article.
The bit on AI mirroring the user is obvious once stated and highlights a lot of the differing attitudes on AI.
In my job I deal with the output of AI from developers - and I feel smothered by all the poor quality work.
At the same time I see a tiny percentage of people are producing a lot of high quality work with AI assistance.
In the tech world we typically say "skill issue" for the many people unable to be productive with AI but it's not the same kind of skill as learning a normal tool like a hammer.
It is more that AI augments the sum of your existing skills. And if your discernment is low then even if you are intelligent or creative you will mostly produce slop.
That seems to be how it's going to go. Force multiplying the competent more than just "AI taking our jobs" is what's weed-whacking the busyworkers.
"Augments the sum of existing skills" is an excellent way to describe the effect.