Self-editing using AI

Lets get back to the concern that an Al editor pushes you towards their preferred style. It does. This doesn't mean you need to follow it. I can only describe how it worked for me.
When I got a suggestion, e. g. to show some emotion instead of telling, two things happened. Since the AI did not tell me how the show should be, I was forced to find a way myself. And I did in my own voice.
However the other effect was more important. In many cases I rejected the suggestion, which caused me to consider why. Effectively answering the question why I wanted my text here to sound this way. And this is one of the key questions when one tries to find one's own voice. I then fed my reasoning back to the AI. The reaction I got was: "You got a point here, my suggestion was not optimal, I will adjust how I make suggestions" And Bingo, suddenly it was me pushing the AI towards my style, not the other way around.
I think casual users are not aware to what degree you can tune an AI to your needs.
 
There are two practical differences between using an AI for feedback and a human, whether professional or amateur.

The AI will always tend to homgonize things. You have some control, if you take it, towards the point at which it is pushing you, but an LLM based AI will never be creative, just ,mirroring what it has said. I can see some advantages in that for beginning writers, before they start developing their own voice.

The second is the moral issue. People try to anthropomorphize how LLMs work. They amalagamate information they extract from the world. People do the same thing, but that simple statement is where the similarity ends. The way our minds work, we extract the essence out of a work, its meaning, its style, its voice. Then we merge out understanding of those. An LLM never has that abstraction; it simply cannot do that. It retains statistical patterns out of the original in a way humans will never be able to. It allows the LLM to regurgitate the original in a way that humans never could do. It just regurgitates mixtures of what is has consumed, nothing original. So 100% theft. For many of us, that reality is a game breaker for using any LLM based technology for anything.
I think you give humans far too much credit for an ability to "invent" new things. You would think Albert Einstein is a genius for developing his Special Theory of Relativity out of thin air, something no one had ever thought of before him.

But if you look back at Einstein's history before 1905, and even further to earlier discoveries by others, you might see that his theory was an amalgamation of work by others. He had Galileo's Theory of Relativity and the Laplace transforms as tools, and Einstein knew about the speed of light being an upper limit. Einstein had the particularly unique combination of math and physics knowledge he extracted from the world to amalgamate at that time. He debated the topic with others and eventually pulled the pieces together with the math to show it.

There was an old example of genius, if I remember correctly: "A thousand monkeys on a thousand typewriters given a thousand years will eventually type out the novel War and Peace." The threat to human authors is that at computer speeds, the AIs can sift through all of the free content on the Internet (as we humans do) and won't need a thousand years to extract and amalgamate the words any one human might proudly think is their own work of genius.
 
I think you give humans far too much credit for an ability to "invent" new things. You would think Albert Einstein is a genius for developing his Special Theory of Relativity out of thin air, something no one had ever thought of before him.

But if you look back at Einstein's history before 1905, and even further to earlier discoveries by others, you might see that his theory was an amalgamation of work by others. He had Galileo's Theory of Relativity and the Laplace transforms as tools, and Einstein knew about the speed of light being an upper limit. Einstein had the particularly unique combination of math and physics knowledge he extracted from the world to amalgamate at that time. He debated the topic with others and eventually pulled the pieces together with the math to show it.

There was an old example of genius, if I remember correctly: "A thousand monkeys on a thousand typewriters given a thousand years will eventually type out the novel War and Peace." The threat to human authors is that at computer speeds, the AIs can sift through all of the free content on the Internet (as we humans do) and won't need a thousand years to extract and amalgamate the words any one human might proudly think is their own work of genius.
I fully understand that people do not create things out of the blue. They are building upon other people's work. Extending what has been done minisculely. Good work done by people does this naturally. For most of us, on a good day, we take the input from earlier things and add them together, bringing just a touch of ourselves, a touch of something new to it. 1+1 =2.01 or something. For a genius, whether Einstein or Picasso, it might be 1+1=2.2 or something.

But LLM's simply cannot do that. In fact they always lose something. A good day is 1+1=1.99. There is never any new. Never anything created. This is something fundamental to LLM's (and some related AI technology) but not all AI or all computer technologies. The Logic Theorist (Herb Simon) in the 50's could do things LLM's can't now.

All LLM's do is recombine regurgitations. There is never a single shred of new in there. It's just not the way they work. I can point you at any number of in depth articles going over this for the big picture or the detailed for specific cases, like LLM's being unable to do even minimal logical inferences. If they have not already seen the inference made, they cannot make it themselves. Not possible in the current manifestation. Fundamental to the entire approach.

You bring up the monkeys on typewriters bit. The problem with that one is the people who said didn't do their math. You could have a trillion moneys and they would not write a short story in many many lifetimes of the universe. Combinatorics is a bitch, but she rules.
 
I fully understand that people do not create things out of the blue. They are building upon other people's work. Extending what has been done minisculely. Good work done by people does this naturally. For most of us, on a good day, we take the input from earlier things and add them together, bringing just a touch of ourselves, a touch of something new to it. 1+1 =2.01 or something. For a genius, whether Einstein or Picasso, it might be 1+1=2.2 or something.

But LLM's simply cannot do that. In fact they always lose something. A good day is 1+1=1.99. There is never any new. Never anything created. This is something fundamental to LLM's (and some related AI technology) but not all AI or all computer technologies. The Logic Theorist (Herb Simon) in the 50's could do things LLM's can't now.

All LLM's do is recombine regurgitations. There is never a single shred of new in there. It's just not the way they work. I can point you at any number of in depth articles going over this for the big picture or the detailed for specific cases, like LLM's being unable to do even minimal logical inferences. If they have not already seen the inference made, they cannot make it themselves. Not possible in the current manifestation. Fundamental to the entire approach.

You bring up the monkeys on typewriters bit. The problem with that one is the people who said didn't do their math. You could have a trillion moneys and they would not write a short story in many many lifetimes of the universe. Combinatorics is a bitch, but she rules.
I still think you're missing my point, that Einstein did not ADD something new, but brought together the correct and then available pieces.

As for your own genius in creating a new and unique story, don't underestimate the amalgamation of all of your past experiences and those tidbits of "forgotten" facts which shaped you to be who you are today.

Many people argue and insist that the other person is "WRONG!" I come from a different view, knowing that my past experiences and long forgotten knowledge is merely different from theirs. So, I never insist that the other person is WRONG. They merely come in with a different set of variables which shaped their opinion, ... so they're not wrong, and at best operating from an incomplete set of data.

In my earlier post about building fiber-channel disk arrays and data replication, while others obfuscate the overwhelming complexity of the Internet as "The Cloud", I see the small droplets of water within it and look for the smaller disk clusters in those RAID-10 arrays which hold the replicated info. That doesn't make others wrong, merely not seeing the details as I do.

EDIT: Yes, I am very much a "geek"! It comes across in my stories with my favorite MMC admitting it.
 
Would anybody not use AI as a labour saving tool? I notice lots of errors in professionally edited works; would you not want your work edited to the highest possible standard?

It’s easy and painless.



Prompt:

If a gave you some writing of mine, could you go through it and ensure the spelling and punctuation are compliant with the Chicago Manual of Style?

Response:

Absolutely, I’d be glad to help with that. Just drop in the text you’d like me to review, and I’ll go through it with a fine-toothed editorial comb, checking spelling, punctuation, and stylistic consistency according to the Chicago Manual of Style. Whenever you're ready, send it over.
 
Some people insist they'll never use technology.

I guess they see themselves as sitting in a cave and eating nuts and berries or tearing at a raw piece of meat with their fingernails. Maybe they're superior to other animals because they picked a better cave.

We all use technology.
 
I still think you're missing my point, that Einstein did not ADD something new, but brought together the correct and then available pieces.

As for your own genius in creating a new and unique story, don't underestimate the amalgamation of all of your past experiences and those tidbits of "forgotten" facts which shaped you to be who you are today.

Many people argue and insist that the other person is "WRONG!" I come from a different view, knowing that my past experiences and long forgotten knowledge is merely different from theirs. So, I never insist that the other person is WRONG. They merely come in with a different set of variables which shaped their opinion, ... so they're not wrong, and at best operating from an incomplete set of data.

In my earlier post about building fiber-channel disk arrays and data replication, while others obfuscate the overwhelming complexity of the Internet as "The Cloud", I see the small droplets of water within it and look for the smaller disk clusters in those RAID-10 arrays which hold the replicated info. That doesn't make others wrong, merely not seeing the details as I do.

EDIT: Yes, I am very much a "geek"! It comes across in my stories with my favorite MMC admitting it.
Then I fundamentally disagree with you on your point. People do add things that do not come from any experience. It is why we advance in our knowledge. Put a group of LLM's together, they will never advance. That's demonstrable. This is not something unique to humans. As I said, Simon had a problem doing this seventy years ago. But LLMs cannot do this. They are purely recombinations of existing pieces. You cannot simplify a complex theory with this. You seem to be hung up on Einstein. I believe there were insights he had, new simplifications he realized. Those would not be possible for an LLM. Not would the advances of Euler or Poincaire or Hilbreth or any number of other great minds in math. Or look at artists or writers. Or philosophers. LLMs cannot and will never be able to do any of these things.

I actually fully expect true AI to emerge some day. I just don't be around to see it. Maybe someone will make a break through tomorrow. Maybe they already have, I just don;t know about it. Maybe Frist's work. All I know is LLM have already had their day in the spotlight. They have uses. But they also have limitations. Tebru and Mitchell coined the term stochastic parrots for them. It;s a good term, other than real parrots are MUCH smarter than any LLM will ever be.

And be aware that almost all their tricks are simply by stealing work from others. In a way that humans are not capable of doing. We are influenced by what we experience. They regurgitate it.
 
Then I fundamentally disagree with you on your point. People do add things that do not come from any experience. It is why we advance in our knowledge. Put a group of LLM's together, they will never advance. That's demonstrable. This is not something unique to humans. As I said, Simon had a problem doing this seventy years ago. But LLMs cannot do this. They are purely recombinations of existing pieces. You cannot simplify a complex theory with this. You seem to be hung up on Einstein. I believe there were insights he had, new simplifications he realized. Those would not be possible for an LLM. Not would the advances of Euler or Poincaire or Hilbreth or any number of other great minds in math. Or look at artists or writers. Or philosophers. LLMs cannot and will never be able to do any of these things.

I actually fully expect true AI to emerge some day. I just don't be around to see it. Maybe someone will make a break through tomorrow. Maybe they already have, I just don;t know about it. Maybe Frist's work. All I know is LLM have already had their day in the spotlight. They have uses. But they also have limitations. Tebru and Mitchell coined the term stochastic parrots for them. It;s a good term, other than real parrots are MUCH smarter than any LLM will ever be.

And be aware that almost all their tricks are simply by stealing work from others. In a way that humans are not capable of doing. We are influenced by what we experience. They regurgitate it.
Gene Rodenberry's "Star Trek" was described as "Wagon Train in the stars." And if you take a critical look at "Superman", how much was he inspired by the Bible, a sky-sent savior of the world.

We all amalgamate and regurgitate. Even the greatest philosophers were inspired or influenced with something they saw and heard in their past.

Think of it as Sir Issac Newton sitting under an apple tree and being hit in the head by a falling apple. It drives some to question "What made the apple move in that direction?" The advance of human knowledge (in my opinion) is the right person with the right experiences and knowledge in the right place and right time to regurgitate it.

But you are not wrong!
 
Put a group of LLM's together, they will never advance. That's demonstrable.
So what? A table saw doesn't create anything on its own, and never will.

It's a force multiplier that allows us to do more, with less effort. You can make things by hand, and people do, but that doesn't meant that using the tool is cheating.

Using a tool to help you is not a moral failing.

And be aware that almost all their tricks are simply by stealing work from others
Would you feel any differently if the inputs were ethically sourced?
 
So what? A table saw doesn't create anything on its own, and never will.

It's a force multiplier that allows us to do more, with less effort. You can make things by hand, and people do, but that doesn't meant that using the tool is cheating.

Using a tool to help you is not a moral failing.
In the context of AI as it exists today, using a tool based on theft is a moral failing. using a tool not based on theft isn’t.

What about using a stolen table saw? Or one bought cheap, because it was stolen?
Would you feel any differently if the inputs were ethically sourced?
Absolutely. Those of us on the “AI and LLMs only work because they are based on the theft of ideas on a grand scale ” would fee very differently about the technology as a whole if it weren’t rooted in harvesting smdata without permission.

For example, AI tools that scan medical charts within a hospital’s OWN data is probably (that’s what the vendors say at least) not stealing the chart data for training the vendors product more. It’s used by and within the hospital and the private healthcare information within its charts. That’s not as bad as wholesale theft, even if the LLM got where it is via unscrupulous methods.

Back to Lit and writing related topics, there would still be other ethical issues such as “did you really write that?” But knowing the tool itself wasn’t a massive stew of plagiarism would make a difference to those of us who don’t deal in stolen goods.
 
Now the discussion curves again towards AI as writer. This is not the topic here. Editing is not a hugely creative task. Generative Al is not even needed for it. You could use an expert system (which is an older, simpler attempt at AI from the 1980ies) Take an input text, find texts in your database that are similar. Search how these texts were rated. Display this rating. No magic there. Just a machine sifting through large volumes of text faster than a human can.
 
The problem with self-editing, is that in my head I have the consistent image of the scene, but when I write it down, it is something else. And when I read it, I do not recognize what is missing or wrong.

Then the AI is very patient and always available. I can rewrite a passage and ask, what effects these changes have. I can do that 10 times and learn from that. Something I would not be able to understand just by myself.
I encourage a process that usually requires at least three of four steps: 1. Write uncritically from a personal POV. Make your people talk like people you know. Sometimes have them do what they want to do and you didn't first see them doing. 2. Put the damn thing away. don't be in a hurry to get your props. 3. READ it at 'gut level' with a 'blue pencil', marking or crossing out, stuff that makes you feel uncomfortable. Maybe hinting to yourself a better idea. (Trust yourself that if you feel it is wrong you have a hint of what is right.) 4. Put the damn thing away. 5. Come back as a 'friendly editor'. Never beat yourself up. Move through it 'bettering' the places that felt wrong. Go down a rabbit hole your character wants to follow. You can always chuck it. 6. Put the damn thing away. This isn't a college paper. 7.
 
Back
Top