I am curious—are you using AI for your work to be read? It can’t begin to express the text, if so. The text becomes wooden and void of the ability to make connection. But I love your writing.
Ok. So I was right then about your comment. And yes I agree. At one point I thought it would be fun to read all of my articles like heather cox Richardson does. But that adds a lot of hours! But maybe as I publish the books I could do my own readings of those. So much to do so little time.
Do you mean have I experimented with using an AI voice to speak my writings such as speechify? If that is what you are saying, Yes, I have played with that. I have even cloned my voice on speechify to try it out. But, like you say, it does not really capture the text very well. The timing and emphisis are off as well as the rhythm and cadence. I have played around with how to type the text in a way that more closely resembles it (as if writing a musical score) so that as AI reads it, it captures the right music of the voice by using punctuation and line breaks. It might be possible with enough effort to get pretty close but I have not spent enough time to figure that out. At the same time, I am a bit clumsy on the delivery myself when I am reading a text out loud but that would be a better choice I think. I sometimes listen to cloned texts, there are plenty of AI youtube videos. But you can tell they are AI and that takes some of the enjoyment out of it and sometimes they are hard to follow becuase there are not enough little silences where they should be for the delivery of the idea to the mind. Have you listened to some of my writings as audio?
I use ChatGPT and Google AI mode to help me out with research for my articles and, as a collage artist I see a lot of the text as collage material. But AI LLMs are just predicting what word should follow the previous word based on its algorithmic training. So I am often leading the AI conversation down a certain path to sharpen and clarify my points I am trying to make. In that way it helps to speed up my train of thought.
Some times I realize I myself am a little off track or a little half baked or am not being clear enough and sometimes other perspectives are presented by AI that I had not gotten to the point of considering.
Sometimes I am experimenting with how far I can push things, testing what it can do under different conditions and sometimes it amazes me the things it comes up with, sometimes I am disappointed at its shortcomings (or are they my shortcomings?).
I have always been pretty sloppy with language. Back in the day I thought the greatest advancement was spell checker. I was often saying to myself "Oh, that is how you spell that?" It greatly improved my spelling. I don't have an MFA in writing arts so consequently there are a lot of rules and techniques in writing that I am still unaware of. AI knows all of the rules and has all of the examples of all of the great writings of all time in every language. So that is like having an entire team of experts at one's disposal if you can figure out how to take advantage of it.
Here is an interesting experiment. At first I wrote my above response. Then I put your original comment into Chat GPT and asked it how to respond to what you said. CHAT assumed you meant my writing was AI generated and made a response based on that assumption it made.
Then I pointed out that you said: "are you using AI for your work to be read? It can’t begin to express the text, if so. The text becomes wooden and void of the ability to make connection."
Then I asked: I am thinking she means am I using AI to read my texts audibly as in speechify. Am I misreading that?
then CHAT said: Yes. On a second reading, I think your interpretation is quite possibly the correct one. Concluding after some analysis: So yes, I think my previous interpretation jumped too quickly to the current controversy over AI-assisted writing.
I then took my comment that I tapped out on my own (above) and asked Chatwick (ChatGPT) to look it over and correct anything I got wrong.
CHAT's main correction was this:
I'd be careful with this:
“AI knows all of the rules and has all of the examples of all of the great writings of all time in every language.”
That's too sweeping and isn't quite accurate. An LLM has learned statistical patterns from enormous amounts of text, but we shouldn't say it necessarily has "all the examples" of great writing. Your underlying point is good, though: you're suddenly able to converse with a tool that has encountered an extraordinary range of writing conventions, styles, structures, and traditions.
So then here was CHAT's corrected response to your comment:
"Do you mean, have I experimented with using an AI voice to speak my writings, such as Speechify? If that is what you are asking, yes, I have played with that. I have even cloned my own voice on Speechify to see what would happen.
But, like you say, it doesn't really capture the text very well. The timing and emphasis are off, as are the rhythm and cadence. I've experimented with typing and punctuating the text in ways that more closely resemble a musical score, using punctuation, spacing, and line breaks to try to get the AI to capture the right music of the voice. It might be possible, with enough effort, to get pretty close, but I haven't spent enough time with it yet to figure that out.
At the same time, I'm a bit clumsy in my own delivery when I'm reading a text aloud! Still, I think that would probably be the better choice.
I sometimes listen to AI-voiced texts, and there are plenty of them on YouTube now. You can usually tell they are AI, and that takes some of the enjoyment out of listening. Sometimes they are actually difficult to follow because there aren't enough of those little silences where they should be - those tiny pauses that give the idea time to arrive in the mind.
Have you listened to some of my writings as audio? I'd be interested to know if that is what prompted your question.
As for the writing itself, I do use ChatGPT and Google AI to help me with research and to think through ideas for articles. As a collage artist, I suppose I also naturally think of a lot of text as collage material. I am usually leading a conversation down a particular path, testing an idea, sharpening it, rearranging things, looking at it from different directions, and trying to clarify the point I am attempting to make. In that sense, AI can greatly speed up my train of thought.
Sometimes in the process I discover that I am a little off track, or that an idea is still half-baked, or that I haven't expressed something clearly enough. Sometimes another perspective comes up that I hadn't yet considered.
And sometimes I'm simply experimenting to see how far I can push the technology and what it can do under different conditions. Sometimes I am amazed by what comes back. Other times I am disappointed by its shortcomings - or perhaps they are my shortcomings in knowing how to use it.
I've always been pretty sloppy with language. Back in the day, I thought spell-check was one of the greatest technological advancements ever. I was constantly saying to myself, “Oh! So that's how you spell that?” Over time, it actually improved my spelling.
I don't have an MFA in writing, so there are plenty of rules, traditions, and techniques of writing that I have never formally studied. What fascinates me about AI is that it has been trained across an enormous range of writing and can recognize patterns, structures, conventions, and approaches that I may not know about. In that sense, it can sometimes feel like having a roomful of extraordinarily well-read assistants available for conversation.
But you still have to know what you are trying to say. And, maybe more importantly, you have to recognize when the words coming back no longer sound like you."
>>>>
You can see that CHAT kept most of my original language and made the corrections needed to be as accurate and clear as possible.
Thanks for your reply. I was referring to the voice used to “read” your poem. (I listened to it as well as read it.). Without breath to infuse life into spoken words, they become wooden and incapable of phrasing. The ancient Greeks knew this in how they wrote their language as well as how they spoke it. The two were intertwined.
Anyway, I am sure you are busy but reading your own work would be a good thing to work toward. There are coaches who could help. (I am not one of them so don’t assume I am trying to sell something!)
I am curious—are you using AI for your work to be read? It can’t begin to express the text, if so. The text becomes wooden and void of the ability to make connection. But I love your writing.
Ok. So I was right then about your comment. And yes I agree. At one point I thought it would be fun to read all of my articles like heather cox Richardson does. But that adds a lot of hours! But maybe as I publish the books I could do my own readings of those. So much to do so little time.
Hi Cate, Thanks for writing.
Do you mean have I experimented with using an AI voice to speak my writings such as speechify? If that is what you are saying, Yes, I have played with that. I have even cloned my voice on speechify to try it out. But, like you say, it does not really capture the text very well. The timing and emphisis are off as well as the rhythm and cadence. I have played around with how to type the text in a way that more closely resembles it (as if writing a musical score) so that as AI reads it, it captures the right music of the voice by using punctuation and line breaks. It might be possible with enough effort to get pretty close but I have not spent enough time to figure that out. At the same time, I am a bit clumsy on the delivery myself when I am reading a text out loud but that would be a better choice I think. I sometimes listen to cloned texts, there are plenty of AI youtube videos. But you can tell they are AI and that takes some of the enjoyment out of it and sometimes they are hard to follow becuase there are not enough little silences where they should be for the delivery of the idea to the mind. Have you listened to some of my writings as audio?
I use ChatGPT and Google AI mode to help me out with research for my articles and, as a collage artist I see a lot of the text as collage material. But AI LLMs are just predicting what word should follow the previous word based on its algorithmic training. So I am often leading the AI conversation down a certain path to sharpen and clarify my points I am trying to make. In that way it helps to speed up my train of thought.
Some times I realize I myself am a little off track or a little half baked or am not being clear enough and sometimes other perspectives are presented by AI that I had not gotten to the point of considering.
Sometimes I am experimenting with how far I can push things, testing what it can do under different conditions and sometimes it amazes me the things it comes up with, sometimes I am disappointed at its shortcomings (or are they my shortcomings?).
I have always been pretty sloppy with language. Back in the day I thought the greatest advancement was spell checker. I was often saying to myself "Oh, that is how you spell that?" It greatly improved my spelling. I don't have an MFA in writing arts so consequently there are a lot of rules and techniques in writing that I am still unaware of. AI knows all of the rules and has all of the examples of all of the great writings of all time in every language. So that is like having an entire team of experts at one's disposal if you can figure out how to take advantage of it.
Here is an interesting experiment. At first I wrote my above response. Then I put your original comment into Chat GPT and asked it how to respond to what you said. CHAT assumed you meant my writing was AI generated and made a response based on that assumption it made.
Then I pointed out that you said: "are you using AI for your work to be read? It can’t begin to express the text, if so. The text becomes wooden and void of the ability to make connection."
Then I asked: I am thinking she means am I using AI to read my texts audibly as in speechify. Am I misreading that?
then CHAT said: Yes. On a second reading, I think your interpretation is quite possibly the correct one. Concluding after some analysis: So yes, I think my previous interpretation jumped too quickly to the current controversy over AI-assisted writing.
I then took my comment that I tapped out on my own (above) and asked Chatwick (ChatGPT) to look it over and correct anything I got wrong.
CHAT's main correction was this:
I'd be careful with this:
“AI knows all of the rules and has all of the examples of all of the great writings of all time in every language.”
That's too sweeping and isn't quite accurate. An LLM has learned statistical patterns from enormous amounts of text, but we shouldn't say it necessarily has "all the examples" of great writing. Your underlying point is good, though: you're suddenly able to converse with a tool that has encountered an extraordinary range of writing conventions, styles, structures, and traditions.
So then here was CHAT's corrected response to your comment:
"Do you mean, have I experimented with using an AI voice to speak my writings, such as Speechify? If that is what you are asking, yes, I have played with that. I have even cloned my own voice on Speechify to see what would happen.
But, like you say, it doesn't really capture the text very well. The timing and emphasis are off, as are the rhythm and cadence. I've experimented with typing and punctuating the text in ways that more closely resemble a musical score, using punctuation, spacing, and line breaks to try to get the AI to capture the right music of the voice. It might be possible, with enough effort, to get pretty close, but I haven't spent enough time with it yet to figure that out.
At the same time, I'm a bit clumsy in my own delivery when I'm reading a text aloud! Still, I think that would probably be the better choice.
I sometimes listen to AI-voiced texts, and there are plenty of them on YouTube now. You can usually tell they are AI, and that takes some of the enjoyment out of listening. Sometimes they are actually difficult to follow because there aren't enough of those little silences where they should be - those tiny pauses that give the idea time to arrive in the mind.
Have you listened to some of my writings as audio? I'd be interested to know if that is what prompted your question.
As for the writing itself, I do use ChatGPT and Google AI to help me with research and to think through ideas for articles. As a collage artist, I suppose I also naturally think of a lot of text as collage material. I am usually leading a conversation down a particular path, testing an idea, sharpening it, rearranging things, looking at it from different directions, and trying to clarify the point I am attempting to make. In that sense, AI can greatly speed up my train of thought.
Sometimes in the process I discover that I am a little off track, or that an idea is still half-baked, or that I haven't expressed something clearly enough. Sometimes another perspective comes up that I hadn't yet considered.
And sometimes I'm simply experimenting to see how far I can push the technology and what it can do under different conditions. Sometimes I am amazed by what comes back. Other times I am disappointed by its shortcomings - or perhaps they are my shortcomings in knowing how to use it.
I've always been pretty sloppy with language. Back in the day, I thought spell-check was one of the greatest technological advancements ever. I was constantly saying to myself, “Oh! So that's how you spell that?” Over time, it actually improved my spelling.
I don't have an MFA in writing, so there are plenty of rules, traditions, and techniques of writing that I have never formally studied. What fascinates me about AI is that it has been trained across an enormous range of writing and can recognize patterns, structures, conventions, and approaches that I may not know about. In that sense, it can sometimes feel like having a roomful of extraordinarily well-read assistants available for conversation.
But you still have to know what you are trying to say. And, maybe more importantly, you have to recognize when the words coming back no longer sound like you."
>>>>
You can see that CHAT kept most of my original language and made the corrections needed to be as accurate and clear as possible.
Hi Cecil,
Thanks for your reply. I was referring to the voice used to “read” your poem. (I listened to it as well as read it.). Without breath to infuse life into spoken words, they become wooden and incapable of phrasing. The ancient Greeks knew this in how they wrote their language as well as how they spoke it. The two were intertwined.
Anyway, I am sure you are busy but reading your own work would be a good thing to work toward. There are coaches who could help. (I am not one of them so don’t assume I am trying to sell something!)