When I got to college, I realized that I couldn't write a coherent paragraph or an intelligible essay. When writing programs such as Microsoft Word came out, it transformed the process of writing for me, taking out much of the "friction" of editing, which was the biggest inhibitor to expressing my thoughts. Access to improving the quality of my writing was challenging too, as Funk and Wagnall and MLA were pretty much the only sources to improve grammar and style.
Word was a revolutionary improvement for me and many others, but the friction that remained was getting feedback, expanding my thoughts, and going deeper into detail. Now with AI models, that friction has been reduced dramatically once again.
Writing is not something everyone is enamored to, but I predict we will see more and better writing in the next few years by those who use AI as a writing partner, critic, coach, and colleague. I don't see that it will have much of an improvement for creative forces of nature like Billy Collins or George Saunders, but for the rest of us, it's a game changer.
I love the alchemy of pulling ideas from my head and unveiling them as words that I can share with others. It's always a struggle, but a good struggle. That process is not replaced by an AI model; rather, it's enhanced, and my ability to be creative has leveled up. I can ask any model to give me suggestions to improve a passage, suggest ways to add more detail, and provide different perspectives on a topic.
Some of this reminded me of young writer I met recently who suggested a method for using AI to improve creative writing that he calls “adversarial creativity.” The basic idea is that a writer generates original work, feeds it to AI, and then asks the AI to generate something similar in tone, perspective, etc. If there is overlap between the two, the writer must rewrite their own work and omit the lame ideas that the AI generated. Each revision of the writer’s work can become more original by removing generic ideas. Much longer explanation here:
I think the podcast feature of Notebook will be much better once there is more power to dictate the tone and voice within the podcast. I find myself tremendously put off the first time I here a redundant 'like' from the AI host.
The funny thing is that none of these tools are particularly new or innovative. Writers and creators have been building these kinds of things for a while ... in ways that are way more useful.
We want writers and creators building AI tools, not software developers or mega-corporations.
Great post. Thanks. I‘m testing NotebookLM and Claude Projects. I also pay for Claude - there‘s a document limit there, while NotebookLM keeps accepting sources. I save all impt chats to Notes in NotebookLM.
When I got to college, I realized that I couldn't write a coherent paragraph or an intelligible essay. When writing programs such as Microsoft Word came out, it transformed the process of writing for me, taking out much of the "friction" of editing, which was the biggest inhibitor to expressing my thoughts. Access to improving the quality of my writing was challenging too, as Funk and Wagnall and MLA were pretty much the only sources to improve grammar and style.
Word was a revolutionary improvement for me and many others, but the friction that remained was getting feedback, expanding my thoughts, and going deeper into detail. Now with AI models, that friction has been reduced dramatically once again.
Writing is not something everyone is enamored to, but I predict we will see more and better writing in the next few years by those who use AI as a writing partner, critic, coach, and colleague. I don't see that it will have much of an improvement for creative forces of nature like Billy Collins or George Saunders, but for the rest of us, it's a game changer.
I love the alchemy of pulling ideas from my head and unveiling them as words that I can share with others. It's always a struggle, but a good struggle. That process is not replaced by an AI model; rather, it's enhanced, and my ability to be creative has leveled up. I can ask any model to give me suggestions to improve a passage, suggest ways to add more detail, and provide different perspectives on a topic.
Some of this reminded me of young writer I met recently who suggested a method for using AI to improve creative writing that he calls “adversarial creativity.” The basic idea is that a writer generates original work, feeds it to AI, and then asks the AI to generate something similar in tone, perspective, etc. If there is overlap between the two, the writer must rewrite their own work and omit the lame ideas that the AI generated. Each revision of the writer’s work can become more original by removing generic ideas. Much longer explanation here:
https://wgcr.substack.com/p/thesis
A great model-by-model analysis. Things get really interesting near the end. Students react in really interesting ways to AI generated content.
I think the podcast feature of Notebook will be much better once there is more power to dictate the tone and voice within the podcast. I find myself tremendously put off the first time I here a redundant 'like' from the AI host.
The funny thing is that none of these tools are particularly new or innovative. Writers and creators have been building these kinds of things for a while ... in ways that are way more useful.
We want writers and creators building AI tools, not software developers or mega-corporations.
Great post. Thanks. I‘m testing NotebookLM and Claude Projects. I also pay for Claude - there‘s a document limit there, while NotebookLM keeps accepting sources. I save all impt chats to Notes in NotebookLM.