Well This Rabbit Hole Goes Deeper... AI Feedback As A Conversation with Google NotebookLM
Over the last few weeks I became aware of this new free Google tool, NotebookLM.
It takes long and complex documents and turns them into a podcast with two people discussing it.
I wondered, what would it do with the feedback give on Sasha’s story, Darkness, My Friend (First draft)?
So I sent it the story, the human feedback left by the community and the feedback from ChatGPT (you can read all about that HERE). And this is what it created.
What do you think? Please leave comments below…
For clarity I have not done any editing, or tweaking of uploads. Literally just the story and feedback. I also did no regeneration or second attempts, this is the first thing it created.
Note: If the audio player is not appearing, try Incognito mode as it may be being blocked by ad blockers.
I alos asked it for some text feedback and it offered this…
The story also depicts the protagonist’s isolation and vulnerability in a way that evokes empathy from the reader, making the horror elements more impactful. The character of Helen is also effective as a creepy and unsettling figure, adding a supernatural element to the story. Additionally, the story builds tension effectively leading up to the climax, particularly with the image of the razor and the blood-red sky.
The dialogue with Helen also lacks depth and could be more sinister and manipulative to heighten the tension. The scene where the protagonist is bullied at the party feels rushed and lacks emotional impact, despite being a pivotal moment in the story. While ambiguity can be effective in horror, the story could provide more clarity about Helen’s true nature to make the ending more meaningful. Lastly, the ending feels rushed, and the protagonist’s decision to harm herself happens too quickly without sufficient exploration of her internal struggle.
Filmmaker, author, screenwriter and founder of Twisted50. Currently directing splinter unit on Mission: Impossible 8.
Want to get a little bit more Twisted? Here's some related posts...
Why YOU should enter Twisted50, Volume 3 with a Horror Short Story by Lucy V Hay 
Where in the world… Meet The Global Writers of Twisted50 vol 3 
Twisted 50: Why The Little Voice In My Head STOPPED Me For Pursuing The Success I Wanted 
Dare YOU Read Any Of Our Twisted Entries? What Feedback Can You Offer Writers? 
How will you win Twisted 3? by Christopher Stanley
14 thoughts on “Welll This Rabbit Hole Goes Deeper… AI Feedback As A Conversation”
That is really fascinating and I think is a really interesting tool. It was easy to listen to and really felt like a true conversation. I’d be curious to listen to more
Wow. It sounds so real. The feedback is good.
I am kind of blown away by this. The actual feedback analysis is not perfect, but TBH, it’s more often than not from a human too. The speed, the access, the way it just lands it really bending my mind.
Both fascinating and scary at the same time! That said, I’m not sure how useful the feedback would be as it’s based on what’s already been given, but it’s an easy listen and engaging.
Where this comes into its own is in large texts. I asked it to use a couple of pages, but you can get it to do hundreds, and distill it down.
Well it was certainly comprehensive, I didn’t expect a critique of such length and thorough description and doing it in the form of a discussion between two people is interesting as it gives it an air of informality, there’s no bombs going off, no cringeworthy OMG moments, it’s very friendly and intimate and sort of breaks it all gently to you. It would be interesting to know what the writer thinks of course and I’m going to have a go with the three Twisted pieces I wrote. Also wonder, though it may seem to defeat the object, whether once we’ve heard it if a transcript is available?
Wow. This sounded so… natural. It did seem also to add meaningful content to what had gone before (for example, the suggestion here to give a glimpse into Helen’s mind and plans). The additional text feedback was a clear summary of the major points, but these were largely the same points made by ChatGPT and in the human feedbacks; agree, therefore, a transcript of the AI dialogue would be a good addition, as it would be easier for the writer to subsequently mull over.
This is really cool! It sounds like a natural, free-flowing conversation, and I’m really impressed with the technology. As someone said before, I’m not sure if all the feedback is 100% on point, but opinions would differ between human readers, too. That said, while I’m enjoying the AI advancements, to a degree, I’m not sure who this would be for. For a writer who doesn’t want to read the (sometimes) too blunt feedback from an AI bot? I get that, as ChatGPT (and the like) can state its opinion like its fact, which gets irritating when you’re looking for what’s relevant.
Or is it for readers who’d like to hear an analysis for each story? Again, this is a cool concept that will scratch that itch, but I’m not sure if it may be too niche a need to be “worth it” beyond an experiment. Anyway, very impressed overall.
It’s great to hear a podcast of thoughts discussing the editor’s notes.
Hearing the nuances of alternative directions is refreshing and opens up the creation of a more intense mystery.
Still, knowing that there was a 2000-word max, I would be curious how many of the suggestions could be manipulated into the story and maintain the word count.
I have a lot to learn.
Thanks for sharing.
Very interesting, but for me too far! Using AI for the feedback and hearing the story read back has been beneficial I would say, but listening to two AI generated characters have an in depth “conversation” is blurring the lines too much. Others may get something from it as other commenters have, but I’m happy with the other existing forms.
OMG! I think the podcast was really good, goes to show how well A.I. can mimic a human because if it wasn’t stated, I would’ve totally assumed it was two real people talking. I’m split over this as I’m well impressed by what A.I. can do but its taking away from human creativity… that could put us all out to pasture unless some restraint is placed upon it. Once A.I. is placed in a fully functional body, where does that leave us?
HERE COMES SKYNET!
I’ve just run my 3 Twisted stories through Notebook…ahem…I have to say its immensely flattering to hear ‘their’ ‘conversation’, they seem to hit the nail on the head and actually apparently read, see things I had never anticipated but maybe thats the value of distancing oneself. I don’t know, it’s weird, I’m not sure how it’s doing it, ot’s like lt’s reading the story but it’s looking deeper into the meaning, the subtext, the themes of the stories, the emotional effect its likely to have or does have on a reader. It’s like going on a date with Charlize Theron and she seems to see inside your soul….and who’s gonna argue with her?
Fascinating! But so creepy at how realistic it all sounded. It gives some of the podcasts i’ve listened to a run for their money. Come to think of it, howe do I know those podcasts really were humans?
I can easily envisage a world where humans have long disappeared yet AI remains, analysing stories and works of art or music and then analysing their own analysis until even the original work of art is no longer the subject but merely a copy of a copy of a rendering of something once thought up by a human.
Yeah it is CRAZY and just going to improve year on year. Or is that month on month. Or week on week. Or maybe by tomorrow. It’s moving FAST.
CJ