For a long time, I was reluctant to use AI. I believed it would stifle creativity and take jobs away from artists and authors. However, I've recently started using it as a research tool, specifically Google's Notebook LM (Language Model), to aid my writing.
What exactly is Notebook LM, you might wonder? NLM is a language model where you provide the sources. Unlike Gemini or Chat GPT, it doesn't scour the web but uses the documents you upload.
But let's skip the technical jargon - what are the advantages of using NLM in writing? Here's an example: I created a dictionary for my ConLang (Constructed Language), similar to Klingon or Elvish. Remembering every word in the Veillen language is impossible for me, so I compiled a dictionary and uploaded it to Notebook LM.
If I ask Notebook LM to translate: Veillen protect humans from the graekull, the response is:
To translate "Veillen protect humans from the graekull" into Veillen, it's crucial to follow the Veillen grammar rules. The typical word order is Subject-Object-Verb-Modifier.
Here's a breakdown of the translation, based on the source text:
Veillen is Vellen.
Protect is tryg.
Humans is Imano.
From is proem.
Graekull has two possible translations: kat'ak (agile) or dakuma (brute).
Assembling it all while keeping the Subject-Object-Verb-Modifier order, possible translations are:
Vellen Imano proem kat'ak tryg.
Vellen Imano proem dakuma tryg.
This eliminates the guesswork from creating content for my third book. Another feature is the new Deep Dive Audio Overview, an AI-generated podcast with two non-human hosts. I've experimented with this feature and thoroughly enjoy listening to them. Interestingly, you can interact with the hosts and ask questions!
A quick disclaimer about the accuracy of NLM - it can only give what it is given - so if the work is incomplete or not specific it will make conjectures, so accuracy can be a bit off it it doesn't have full context.
Comentários