Thanks for this article! Hard to find content that helps you do things better!!
One interesting thing I found with Grok is that I have a rough outline/timeline for a story and it rephrased it slightly incorrectly like it was a human where it got 90% correct in the correct order but the other 10% was either incorrect by jumbling events together or it got things out of sequence.
Welcome aboard. Great initial post! We've got text specialists, music specialists, and roll-your-own specialists, now we just need image and video specialists.
Gemini can handle up to ~1M tokens, which is about 4-5M characters. By way of comparison, "A Game of Thrones" apparently has 9,724,851 characters including spaces. Of course, the context tokens are also used for any prompts you give it etc, so best to stick to shorter works all the same.
Thanks for this article! Hard to find content that helps you do things better!!
One interesting thing I found with Grok is that I have a rough outline/timeline for a story and it rephrased it slightly incorrectly like it was a human where it got 90% correct in the correct order but the other 10% was either incorrect by jumbling events together or it got things out of sequence.
Anytime.
Welcome aboard. Great initial post! We've got text specialists, music specialists, and roll-your-own specialists, now we just need image and video specialists.
Interesting post. Thanks.
Sure thing!
Gemini may do better with its larger context window. Larger chunks., maybe not the entire novel though.
Deepseek has problems with long outputs.
Gemini can handle up to ~1M tokens, which is about 4-5M characters. By way of comparison, "A Game of Thrones" apparently has 9,724,851 characters including spaces. Of course, the context tokens are also used for any prompts you give it etc, so best to stick to shorter works all the same.