What if the problem with AI writing novels isn't context length, but how we're using memory?

Description and use

<p>I keep seeing the same argument whenever people talk about AI writing long-form fiction:</p> <blockquote> <p>AI can't really write a 100,000-word novel because eventually the context becomes too large, details get lost, and the model starts contradicting itself.</p> </blockquote> <p>And I think we might be looking at this as an <strong>LLM problem when it's actually an architecture problem.</strong></p> <p>Hear me out.</p> <p>We've already solved something surprisingly similar in software engineering.</p> <p>If I'm working on a 200,000-line codebase, I don't need to hold the entire codebase

Problem

<p>I keep seeing the same argument whenever people talk about AI writing long-form fiction:</p> <blockquote> <p>AI can't really write a 100,000-word novel because eventually the context becomes too large, details get lost, and the model starts contradicting itself.</p> </blockquote> <p>And I think we might be looking at this as an <strong>LLM problem when it's actually an architecture problem.</strong></p> <p>Hear me out.</p> <p>We've already solved something surprisingly similar in software engineering.</p> <p>If I'm working on a 200,000-line codebase, I don't need to hold the entire codebase

Outcome

<p>I keep seeing the same argument whenever people talk about AI writing long-form fiction:</p> <blockquote> <p>AI can't really write a 100,000-word novel because eventually the context becomes too large, details get lost, and the model starts contradicting itself.</p> </blockquote> <p>And I think we might be looking at this as an <strong>LLM problem when it's actually an architecture problem.</strong></p> <p>Hear me out.</p> <p>We've already solved something surprisingly similar in software engineering.</p> <p>If I'm working on a 200,000-line codebase, I don't need to hold the entire codebase

Evidence

What if the problem with AI writing novels isn't context length, but how we're using memory?

Source