> Add another aspect here that LaTeX is a bit outdated in 2024 (I know that’s controversial! Sorry)
I met Leslie Lamport seven or eight years ago and asked him what a completely modern LaTeX might look like. He replied
“well, we won’t be using PDFs in twenty years” and so it would need to be something completely different. Something interactive, with depth. Remembering, of course, to focus on quality content first and quality presentation second.
In a world with LLMs, this question becomes ever more interesting - why write a literature review if one can be generated?
I think the important thing is capturing information in basic blocks (text, images, etc) and having the flexibility to reflow it later for any modern presentation mode, be it ingestion by LLM, listening to it, or just rendering it on desktop, mobile.
Translation and a11y is another important consideration here.
I met Leslie Lamport seven or eight years ago and asked him what a completely modern LaTeX might look like. He replied “well, we won’t be using PDFs in twenty years” and so it would need to be something completely different. Something interactive, with depth. Remembering, of course, to focus on quality content first and quality presentation second.
In a world with LLMs, this question becomes ever more interesting - why write a literature review if one can be generated?