I don’t want to sound like a downer, but I’m skeptical about Armada: never heard of it before, repo created last June, tons of features implemented in that short span and already “ready to go”… Seems like it has been heavily vibe-coded.
Yes, if you block ads, then you do not generate revenue for them.
For those looking for a showcase, take a look at this recent publication in an academic journal.
Typst is way simpler to learn, especially if you’re used to markdown. You can first approach it as a markup language like markdown, but with some scripting.
Package management in LaTeX is infuriatingly bad, you can’t even specify the version you want to use.
Some journals and conferences are starting to accept Typst.
Personally, the huge difference with LaTeX is that if I want to do something slightly complex that is not covered by a third-party plugin, I’m not afraid to get my hands dirty and script it myself, whereas in LaTeX, I dread writing macros that do more than being notations/shorthands.
It’s not the same use case. Asciidoc is closer to markdown/org/reST, i.e., simple markup languages, whereas Typst also emphasizes on presentation (layouting, element positioning, creation of complex figures, etc.). You can reproduce features from Asciidoc in Typst using scripting.
As for editors, aside from the official webapp, and the community LSP (tinymist), there aren’t that many available.
The term you’re looking for is GPGPU (General Purpose computing on GPU)
700+ preprints is a lot and takes time to be ingested, especially when the text is generated. For example, Communication in Mathematics had less than a hundred submissions in 2025 and accepted a fourth of them.
So, the small correction rate might be due to the bottleneck at the verification stage, which OpenAI should have done in the first place, rather than throwing everything at mathematicians and expecting them to do the ungratifying work whilst they reap the fame.