One of the under-appreciated°hard word benefits°hard word of using LLMs°hard word for coding°hard word is that they are experts on most topics°hard word.
Writing Trakaido°hard word would have been impossible if the LLM°hard word did not already have familiarity°hard word with Lithuanian. 💡 It was also familiar with dozens of other languages.
Languages are a strong point for large language models, but there are many other software projects that can benefit°hard word from having domain°hard word experts available°hard word to all developers°hard word at all times at a minimal°hard word cost°hard word.
----
This applies°hard word to non-coding°hard word projects as well. Trying to use LLMs°hard word to help file°hard word my taxes was ... still a bit optimistic°hard word in early 2026 🌎 there were several mistakes it made. But it generally understood what was going on. Without context°hard word.
💡 I have discussed before the question of context°hard word v. built-in°hard word training data. Should the machine be able to play chess without instructions°hard word (or external°hard word tools)? What about backgammon°hard word? Or a game I just made up? Is there any reason to believe/expect the universal°hard word game-playing°hard word machine will be a Large Language Model?