Did Machine Translation peak?
I mean, not that it has reached its full potential, just that it won't be developed anymore because of AI translation tools... I'm trying to use a local AI to translate subtitles and it's just using 97% of RAM and 50% of CPU and doing shit lol
6 replies
I don't understand.
Why would AI translation stop translation development?
What does that have to do with what your specific computer is doing?
The current wave of LLM development did put a stop to a lot of AI research. LLM is this big hammer that you can hit everything with, and kinda solve it. But it's rarely the most efficient and correct. But because it's so much easy to use, that's what people are using now.
LLM translation is waaaay more advanced than machine translation.
My fear is that they will stop machine translation because it has been so outclassed.
My special case is that running it locally is very resource consuming (unlike machine translation).
Oh! I didn't realize you were talking about two different things. LLMs are still machines. Now I get it.
Yah I would expect since LLMs do a much better job, the older methods are going to slowly disappear.
LLMs are literally the evolution of the machine translators you are talking about.
but it's like taking a bazooka to a pistol duel... if they would use LLMs to build NMT database it would be great.