@Talia This is why I don't use "AI" in my translation...

@Talia This is why I don't use "AI" in my translation work. Yeah I'm supposed to edit its output or whatever, but the initial (crappy) output still has an anchoring effect and psychological/physical inertia works against making too many changes. Yeah there's value in seeing a variety of suggestions and examples, but I can get those through dictionaries and database searches WITHOUT destroying the planet.

Like 31 July at 2:30 | Wall-to-wall | Open on writeout.ink

3 comments

Talia Hussain

@ljwrites well said. In creative work, I often find inspiration and ideas come from apparently unrelated areas. Look around with a prepared and open mind and ideas present themselves

31 July at 11:55 | Open on mstdn.social

Richard Varga :esk:

@ljwrites @Talia Precisely. I don't use AI nor MT in my translation work because the initial output sets up a bias that I could overcome only if I had the ability to literally unsee things. I want the result to be the result of *MY* work.

31 July at 18:27 | Open on localization.cafe

L.J., un-author-ized writer

@larsimore This! LLM/MT outputs are mediocre at best and wildly incorrect at worst, nothing I want to be putting my name on. Back when I was getting post-editing jobs I edited so much the screen was a wall of color from tracked changes, and even then errors and inconsistencies still slipped through. Thankfully these jobs have petered out lately and I'm getting real translation work again, maybe companies are finally realizing the quality hit isn't worth it.

31 July at 23:29 | Open on writeout.ink