I go a lot faster this AI way.
Before I was writing a lot of definitions, now the AI contextual is often correct in my scenario. If I encounter the word again, I might choose another contextual, and if I encounter the word more times, I might improve it.
I’ve written tons of definitions so far, and I go a lot faster now. AI seems pretty cool and can definitely improve with time.
Before, many popular meanings were wrong, many were just from Google Translate, and many that were “popular” weren’t actually the better option among the various translations.
But I’ve never used the auto-lingq for the exact reason that many popular meanings were incorrect.
@mark Maybe, an option could be for who is using auto-lingq only. The possibility to choose the auto-lingq from AI, or from the most popular meaning like before. This could fix it for now.
@davideroccato Perhaps I wasn’t clear. I am not advocating for using Popular Meanings. On the contrary. I am advocating for doubling down on AI-created Contextual Definitions. I would just like them to work better on Android to deal with words with multiple definitions by also showing the AI-created Contextual Definition for yellow words in the pop-up without all the extra steps. You see the AI-created Contextual Definitions in the pop-up for New Words, Ignored Words, and Known Words. I just want the exact same interface but also for lingQs/yellow words. I want to see the AI-created Contextual Definition every time I click a word, no matter what colour the word is.
Also, increasing the speed of the Contextual Definitions would not go astray. They kinda take a while.
It’s an interesting suggestion. Although, not exactly the interface for Ignored words but I get what you mean. We’ll see if we can add a setting to show the contextual meaning on LingQs.
Yes, I agree with you and @nfera. Since these are contextual clues, I would like them to be displayed on yellow (and possibly even white) words, too - after all, these are not always going to be in the same context I first saved them, either!
my native language is Hebrew and it’s the same situation - No community definitions at all but AI word translations are perfect for Russian-Hebrew’ English-Hebrew and Persian-Hebew. Thanks to AI new function I can read a lot without opening external dictionaries or google for every word.
vice versa. without AI (if your mother language is not English) it was annoying. because there are languages without any community translations and AI stanslations are accurate. The reason why the AI translates an entire phrase and not a single word is that most words do not have a single meaning when they are alone. Their meaning is determined only in the context, therefore usually the translation that is correct in the context will not be accurate when the word appears in its single form. Therefore, although it is annoying, it would not be correct to present a translation for a single word.
@Cinderela You know this is no all or nothing debate. Some users asked for an option to disable it, not to remove it, and everyone who likes the AI feature suddenly goes bananas as if something will be taken away from them. The ability to disable the feature was added, and you still can use the ai feature as before as it is just optional.
The same applies to other suggestion made by those who dislike the feature in its current state or see it as more of a hinderance as a support. Noone is asking for a removal of the feature, left alone the LingQ team won’t do that anyway after all the time invested. All they are suggesting are improvements so the system does create better results for as many people as possible. Should development be stopped just because some think it is okay as it is?
We are all aware that depending on ones mother tongue the user created definitions and the dictionaries available aren’t optimal, and that the quality of the ai generated translations depends on the language pair choosen. This has been stated several times now and is clear. But you could also acknowledge the experiences made by others.
In regards to dictionaries I would suggest that you check the internet for some and kindly ask the LingQ team on whether it is possible for them to be added. If you check out the forum you will see that this happens every now and than.
@mark it would be nice if they weren’t cut too. There’s a new problem where some translations, especially the ones that have more than one word as a result, are cut in half.
For example, instead of delivering “train station”, it delivers “train stat”.
Or instead of delivering “apple pie”, it delivers “apple p”, and so on.
It would be nice if it could be fixed.
Same thing with the generated audio from an English text.
Yes; I use the web because, with a macro and keypad, I can get comprehensible input in one keypress: the next sentence + the translation + the audio. This is fundamental to the comprehensible input approach (that you know what you’re reading) and I’m amazed it isn’t one click on the app.
Oh ok, I didn’t realise that. For some reason, when I copy and paste text into ChatGPT 4o, I tend to get a better translation. I’m not sure what’s going on.