lil_csom - 18d
I'm in my 4th year living in Denmark as an expat, and I finally decided it’s time to properly learn Danish. I do have a Danish girlfriend, after all. One way I’ve been practicing is by trying to text only in Danish, but I often find myself stuck. I start my message in Danish, then hit a wall because I don’t know a word or how to fit something naturally into the sentence.
Especially in those cases, I used to give up and translate the entire message from English, which kind of defeats the purpose and interrupts the learning process.
So I started prompting GPT. I’d write my message with wildcards or notes for the parts I didn’t know, and it would return a corrected version. That worked well, but reusing the prompt each time became tedious.
So I built a wrapper around it.
Now I can type in the target language, mark unclear parts with curly braces {like this}, and get an instant corrected version with explanations. I also added a history feature so I can review what I got wrong, and I plan to build more on that soon (eg. summary of areas or words to review).
This app is for language learners who want to practice writing without feeling insecure about mistakes or breaking their flow by switching to a translator.
I hope you find it useful!
This is an interesting idea.
But, besides not being able to use your site due to the errors mentioned by other posters, both the examples on your website give me pause:
The example input "Kan you hjelmpe mig {yesterday}?" reads to me as "Can you help me yesterday?", but that's just a nonsense sentence and an odd choice for an example. The word for "help" is also misspelled, but presumably that would get corrected.
And the suggestion of "Jeg vil gerne handle i morgen" for "Jeg vil gerne go shopping i morgen", instead reads to me as "I would like to act tomorrow". A more idiomatic translation would be "Jeg vil gerne købe ind i morgen".
> And the suggestion of "Jeg vil gerne handle i morgen" for "Jeg vil gerne go shopping i morgen", instead reads to me as "I would like to act tomorrow". A more idiomatic translation would be "Jeg vil gerne købe ind i morgen".
It's a little better, but I would never expect anyone to translate "shopping" to "købe ind". "købe ind" is about getting groceries for the week, shopping is about walking the strip and dreaming of buying random clothing. As a native speaker I'd be less surprised if the you just used the borrowed word "shopping" directly. Basically "Jeg vil gerne shoppe i morgen".
In British english at least, 'going shopping' is a normal way to say 'getting groceries for the week'.
Yeah groceries is "shopping"
"walking the strip and dreaming of buying random clothing" is "window shopping"
In Swedish, shopping has been loaned to mean window shopping. Buying groceries is commonly called att handla.
I think "go/do shopping" is the same in French, Spanish, Italian and German, as opposed to faire les courses / ir de compras / fare la spesa / einkaufen.
In US english (at least as I speak/understand it) "shopping" is any act of browsing/looking at/selecting something for purchase. It can be groceries, clothes, a car, anything really and it can be online or in person at a store.
"I'm going shopping" with no other specifics would normally mean "for groceries" or other general household supplies, though.
If a European language gained a new everyday word in the last fifty years, there’s a solid chance that it’s a loan word from English. A little odd to learn a “foreign” language filled with that stuff.
> European
It's not restricted to European languages. 贝果 is bagel, just sounded out phonetically, and 三明治 is sandwich.
Idk if there's anything super odd about it.
Of course, English is the worst offender of loan words. As someone else said somewhere, "[English doesn't] just borrow words; on occasion, English has pursued other languages down alleyways to beat them unconscious and rifle their pockets for new vocabulary."
Gung ho, monsoon, filibuster, herbivore, vacation, etc etc etc. Thousands upon thousands of loanwords.
Fun fact, 99% of words ending in -tion are the exact same in French. Every English speaker has a head start of hundreds of word vocabulary in French.
Those words also exist in Spanish, there the ending is "cion", and in Portuguese with ending "cão"
And in Polish: -cja. Lots of languages have had a deep relationship with Latin, not just Romance languages.
On the other hand, English is a Germanic language and got influenced by French. A nice recent article talked about why English doesn't use accents (because... of French) [1].
Just to say it's not uncommon at all to have languages influence each other.
[1]: https://www.deadlanguagesociety.com/p/why-english-doesnt-use...
With Scandinavian languages it went full circle, there are lots of everyday English words stemming from old Norse :)
yes, a lot indeed. Even some very rare adoptions that almost never happen in languages (like the pronoun they). My most favorite has to be window though from the Old Norse vindauga (vind = wind, auga = eye).
Same in French. When my colleagues tell me about "le repository Git", I love to answer about "le repositoire git" - sounds mightily quaint but that French word is actually the one through which the Latin repositorium percolated to English.
As a Swede, I can recognize the norse root of most English word for things that existed 1000 years ago.
You'll love Flanders.
Hey, thank you for your input! I definitely have to improve the example texts and make sure that the way the tool should be used is better understood!
Interesting use case. By coincidence I am also learning Danish quite intensively. What I do (and that tends to work for me) is the following:
- I try to actively write something and when I hit a wall, I try in other languages: the important in this part is to get the text flowing from me. Eg, "du could har hjælpe mig, når jeg called du" (trying to write in Danish, "you could have helped me when I called you")
- I then paste it on ChatGPT and ask for a C2-level correction with explanations on vocabulary and grammar, and translations into a few other languages. Eg, my prompt would be """ "du could har hjælpe mig, når jeg called du"
correct to c2-level and explain grammar and vocabulary. give related examples. then translate to German, English, French and Italian. """
This is incredibly helpful to do everyday, especially if you are also learning passively by reading/studying by yourself. By the way, for those curious, here's how ChatGPT would correct the sentence: "Du kunne have hjulpet mig, da jeg ringede til dig." Interestingly, it assumed I meant "called" as in, "telephoned" (not my original intention). Translating into other languages helps you spot and get a sensibility for such cases.
As an aside, for me the most challenging part of learning Danish is the pronunciation. It's beautiful, but it doesn't map out too easily to written words :)
> As an aside, for me the most challenging part of learning Danish is the pronunciation. It's beautiful...
As a Dane, all I have to say is; lol
The peoples of Sweden and Norway stand behind you on the pronunciation issue!
This is really cool!
I've done a lot of research into LLM translation for my product[0], and I'm currently working on a deep translation service that provides reliably human-level translations.
I don't know what model you're using, but GPT-4.1 is probably the best for your use case - it's in the top few % for nearly every language, and has a low standard deviation, while also being relatively low latency and low cost.
Thank you so much for the suggestion! I'll later also check out your app!
It doesn't seem to work with "any language", as claimed. It cannot process Chinese writing (nor can it process pinyin). "Error processing your text", is the message. Though it's certainly possible the volume of traffic has overwhelmed things and it would work under normal circumstances?
I tried with Korean and had the same thought but English seems to trigger the error as well as I'd say the poor app received HN hug of death.
I tried Spanish. I wanted to test a few words that are different in Spain versus Mexico. I didn't get any results for anything I tired.
> works with any language
Does it work in Tojolab'al (Mayan language spoken by 70,000 in Chiapas)?
I get mildly irked by apps saying they work in "all/any" language. Obviously, it works in whatever languages ChatGPT "knows". Although for marketing, I understand the former is easier to say.
fwiw - ChapGPT acts like it knows these kinds of low-resource languages, but it's difficult to tell how much of it is just hallucinating. (I am trying to do work in this area).
First, I applaud you building something and encourage it.
Next, I want to say I don’t understand these llm “apps”. I tried this in ChatGPT with the minimal prompt (that you can keyboard shortcut) “fill in curly” and it gave me exactly the correct full sentence plus even better alternatives.
Why would I use a separate app?
In general why would I use any of these separate apps vs my included ChatGPT that I’ll have a button away with included unlimited subscription
I imagine once this app becomes more mature, it'll have more options and functionality around its specific use-case, and tailored UX.
ChatGPT just has a text input and output. They can't build ChatGPT to have the perfect UX for every single eventuality. TLDR - UX, convenience, and ease of use.
ChatGPT can literally build its UX on the fly
> Why would I use a separate app?
It's not about whether you'd use it or not. It's just so that people have something to build, managers have tasks to distribute, and the machine can keep on turning.
Not to be nag but the example translation at the bottom of the page is not correct. Shopping is a word used in danish.[0] The translation with the word handle does not correctly convey the suggestion of going out to multiple stores looking at clothes etc. to buy. It more comes across as wanting to go buy groceries, or doing some kind of general commerce.
A translation to standard spoken danish would be jeg vil gerne shoppe i morgen.
I entered the text that was present in the placeholder, but I got an error message: "Unable to process correction Error processing your text. Please try again."
Quick feedback, the idea sounds interesting to me and I wanted to learn more. Feature request: more examples, or cached examples that would still work if the server crashes.
Take an intensive course. That's what I did to learn Dutch with my Dutch then girlfriend (now wife and mother of our kids). Intense, focussed and I was fluent at a higher level than many natives within a year (which says more about 50% of the country than it does me, but I digress).
What was the time/format of this course?
[dead]
Tried it with some Dutch, only getting "Unable to process correction" errors. Good idea but seems it need a bit more error handling.
Hey yeah, I have ran out of API credits due to popular demand! Sorry that you couldn't try it this time :(
That the modern way of getting out of office messages on Welsh signs. See: https://imgur.com/a/Vr3WGKp
Not to hate or anything but, the site didn't work at all. I thought "hmm, maybe it's not made to handle non-ascii characters, but that's a very rare problem to have", so I switched to another language that used only ascii characters but it still didn't work. I guess it simply fails.
It seems that my GPT credits have been deplated, thus the error :) Sorry about that! I decided that if I do ran out, I will continue the development - showing that is has popularity!
In the meantime, I am at work, so I don't think I can add the rate limiting now... but thank you for everyone for the input!
it would presumably require quite a large rework of what you have, but you could imagine a version of this where the client was talking to the APIs, so that people could input their own API keys without sending them anywhere.
the core idea here is really cool, and the UX is (or seems like it would be) impressive - no need to select a language, for instance.
Good luck getting it back on its feet!
Sounds cool!
(Why doesn't ChatGPT make something like this, essentially plugins, for their web interface? You make a script with macros and put that on a "Marketplace" and people can run it... would remove the need for setting up a whole nother website just to wrap around prompts)
In the example text hjelmpe is not a word in Danish.
I assume the stuff about the AI giving corrections might apply to more than just what's in curly braces, i.e. that this might be intentional, but I'm unable to test due to a "Unable to process correction" error on the demo.
Haha, nope, that's just a genuine typo! :D Good catch!
A long time ago, before SEO spam, I would Google parts of what I had written phrase by phrase (with proper nouns removed or generalized) and if the number of results was under a certain threshold, I'd assume I was wrong.
I found https://ludwig.guru to be a good replacement for that. But, yes, Google sucked a lot less back then & you even got relevant & longer search suggestions. Someone had to 'fix' that. Everything for the bottom line.
I made something along the same lines, but just targeting WhatsApp Web specifically. I came up with the idea when I ended up in a WhatsApp group chat that was very MultiLingual. So this allowed me to participate in the conversation. It's also been handy to talk to contractors and service providers. https://ai-polyglot.com
Me and my wife made something very similar (https://nativi.sh/). I find it's a really good use case for LLMs to actually help you learn rather than doing everything for you.
Also, not sure if you're getting hugged to death but I'm getting this in the interface but not seeing any network failures.
``` Unable to process correction
Error processing your text. Please try again. ```
I did a similar thing with a personal spaced repetition system I am building. Instead of braces I used ?question marks? and I added daily writing quota to force myself to write instead of going straight to card reviews.
https://github.com/RickCarlino/KoalaCards
(You must add cards before accessing writing)
Interesting idea. Also got errors.
That withstanding, surely having the user select what their target language is, would help with processing and accuracy? It's a one time lift for the user, and would ensure I'm definitely getting back my target language, not a dialect or totally different (norwegian/swedish come to mind immediately)
Really good idea.
I also really like the name.
Maybe add a language selector such that the AI has a better context of what language the user is interested in.
Thank you! Yes, I think that would be the best really, so the AI has an easier time.