fabiensnauwaert

Posted about Gliglish.com a few months back which does the same thing.

Have you thought about adding speech synthesis to it? For languages like such as English and French, with very opaque spelling, this sounds critical.

Hey, thanks! How can I contacter toi ?

Yes, one step at a time :) Definitely high on the list, pronunciation is very dear to my heart.

Hey, Could you please send me an email on how you did it at [email protected] ? I take this seriously.

It's the default model from the OpenAI API endpoint.

There's a translation button. Should I make it more prominent?

I'll also add voice and translations to the suggestions provided on every screen.

It's also a matter of learning style... I don't speak a word of Portuguese or Bosnian, but diving into those was a lot of fun! Language learning can be seen like a detective game and it's super exciting then.

In any case, will definitely try to accommodate more learning styles and personality types.

Cool project! I always meant to do something like this but never got around to playing with the tech.

r/languagelearning/ does not currently accept anything relating to AI. Here's the messages I got:

Automatically:

> Thank you for posting on r/languagelearning. Due to their frequency, posts relating to AI and chat bots have been banned from the sub.

And from the mods:

> Hello--it does look good. Unfortunately, we currently are severely curbing chatbot posts.

Hey! Thanks for the balanced feedback, all totally valid.

(1) I'm self-directed in my learning and always knew what I wanted my teachers to do, but everyone's different. Definitely going to add more choice of content (and not just roleplay.) I have a very structured and successful program to learn English and will try to adapt what works well in it to Gliglish.

(2) You're right! The speech recognition and LLM layers are separate. I actually do have data about the quality of pronunciation, I'm just not using it yet. It's a matter of adding a layer to handle it. Basically, all the building blocks are here and I need to get them to play nicely together.

(3) It's the same issue as with ChatGPT on which it's built. Part of the progress will come from OpenAI or similar endeavors themselves, part from further checks on my end as in (2).

I just don't share your pessimistic view about the future of such tech though (well, if I did I probably wouldn't have started the whole thing! ) Once you put ChatGPT into a corner, there's usually no hope of it getting back on track. BUT with clear context and prompts on things it does well it's just incredibly useful! Again, is the glass half empty or half full?

As for this use case, there's tremendous value in being able to practice. Just talking to oneself in the mirror or journaling in a foreign language is already super productive and this is it… on steroids.

I've seen people terrified at the idea of speaking (hell, I experienced it myself decades ago.) It's not always even a matter of level. People with good grammar and a lot of vocabulary will have zero ease speaking for lack of practice.

On the other hand, at some point, my super limited Hungarian (of maybe 300 to 600 words?) was getting me praise just because I could use the little I knew without thinking.

”Practice, practice, practice!…”

Yes, definitely want to add something in those lines, where the AI detects your level and adapts in more than one way.

Hey, thanks for the input and happy to have a healthy debate!

I'm a bit confused though:

> all the input I got was some lines of text

Didn't you hear the audio? Or what else were you expecting? I've experimented with the length of answers a bit and people are put off by long answers.

·

I think people overcomplicate language acquisition. Or at least here's my two cents from learning four foreign languages to fluency and now seeing my kids grow up trilingual. There's a natural language to learning a language:

1. Listen 2. Speak 3. Read 4. Write

Foreign language education has been switching the first two with the last two and keeps wondering why the majority of learners have such thick accents. (Meanwhile, kids who grew up with YouTube sounds great.)

You need input to observe and 'pick up' the language. But you need experience to grow confident and fluent. No speaking practice means no confidence speaking. e.g. if you never spoke on the phone after years of learning you're likely to get cold sweats on your first phone call for work.

Time and motivation play a big role. You just cannot squeeze the equivalent of years of learning as a kid into a couple of months (usually before giving up) as an adult. Making language rewarding (which means seeing progress and having fun) is critical.

·

Now my observation is that people just don't talk much in most classes and that people should be able to learn foreign languages in the most natural way possible: conversations (listening and speaking.) It's somewhat ironic that ARTIFICIAL intelligence is now enabling this.

I'll put my money where my mouth is: I'm still working on the product and when it's ready will use it to learn a language FROM SCRATCH :)

I should have disabled 'Multilingual speech recognition' by default. Lesson learned.

Thank you for the encouragements!

> you're probably in need of content that users will feel compelled to share on social media and other places

That's a very good point.

> It was slow to send the recording then provide a response, the speed would make it impractical to use in a practice setting

Thank you for pointing that out. I haven't had any issues so far, and now realize the traffic PEAKED after this post. I should have scaled the server, my bad.

> I'd rather it ask me what I wanted to talk

Good point.