> 1. That Duolingo has the ability to teach a language to the US Foreign Service Institute standard. I have reasons to doubt that.
My point is that how many people here have put in the requisite time to even find out. I’d wager less than 0.01%. Duolingo have come out and said their long term aim is for each course to get you to B2 standard eventually. I’d say they’re not there yet but they’re on their way. The other thing to bear in mind is that we’re looking at this from an English perspective. The Duolingo English language test is widely accepted at universities as a demonstration of English proficiency so they must be doing something right in that regard.
> Several years ago I completed the entire French tree in Duolingo and I am in no way proficient: I can understand a bit but I can't really formulate a sentence beyond the absolute basics.
What do you mean by “completed the French tree”. Do you mean you went through and unlocked every lesson once. Or did you gold every single lesson? There is a significant difference. You unlock extra vocabulary and the lessons became more focused on writing and listening comprehension when you complete the higher levels of each lesson. This is now baked into the new format as you go along.
> Now, there's clearly a significant cost difference between the two approaches. My point is that an hour of study isn't really a good metric when it comes to comparing methods of study.
I disagree with this. Hours of study is an important metric. You need proficiency on a standardised test along with hours studied for a true measure of method efficiency. And as you mentioned cost is also a massive factor. Most people cannot afford 65 hours of one to one tuition.
I really don't understand how any method without regular conversational practice can hope to get a learner to B2-level proficiency. (As measured by a standardised test.)
They've added written conversational practice on the new AI plan. The courses aren't at B2 level yet, but it doesn't take much imagination to see that a complete conversational chatbot with verbal and written modes is not very far off. It's pretty much already doable with current off the shelf AI offerings. Eventually you'd probably even be able to tailor the accent to the exact region you're planning on visiting.
My point is that how many people here have put in the requisite time to even find out. I’d wager less than 0.01%. Duolingo have come out and said their long term aim is for each course to get you to B2 standard eventually. I’d say they’re not there yet but they’re on their way. The other thing to bear in mind is that we’re looking at this from an English perspective. The Duolingo English language test is widely accepted at universities as a demonstration of English proficiency so they must be doing something right in that regard.
> Several years ago I completed the entire French tree in Duolingo and I am in no way proficient: I can understand a bit but I can't really formulate a sentence beyond the absolute basics.
What do you mean by “completed the French tree”. Do you mean you went through and unlocked every lesson once. Or did you gold every single lesson? There is a significant difference. You unlock extra vocabulary and the lessons became more focused on writing and listening comprehension when you complete the higher levels of each lesson. This is now baked into the new format as you go along.
> Now, there's clearly a significant cost difference between the two approaches. My point is that an hour of study isn't really a good metric when it comes to comparing methods of study.
I disagree with this. Hours of study is an important metric. You need proficiency on a standardised test along with hours studied for a true measure of method efficiency. And as you mentioned cost is also a massive factor. Most people cannot afford 65 hours of one to one tuition.