Duolingo is losing one of its hardcore fans

Evan Edinger has earned quite a reputation in the YouTube language community as a convinced Duolingo fan. As an early adopter, he has been using Duolingo for over nine years continuously. Evan is an American living in London. 

His first choice language was German and he claims having become fluent in German using mainly Duolingo, but also supplementing it with different content (film, TV and books) to get more exposure, crank up his vocabulary and getting more acquainted with the German culture, society and even politics. These subjects fail almost completely in the Duolino environment.

Several years ago, he also started learning Spanish. Watch the video: I did Duolingo for 2000 days. Can I speak Spanish? Note that the Duolingo Spanish course is about the longest and most detailed one you can find. This is why Duolingo claims it can bring you to B2 level fluency at the end of the course. Considering the time and effort you will have spent by the time you finsish that course, anything short of B1+ would be a system failure for Duolingo.

But lately, Evan has become dissatisfied about Duolingo, more precisely the fact that Duolingo’s AI Update is Quietly Ruining Everything

Why is everyone Deleting Duolingo?

Recently Evan got himself a reputation for distinguishing between a human written text and an AI generated one. In following recent video he goes into detail on this.

I can spot AI writing instantly: Here is how you can too!

There is one more video I would like to include here, not only since it links to the first video of Evan Edinger, but because it invalidates the decision making process by Duolingo to opt for 'AI-first'.

The decision was taken based on a flawed statistical study. Flaws are manifold: not including a 'control group' which is not using AI. There is a 'survivorship' bias, since only confirmed Duolingo users (with moderate to long streaks) are include in the study. He didn't quite go as far as to claim that the statistical study was tampered with and that the selections made were such as to corroborate a decision already taken. 

Just a few remarks: 

• The video is made against an artificial background, causing some flaws when the author is moving his hands.
• The video ends with a commerical section (take it or leave it).

And one more caveat: I might not have included this final video if I did not have a background in statistical methods.

Duolingo's Big MISTAKE - The Flawed Study Behind "AI First"