Today, I came across an internal email from Duolingo’s all-hands, announcing that Duolingo is going to make a significant platform shift – AI-first. Luis, the CEO, believes that AI is happening and changing much of how they work today, they can’t wait until the AI is 100% perfect, and would rather move with urgency and take occasional small hits on quality. Also some constructive constraints are raised from him, one of which is:

We'll gradually stop using contractors to do work that AI can handle. 

Along with that, I recall that Meta was also planning to replace the junior software with AI.

Moreover, Luis said, “Duolingo will remain a company that cares about deeply its employees”. However, I believe the role of current engineers may no longer exist if we don’t adapt how we work. In other words, if we are able to produce 10 units per day, we, as engineers, should shift to produce X times units per day with AI.

It’s really crucial for us to learn to enhance our skills and improve work efficiency with AI to the greatest extent possible. Recently, I’ve been doing some development work with Cursor, but it hasn’t been as efficiency as I expect. The main reasons are that I’m not familiar the frontend work and a lack of expertise in working with AI tools. That said, over the past few weeks, I’ve found that working with AI has become increasingly easier. I believe this is where the real change lies.

Also, it makes me feel somewhat excited and optimistic about the future. The idea of a one-person company seems more achievable than ever, and AI is opening up new possibilities for creating products that we could never have imagined.

That’s all for now.

The original email is shown below: