Dear backers,
We wish you and your family a happy new year. We also want to start the year by showing you what we’ve been working on for the past 6 months: the new Lily.
As you saw in the video, we’ve turned Lily the smart speaker into a 3D animated character that can move, speak, listen to you, and teach you Chinese. The video was not made by a video company, it’s a full 3D rendering made by our team and it runs on any phone without powerful CPU/GPU.
Later in the update, we’ll tell you more about the new Lily. But before that, I want to take some time to reflect on the past year.
2020
The past year has been a tough year. For the world and for us too. Our hardware efforts combined with the pandemic’s economic turndown put lot of pressure on our finances. We had to pause the development of our smart speaker. Throughout the year, our team went from 20 people to 10. We had to let go of our hardware team, our embedded team, our community manager and our cartoonist. Alexis and I stayed in Shenzhen for the project — we haven’t seen our family for 18 months — and in July, we decided to pivot Lily to an app. I know that we’ve generated lot of frustration with that decision and I want to apologize again for that. But looking back, if we hadn’t done it, we would have shut down the company by now.
It’s the hardest and the best decision we’ve ever made for the Lily project and let me tell you why.
The new Lily
We had to restart. Designing an app is so different from designing a smart speaker. You want users to focus on the screen instead of just speaking in the air.
We came up with the idea of turning Lily into a “real person”, a teacher with flesh and bones who can speak/listen and teach with gestures, facial expressions or by manipulating objects.
But we didn’t change the core concept of Lily: you learn Chinese with your voice, no need to touch the screen of your phone.
We quickly hired a 3D animation team and we started giving Lily her new shape:
The more we were progressing, the more obvious it became that Lily 3D would offer a better learning experience that Lily the smart speaker. We could retain the voice-control experience while using visual animations to make the lessons more memorable.
For example, it’s easier to understand that Lily is saying 现在小雨 (it’s raining now) with this animation:
Every time Lily speaks, we integrate live Pinyin subtitling on the screen which makes it easier for beginners to follow what Lily says.
Another advantage is that we can make the new Lily only speak Chinese, which is more immersion into the language. What Lily was saying in English can be replaced by an animation or text on the screen.
Here a quick comparison of the two Lilys:
Some plans for 2021
The 1st thing we’re going to change is to stop announcing timeline we’re not sure to reach. We’ve been working on the new Lily for 6 months now, it looks really good and the AI is working well. We’re still fine tuning the app beta and we’re targeting to launch it a few weeks after the Chinese New Year. We’ll reach out to you, beta testers, directly by mail.
We hired an Android engineer and we’re already working on the Android version.
We’re looking for some help in acoustic modeling (a part of Speech Recognition) to improve the accuracy of our pronunciation detection. If you know anyone who is working on this, we could love to collaborate or do some remote freelancing.
A few words about crowdfunding.
Lily crowdfunding campaign ended in February 2019 and despite our struggles, you’ve been incredibly supportive: we still receive 500-800 likes on each update and more than 100 backers have asked to join Lily’s beta.
However, we also have backers asking for refunds. We did refund some people over the course of the campaign, especially people who backed a pack of 5+ Lilys, but today we are not in the capacity to do more refunds.
A crowdfunding project is very similar to a startup. It has potential but it’s risky and you can lose your money. When you backed the project on Indiegogo, you had a warning button. The project is only at seed stage, it goes through trials, it morphs along the way, and survival is always on the line.
I’ve made so many errors running this project: hiring too fast, focusing too much on the hardware, setting unrealistic timelines, etc… But during the past 2 years, we’ve spent every working day and every dime of the fundraise on developing Lily. We still do and our focus is still the same: build a better Chinese learning tool with voice technology and AI.
We’re hopeful we can get there in 2021.
Jie and the Maybe team.
