After meeting Sun Yang and Zang AI in person, he said DJI also faced a lot of haters back in the day

"Facing the Doubts Head-On 👊"

"**Zang AI is launching a response column starting today. We want founders who've been critiqued to come face the questions directly — keep it real.

Last weekend, whether sensing a critique was coming or hoping to connect deeply with core users, Looki founder Sun Yang spent over two hours with Muqiu and Luozima.

To prevent any sparks of camaraderie from softening the review, I specifically asked Muqiu to finish her evaluation before meeting Sun Yang.

During the meeting, Muqiu and Luozima went hard — critiquing Looki's problems one by one. Sun Yang responded fairly candidly.

Given how many people actually asked me whether roasting Looki was part of their multi-channel marketing strategy, whether Zang AI got paid to trash them.

I'm floored. You people in AI are insane. We roast this hard and you still suspect we got paid. Should I run through every Club deal financing story next time to put this to bed completely?

Let me restate: Zang AI doesn't take startup ads. Our critiques are always genuine evaluations after actually using the product.

Enough talk. Here's our conversation.**"

Muqiu: So have users given you feedback about Looki getting lost?

Sun Yang: Yes.

We have two wearing methods. One is clipping it to clothes with the retractable cord shortened, so it doesn't swing as much. The other is magnetic attachment through clothing, but it can indeed get lost because different fabrics have different textures and surfaces. We considered lapel clips and such, but those don't work either. It's not just us — DJI and Insta360 devices get lost too.

I've tested the magnetic force. Looki's is about half that of a thumb camera. Given the form factor and weight constraints, we really can't add more.

The current plan is to add a location feature so users know where it was last used.

Muqiu: From my experience, Looki's approach of recording 9 seconds every 2 minutes doesn't feel intelligent — it's purely random recording.

Sun Yang: Around late December or early January, Story mode will launch with automatic mode, so you won't have to manually select anymore. Looki will assess the scene the moment you open your eyes, and if it has no value, it won't continue recording.

Muqiu: Looki's Q&A feature felt pretty useless to me — it doesn't handle complex scenarios well.

Sun Yang: One reason might be that you haven't used it long enough. I've worn it for about half a year. Without me telling it, it figured out my MBTI, knew I have two kids. When I said I had a headache today, it listed things like photos showing I worked late yesterday.

What you see with your eyes requires your brain to edit and filter. Looki gets information without human involvement — it has an additional perspective.

Today a user actually set Looki on their desk to film themselves from a third-person angle.

Muqiu: Why is Looki's privacy light designed to be so discreet, with the option to turn it on or off?

Sun Yang: Privacy lights aren't mandatory. Phones don't have privacy lights. For our product, the privacy light doesn't mean much because the device is right there. It's not like glasses where it needs to be obvious.

Muqiu: Could Looki be considered a covert recording device?

Sun Yang: It's already this big. If you insist it's a spy device, I can't help that. I actually looked into it — when DJI launched drones, when phones first got cameras, there were plenty of haters too.

Muqiu: Why didn't you choose glasses as the form factor?

Sun Yang: Today's supply chain can't support it. I researched glasses supply chains for half a year. Weight and battery life can't meet requirements, and glasses have overheating issues.

When we chose the form factor, I looked at hardware above the chest shipping over 1 million units — there were only three categories: action cameras, earphones, and glasses. This shows plenty of people accept this wearing style, so we ended up with Looki's current design.

Muqiu: Let me ask a serious question — why focus on the domestic market?

Sun Yang: Partly because of Plaud. Xu Gao wanted to do domestic, thought about it for over a year. By the time he figured it out, Mobvoi had copied it, DingTalk had copied it, so the market leverage was gone and new products didn't sell particularly well.

For Looki, there's a window right now. Regardless of how the domestic market turns out, we need to do it today, with limited resources. As for the future, we'll evaluate dynamically.

Domestic is also a hassle. From August until now, we've been doing various filings, adapting to different phone models, different app stores — and nobody cares, plus people骂你.

Not you guys though. (Muqiu: So you're pointing at me then.)

Luozima: Domestic users are broke too. I don't actually believe anyone would spend 40 yuan monthly on your storage.

Sun Yang: We have 30GB free. You can use it one day, delete the next.

Luozima: Like a fish recycling 7 days of memory.

Sun Yang: Honestly we'd lose money on that because there's still compute costs. But if you go through that much trouble and still like using it, you genuinely like Looki.

Luozima: So you're accommodating broke users.

Muqiu: You previously mentioned a feature where Looki detects your calorie intake to remind you about diet, but Looki can't accurately measure Chinese food calories.

Sun Yang: This scenario targets overseas more. Overseas it's all portion-controlled Western food, no shared dishes, and overseas models understand Western food better — not much Chinese food data. Chinese food has shared dishes, which is genuinely hard. But single-person Chinese takeout can give a rough estimate.

Actually, today's AI is better suited for things that are broad rather than precise — scenarios with high tolerance for error.

Muqiu: This is the gap between user perspective and founder perspective, right? From the user perspective, it's about what's useful to me. If Plaud is useful, I'll buy it. If it's not useful enough, it needs to be fun enough. But for you, maybe you first think about what AI can do today.

Sun Yang: Yes and no. What users want is based on current inertia. But when massive variables emerge, you can't build products by asking users, because paradigms shift — the meaning of products to users changes.

Today everyone is still living on dopamine products, still looking for the dopamine hit from AI products. But what AI is actually better at delivering is endorphins.

Whether Looki survives to see that day, I don't know. But the direction isn't wrong. Trust me.

Muqiu: From my usage, the Vlog scenario is pretty mediocre too. Why does most of your marketing feature this use case?

Sun Yang: If you ask me, Vlog definitely isn't Looki's core function — AI understanding is. But Vlog is the most tangible consumer-facing application. Users know what a Vlog scenario is.

You can think of it this way: when Apple released the first iPhone in 2007, there was no mobile internet story. They just told users it was a better iPod.

Muqiu: Then for most white-collar workers, their daily work routine might be identical every day. Is there value in generating Vlogs from such scenarios?

Sun Yang: I think there are still differences every day.

If your life is truly identical every single day, shouldn't we reflect on our lives?

Muqiu: You should put that in your marketing copy. Pretty inspiring.

(Article images generated by ChatGPT, purely human-written responses 🤓)