In Otter Words: Edith Yeung

10:59PM Sep 15, 2019

Speakers:

Ross Rubin

Edith Yeung

Keywords:

otter

china

people

wechat

edith

world

satya

literally

chinese

nuance

startups

notes

ai

transcription

voice

case

understanding

super

google

called

Hello and welcome to In Otter Words, an audio program where we ask people in different roles how they make use of real time transcription as it is enabled by Otter. My name is Ross Rubin. I am an industry analyst who has followed consumer and emerging technologies for 20 years. Our guest today is Edith Yeung, a well known VC with 500 Startups who has spent extensive time both here in the US and in China. Prior to joining 500 Startups, Edith was the head of marketing at Dolphin browser, where she amassed millions of users while competing against some pretty well known companies such as Apple and Google with their default browsers. So, let's just get into it: Edith, could you tell us a bit about how you use Otter?

Yeah, so I am definitely an avid user of Otter. And I actually... So my day to day. I either spend time talking to CEOs and founders for potential investment or other fellow investors to catch up on deals. So, a lot of times there's literally information overload. So, especially if it's a first time pitch, I would usually think and I tell them, "Hey, you know, this is our first time meeting and I would love to take notes." But at the same time, especially when you're taking notes, let's say on your phone, it seems like... it just seems, like, distracting. People thinking that you're probably texting other people and not being focused. Instead, I will focus and tell them and usually there's a wow factor, because they usually never heard of Otter before. The fact that I am taking notes and our conversation is literally just right there. So, it gives people a little bit of wow factor in what's going on.

It's pretty amazing when you see it in action and not only is it doing such a great job as it goes but it's correcting itself in real time as it figures out what people mean, which is very, very neat.

Yeah, it's amazing. So I usually use it for two main things. One is all my professional meetings, especially face to face. I will just tell them, "Hey, I'm taking notes but I want to look at you in the eyes [rather] than me looking at my phone", or, like I have a notebook with me but it's just not the same because then I have to retype everything back to, you know, my notes. So, either you use actual notes or Evernote. In this case, it's super convenient. That's one. And second, it's actually conferences and events, It's a strange use case because although there's no conversation going on per se, It's mainly just I'm looking at somebody speaking and I really, really want to take notes but even though, like, I don't listen fast enough so I use Otter for that.

The conference application is a very cool one. I haven't heard that one before but it makes a lot of sense. I think I've used it that way. I think in both cases, when I'm getting a sense of is that, you know, you're trying to take some notes and this just provides some level of insurance in case you didn't capture something that you later realize maybe was more important than you thought, right?

Yeah, exactly. In that sense, it's awesome. And, like, I wish... I do have a feature request. I haven't talked to [Otter CEO] Sam {Liang} about it but which is, like, this conversation, right? When you're actually on the phone, you can't actually use Otter for that. I get it: Like, technically, it's definitely much harder but I would love that. Then, everything can be put back into one place.

Right! So as I mentioned earlier, one reason I wanted to use Zoom for the call is because I can take the recordings of everything you've said and everything I've said and then feed that into Otter and then it will do it off the file. But yes, it would be cool if it were built right into some kind of calling program and you could see the transcription happening in real time.

Yeah, totally.

Great. So, I know some of the work you do with with 500 Startups and I've been looking at some of their efforts. I see that they are becoming more active in terms of entrepreneurial education efforts. Have you seen a potential application in that, to your point about conferences, and potentially even trying to expand the knowledge or learnings beyond the folks who might be in the room at that particular time?

Yeah, I think a few things. I absolutely think so and I completely agree with you: That's a great suggestion. I think, in general, if you think about the form factor of learning, right, 500 certainly do a lot of face to face workshops. We also have done webinars and, like,, podcasts -- there's a lot in the investor community that do all the above. It's one of those things we're transcribing into text format. And actually, if you think about it, it's interesting. Different people have different ways of learning. Some are more auditory, some of them by seeing, some of them by doing, and some people are more kind of static.

I think, you know, being able to use Otter, which is voice and text, you at least attract two types of populations in that you're not just listening or not just watching, you're also reading. [unintelligible] So, in this sense, yeah, absolutely, I think more and more education. You should ask Sam for I actually recently went to really, like, a very prestigious conference and they had a team that actually, like, with the original talk videos? They actually have a team not only do transcriptions but also transcriptions into multiple languages. So I'm like, I say, "How do you... What do you guys, like, use?" And they're like "Oh, humans!" And I'm like "You should not use humans!" So, I showed them Otter and they were like "Oh, wow, this is amazing!" So, in my head one is, obviously, much easier to talk in English.

So, this is what Otter is great for: to start. I can totally see, as another feature added in the future is also, like, simultaneously be able to translate into multiple languages, text-wise. If that happened, it would be amazing! Well, that's just a little bit more personal to my family, actually. My parents, in particular my mother, she doesn't speak English and a lot of my work is in English. So I felt like, she's definitely much better reading so the fact that... if I'm going to take Google Translate, which sucks anyway -- it's completely lost, literally lost in translation. But there's so much more. Yeah, like, that's what Otter is really good at because it's very conversational. It felt correct, You're not talking to a robot, you know?

So, those are some really interesting points. Two things I wanted to follow up on in terms of comparing Otter translation and human translation is that, even though Otter may not grab 100% of what's said or it may, you know, not be able to capture all the punctuation nuances that a human might know, I think one of the big advantages is just the speed, right? And so you're getting 90%, maybe more, of the way there and then if you want to refine it so that it's more like, you know, human writing, you can do that after the fact, but it's just going to get you to that endpoint a lot faster, seems to be, yeah, one of the arguments.

I know that you really have your finger on the pulse of the startup landscape in China, so I wanted to ask: what are some of the trends there that you're seeing that are having an impact on how colleagues communicate in the office or telecommute or how meetings may be changing or how the workspace may be changing that's relevant to this kind of technology?

This is a little bit weird. So first off, in China, with office or professional communication, no one uses email. And most people, 100 percent is on WeChat. Even though Alibaba tried to have their own version of WeChat for office, like a Slack, it doesn't even work. Everybody kind of defaulted back to WeChat. There's another nuance, which is kind of weird, is sometimes, as we know, the Chinese government, they do monitor what you say. So I found that sometimes, like, some of my friends, honestly it doesn't really matter, because I think the Chinese governement will figure out your voice anyway. But I do think that I have seen more and more, sometimes people communicating not necessarily through text because you do actually can do machine learning faster with text. You can do more, like, with voice, because these voice messages on WeChat people I can I can just chat rather than text. The core communication was all WeChat and it's 100 percent identity link because of many reasons. But, I think WeChat is also very smart with their payments. A lot of it is literally is like one gigantic, super app, you do everything on there. Having said that, I do see actually more and more folks, even from China, are starting to use Zoom as well, particularly for, like, international calls becasue not everybody overseas uses WeChat. Plus, when you do live calls on WeChat, it's definitely not as stable. So, I do actually see some popularity pick up for Zoom in the Chinese internet population, more or less, because there's so many international friends and partners that they work with.

Right, right. And then, to me, that sounds like an interesting contrast between how things are maybe a little different in terms of work versus personal tools in China versus here where Zoom is a very corporate-focused this tool here and WeChat, of course, is a huge consumer tool there but it sounds like it really has crossed over into the business world as well and is ubiquitous there. You described kind of the integration of all this functionality and WeChat. And I was going to ask if you saw a similar kind of integration happening in other parts of the world. Maybe not in one app but sort of this bringing together of commerce and messaging and breaking down these barriers of translation and voice.

Yeah, so, around the world. In Japan is Line: Japan, Taiwan, Thailand, and Indonesia. These are, like, Line countries. They don't use WeChat. They use mainly Line. And then, in Korea, its KakaoTalk. Both of these have the payments built in. They also, like, use a lot of text, also have voice functionality and they're actually really, really similar to WeChat, except, I mean, obviously the China population is at 1.4 billion, so there's just more people, more data. But I do find Line is actually fairly strong in certain countries. And then, of course, WhatsApp, to me, in places like India, Philippines, and also actually very strong in Indonesia. Hong Kong, even though it's actually part of China, but most people in Hong Kong use WhatApp. So I really think there's some geographic, like WeChat-like WhatsApp-like version, but usually compared to WhatApp... I don't know why but Asians in particular are just really into to stickers and all kinds of GIFs or GIFs, however you pronounce it. They're just really into, like, more animation. So you can see a lot of really funny, cute features and things that you don't normally see in the American versions of Messenger.

Right, right. And yet here in the US, I think Facebook has been trying very hard to turn Messenger into that kind of super platform.

Yeah, they've tried and I do think.... I have a lot of admiration for the Facebook team. I mean, Facebook wouldn't be Facebook today even though you can criticize them about privacy and all that. But they still absolutely have done an amazing, amazing job. Like if you say, China captured the 1.4 billion people, Facebook has captured the rest of the world. So, however I do I think that, just, the Asian behavior is different. There is no right or wrong. It's just people like cute things and they like to idolize; here's a very, very strong, like, pop culture. I think stronger than. The US has a very diverse [set of] interests in things. It's different. I've found, like, a lot of the young people in China, they all, like, if they're crazy about this one person, it's literally insane.

Well, that sounds like a very cool nuance in terms of understanding some of the cultural differences.

There's this super young singer in China but their fans are, they're called Da Ma (中國大媽, literally translated 'Chinese Auntie), is the "Auntie" and they're much older and they literally love that particular pop star so much, they bought out the whole plane to travel with him. I'm like, what? That's beyond fan club!

Wow. That's experiential marketing right there, yeah. That's great.

They just want to, like, follow him wherever he goes when he does his concert so, like, these ladies are like "Yeah, we'll go with him!" Like, on the same plane. What? It's strange.

That sounds like a high level of trust as well! All right. So, just to sort of close things out, I mean, where do you kind of see this this all going? As the accuracy improves, do you think that we'll get used to a world in which recording things sort of becomes the norm and the default and people will recognize that there's a benefit that outweighs some of the privacy concerns? Or do you think that's going to continue to be an issue?

Related to privacy, I was reading this book by Microsoft's current CEO Satya [Nadella] I think... I forgot the name of the title. But I read the book and one of the messages in the book which I found interesting—I don't know. Maybe I'm reading too many books but—I think Satya did mention is, there's this whole concept about, like, ethics for AI. And, I actually don't... Maybe I do spend a lot of time in China. You always have this understanding:, Sure, the government is monitoring and watching but most of the Chinese population take the attitude of "I have nothing to hide. I know that they're watching, but so what?" Versus here, I think in the US, obviously, there's a lot of controversy.

But I really liked what Satya said, which is, like, to have that sort of framework to build. Like, what is, like, a "right" AI, is really having certain ethics even for the people who are building it and not only seeing it as a pure technology play but obviously, like, you need to actually have certain cultural understanding, legal understanding, Like, respect for religion or different voices in the world to be able to build the right kind of engine. And so, in this case, I mean, I come from tech so I don't think about all these things but there are a lot of psychological and social impacts now being, like, being impacted, like, because of social media and, like, AI power generated. Just look at Cambridge Analytica: It's complete manipulation. So there's all these things where It's a really fine line but it's a really good thing, like, to put it out there and with different kinds of backgrounds to help set the right sort of rules.

I don't actually completely agree with GDPR, either. I think that's too strenuous. But there's no right or worng because I do think that every culture has a different bar in terms of privacy and access and all that but it's good to, like, talk through. I'll tell you one example to end it. There's when, whenever... I have to say this there's a lot of misunderstanding of what China AI is about because it's really tough for AI practitioners in China. They feel like they don't have enough talent, they don't have enough data. Just because, like, the world sort of portrays you can have access to data -- not really. Like, not individual companies don't have access to everything. They don't.

So, like, if I speak Chinese, the voice engine, AI engine, the one that I use, is called iFlyTek. They're amazing in interpreting Mandarin. But when I speak English, it's horrible. They don't get it. And then also, let's say In comparison, if I use the Google... There's this thing I keep playing... There's this image where you take your own photo. You upload on it on this Google thing and then it will sort of, like, match you with a painting. I don't know what it's called.

Oh yes, I've seen that. Yeah, yeah.

But they always think I'm like a Japanese painting. And I'm not Japanese. So, just this simple example is, you know what, like, the engine is only going to be as good as—regardless if it's voice recognition or image processing—it's only going to be as good as the understanding culturally, region, religion makes you think to be able to build the right type of AI. So I think the couple of examples I just gave is more like, yeah, because the developers either don't understand the English language, in this case, or they don't have enough Asian data for images. So, it's only as good as the data they feed them. So, that's all.

Oh! No! That's great! Well, thank you so much for your time today. It was really great to hear your thoughts and I'm so glad we could connect. That's going to wrap up our show for today. I want to thank our guest, Edith Yeung. You can visit Edith's website at edith.co and follow her on Twitter @edithyeung. That's Y-E-U-N-G. And you can also visit the website of 500 Startups at 500.co. You can follow me on Twitter @rossrubin. That's R-U-B-I-N. Visit my company's website at reticleresearch.com or listen to the podcast I co-host at techspansive.com. And of course, please check out Otter on the web at otter.ai. Follow the team on Twitter @otter_AI and be sure to download the app if you have not at the Apple App Store or Google Play For In Otter Words, I'm Ross Rubin. Thanks for listening.