In Otter Words: Sam Lessin
11:59PM Aug 28, 2019
Speakers:
Ross Rubin
Sam Lessin
Keywords:
otter
people
coaching
read
tool
fin
transcript
interactions
building
knowledge
company
humans
tasks
faster
technology
speak
wrote
future
sam
happening

Hello, and welcome to In Otter Words, a podcast that looks at the way people are making use of real-time transcription. I'm Ross Rubin principal analyst at Reticle Research, and I've been tracking and writing about the tech industry for more than 20 years. Continuing our focus on investors, today we're going to speak with Sam Lessin of Slow Ventures, Sam first became known for developing a file transfer product called Drop.io that was sold to Facebook. And more recently, he's been looking at ways to improve the performance of knowledge workers, which we're going to get into in a little bit. But for now, Sam, thanks so much for joining us today. Why don't we start out if you could please share some of the ways that you personally use otter
Remembering to dos and remembering things to get done or just ideas or strategy ideas you have, or thoughts that come up, you know. I frequently will, like, just speak rather than type. And then share that with, you know, with a few friends or partners or colleagues, depending on what we're working on. And the nice part is it used to be that speaking is faster than writing for most people, but listening is much slower, right, then reading. And so the nice thing about Otter is it does this great thing of excellent translation of speech to text, but then, also, kind of with a hybrid approach of having the text plus the audio to refer to, you basically make sure that it actually is faster for the person generating content and faster for the person receiving the content to get to the key ideas and go and kind of get things done when it's kind of a memo style use of Otter. And then, you know, or like, you know, interviews, like, you know, like this, the ability is, like, so much faster, to kind of both speak, and get full concepts out and then also be able to. like, parse data, text jump to the right point, review what was sad, and use as data. And then I think for the team use cases is, like record everything: your company meetings, etc. I mean, like, look, we've gotten to a world where what percentage of meetings are on Zoom? It's, like, so high among, like, tech-forward companies, right? And that's because people are remote. That's because you know, people are dialing in, etc. And that's good. But the context is still lost; you still have to like, be there. And then everyone forgets what was said. So I think the vision and saying, look, we can get to a world where meetings happen. We don't have the same meeting 16 times, you know, context isn't lost. Like this is an incredible goal and dream, I think that, you know, kind of relies upon having great, both recording and then great speech to text.
Yeah, I love that dichotomy about optimizing both for the capture of the information through voice and the consumption of it through reading. That's really an interesting point.
If you personally were just optimizing for your own experience and no one else's, you'd probably always want to speak and read. Right? Like, that's what you'd want. I don't want to... People are slow to listen to. Like, if you could just read. If I just had already said everything I was going to say and you can read it, you'd get much faster through this and cut to the heart. We're much faster readers, visually can do more. So the problem is, obviously, we live as humans and society where you can't just do what you want, right? You have to kind of also model to what other people want. And so, you know. But we can have that.
So I suppose one of the advantages of the reading is that, sort of semi-consciously, you're always kind of, like, looking ahead a little bit of where you are. And that's impossible to do with voice. So...
Correct.
Yeah. Cool. You touched on some of the sharing and collaboration you do with your teams, as well as some of the collaboration that other teams --, forward-looking companies, as you refer to them -- are doing. You talk about some of the benefits. But could you maybe talk a little bit about maybe more of the different kinds of scenarios? So, for example, one of the different kinds of tasks we've seen Otter used in is kind of sales review meetings. So it's easy for the team to capture the kinds of current issues that may be happening with the prospect pipeline, I imagine. CRM kinds of applications are another one. Any other kinds of tasks that you're seeing in that vein?
Yeah, I think there's... What I call what you're describing is coaching. So, if you think about it, there's a lot of people who have interactions externally, you know. That can be salespeople or BD conversations, or whatever. And, you know, especially in fast-paced environments, where things are changing, and there isn't one set scripts, but there are goals, and people are trying to improve, you know, it's one of those things where, like... I actually spent a lot of time on this in a company I'm building called Fin. But, like, the upside is, like, knowledge workers in general get the worst coaching. You're on a sports team, or you're you know, doing a job that is kind of more, you know, manual, like, or, you know, manufacturing, whatever. It's, like, there's data; that data lets you be coached, right?
If you think about how most knowledge workers are coached, what happens? If you're lucky, you have a one-on-one once every... once in a while. And your manager, your boss, asks you how it's going and maybe you review some outcome metrics, but that doesn't really help you. If you're a salesperson [and] you're not closing deals, right? Your boss saying, like, "Close more deals. Like, hit the number." doesn't really do much for you. You need help on in the process. And the problem is that you can't have someone shadowing you all the time, right? It's just impossible. And so, you know, it's too inefficient. But, like, if what you do is say, hey, I can record everything and get feedback on my process. right, that is, like, hugely valuable. And I think that's definitely part of the future.
So you mentioned Finn, I was actually a customer of Fin when it was more of a consumer-facing offering. And I've had some discussions about the potential for tools like Otter in this sort of personal assistant space, it really seems to be a very good kind of match, yeah.
Yeah, I use it with my assistant for exactly that. Part of what Fin you know, did. Fin is a business. We started out with this thesis, which is the future of work is human plus computer, for sure. And we built an assistant product because it forces you to figure out the broader set of how do you create the, you know, the ranking and the routing of who should work on what and and it helps you understand and think about the workflow, how you build and classify the workflow? And then it also thinks about you as a holistic system. So how do you create, basically, like, industrialize, is the way I would think about it, like, knowledge work, which is, you know, I would argue, knowledge work is largely pre -ndustrial right now. You know, we've moved that business, to use to the point of about coaching, very much towards the biggest insight we had, which is, again, we have all these assistants and they're doing good work. But unless we understand exactly what they're doing and their process, right, moment by moment, click by click, you know, interaction by interaction, we can't possibly help them do a better job and optimize the overall system.
And so we basically ended up building this very deep analytics, and, like, QA toolkit that other teams can use, right, to basically optimize any sort of knowledge worker, which could be, you know, sales or some kind of support, etc. I think, you know, Otter is, like ,a key input to that, right, in terms of, like, that future ecosystem where, like, if you want knowledge workers in general to have coaching and improve at their jobs, and also you want great interfaces, where, like, as someone who's asking for an personal system task, I just want to ask for it, right? And I want to potentially give you a lot of context as I'm thinking about it, but I don't want to have to, like type it out, right? I just want it to go. Like Otter is great for that. I have, you know, someone to help me out investing on the inside investing side of what I do. And, you know, I frequently just send him Otters, you know, to get a thing done or to, you know, type up a note on a company or share something with someone or do something.
Cool. Cool. So I'd like to switch gears a bit and talk about possible ways that technology could help coach people, even before they enter the workforce, because it seems like there's a number of applications for students, for example.
Certainly, I mean, I was I'd had it right,?. You know, because, again, it's so slow to, like, rewatch lectures, right? And even if you're rewatching, you know, it's interesting. So I'm quite dyslexic. I didn't learn how to read until, like, third grade. I had to do it by flash cards because I couldn't sound things out, because, like, you know, that's part of what it was. And I think to myself, man, if I had a tool that could have basically hybridized spoken word, right, what I'm hearing with actual reading, right, I think would have helped me a lot, both, you know, learning to read probably. But even more importantly, then, when I was in class, or doing an activity, you know, the reinforcement on two channels is super helpful for like, efficiency of knowledge acquisition, you know, watching a movie and putting the subtitles on. Sometimes, people do that for when there's tough accents, you know what I mean?, Like, you listen to some of these shows, like BBC shows, and I know a lot of people will turn on the subtitles. And they're like, look, I mostly understand it. But if I really want to understand a show, it's good to have both streams of understanding, like, that's clearly a better way to consume information you really care about.
That's a great point. And one of the interesting things there is that even if the accent is so extreme or so pronounced, that Otter may not be able to do the dead-on transcription out of the box, it's just one of those side benefits that I think you get by committing to the tool. So, if you're in the habit of using the tool, then you're in the habit of recording, and you'll still be capturing that and can review it and send it to humans or whatever.
Yeah, and, look, one of the things I'd like to call out about the company I really liked about their approach is: There are a lot of companies out there that are working on, obviously, you know, speech to text. And Otter is more accurate by every test I've done, which is great, right? Like, their technology is better than any other technology out there. But here's the other thing that I really like about the company, and the approach is it's honest. So if you go and talk to the kind of people who are working on speech translation tools, you talk Microsoft (not that there's anything wrong with, like, Microsoft or Google or whatever), they'll tell you that like, oh, their speech to text translation is 99.8%. They'll give you some ridiculous statistics about how their speech translation is,, like, better than humans, right? And the reality is, I don't question the science, I'm sure that, like, there's some slice in the universe where they're technically correct.
But from a practical perspective, right, if you just had a transcript and nothing else, they're pretty hard to read, right? And so I like that Otter is, like: Look, our technology is better. But also we're hybridizing the consumption experience in a way that, like, we acknowledge the fact that, like, it's not perfect, right? And that means that, like, you can still pick up and understand what's going on, right, as opposed to companies, I think, that are overly technically there's like some hubris to it, but it's not accurate. But if you ask them if it's accurate, they'll tell you, it's accurate. It's not. Like, look at it.
Academically accurate, right. But I love the point you're making about readability. Because, you know, having looked at a number of these transcripts, I would agree that the tool does a great job on kind of a word-for-word basis, or phrases or what have you. But there's still maybe some work to be done in kind of understanding what a sentence is and punctuation -- those other things that we kind of take for granted when we're reading, right?
The other thing is, like ,actually, humans, if you listen to them, sometimes, if you, like, just wrote down everything I said, it wouldn't make sense. Seriously, it's like there's this thing is, like, so much of language is, like, in context. And we don't fully understand that, like, literally, there's a huge difference between structure, written language, which is meant for legibility and consumability, and spoken word, which again, like, if you just wrote just purely what I said, and the pauses I put in, it would be very confusing to read.
I wanted to maybe hear your thoughts on some complementary technologies that you're seeing. We talked about Zoom, when you mentioned the idea of someone coming back into a conversation that was recorded and viewing that transcript. I've thought about kind of a Slack or a, you know, Microsoft Teams kind of scenario where someone could jump in on the transcript. Obviously, a lot happening right now with language, real-time language translation. Anything you're kind of seeing that you view as particularly complementary in terms of evolving technologies?
I mean -- I'm trying to think off the top of my head -- not directly. It's not that those things aren't important. I just think what's happening is that people are all sorts of, you know, developing their own tool chains. You know, what I always like to say, you know, people who think that the future of work software is the Microsoft Office suite, you know, is just sorely missing the point. Today, what we've done is we've taken technology, and we basically recreated a bunch of tools that we had that were paper tools, that Hooray. Ask yourself the question, which is: In the future, will all of your work interactions be recorded so that you can increase sharing, like, which increases productivity by getting people on the same page? And it gives you the ability to get coaching, like on the interactions and how to improve them more rapidly rather than just spending your whole career doing it as it's done without any feedback? Or good coaching? Like, will you end up in a world? The answer is, in my mind, definitely. Like, there's no way that's not the future. Right. And the question is, if that's the future, what are going to be the key collaboration tools and products that bring you there? Right. And, you know, to me, Otter is on the right path, for sure, which says, look, it is designed, you know, with voice and the natural conversation at the center, but you know, fundamentally is building a collaboration platform. And that's exciting.
Definitely a lot to get excited about there. Very cool. All right. Well, that's going to wrap up our episode for today. Sam, thanks so much for joining us. If you'd like more information about Fin, please check out fin.com. For more about Sam or Slow Ventures, please check out slow.co You can find more about my company Reticle Research at reticleresearch.com. You can also follow me on Twitter at @rossrubin. Or if you'd like to check out my podcasts to get our take on what's happening in tech, you can do that at techspansive.com. And of course, check out Otter on the web at otter.ai, follow the team at @otter_ai on Twitter. And be sure to download the app from the Apple App Store or Google Play if you haven't already. For In Otter Words. I'm Ross Rubin. Thanks so much for listening.