Matthew Wise: How did you meet your co-founders?
Matt Mireles: Well, so that’s a funny story. Initially, I started the company with an old buddy of mine from, so I dropped out of Berkeley twice, I fought forest fires, I was a paramedic, I did all of this other stuff, then I ended up going back to school at Columbia and graduated from Columbia, and then after I graduated is when I started SpeakerText. I talked a buddy of mine from back in the day at Berkeley to help me build a sort of prototype. He was initially, yeah, I know you’re my friend. He was in grad school. He was kind of bored so he said, “Yeah, sure, whatever, sounds cool, I’ll do it for you as a favor.” And I don’t think he ever thought it was actually going to go anywhere. I was like this crazy friend who blew up an M80 in the dorms and got the bomb squad called, so he was like, “Matt, my crazy friend, sure, whatever.”
And so, we got like the most basic prototype out the door, and then I hustled it and showed it to some people, like I pitched Fred Wilson two years ago or something, and he was like, “Oh, hey, this is really cool.” And so, then that external third-party proof was saying this is really cool but like, dude, look at this guy. If other people think it’s cool, you should put more resources and time into it. Then, after we launched, it was a similar thing where we had this product that we had built.
In the meantime, I had gone to a job fair at Columbia, and I got half off for alumni discount and I got a tape and there were like real companies there and I show up in jeans and T-shirt and with 200 printouts of the grand SpeakerText plan for world domination. We had just the most basic prototype, $4000 in the bank, but I had a Mohawk at the time, and the little blurb in the job fair thing it said that our CEO will be there. He will be the guy with the Mohawk, and so the table just got flooded.
Literally there was a real company with real jobs next to me, and people were actually lined up to see SpeakerText and to meet me. They were blocking off access to the two tables next to us because of the Mohawk, or whatever. I got 150 resumes and there was a couple of undergrads and I said, “Hey look, we don’t have any money but we can give you an iPhone and pay your phone bill for the two months of work or whatever.” And it was an interesting project. They thought it was cool and then we got some other developers.
Then, after we launched, we got a bunch of big companies pinging us, corp dev departments at major networks who were asking if we had an enterprise version of this version. We were like, [makes laughing sound]. We literally built the thing for $4000. At that point, it was just totally vaporware but then I was sort of able to leverage that and that plus press that we got. Because I have a journalism background, I’m fairly good at making those relationships and whatnot. That, we were able to get further along and get some buzz, and then when it became clear that the original guy was not… he had never signed up to do a serious startup, well he drops everything and he goes for it.
Matthew Wise: And this is the guy who did the [inaudible 04:02] deal for you to help you get the product, the initial prototype?
Matt Mireles: Exactly. And so, one of the undergrads that we hired is Tyler Kieft, and he is truly exceptional just as a human being and a programmer. He had passed up Harvard for a scholarship to the University of Rochester and then had been working as a Python programmer for IBM since he was in high school. So, he’s like this totally bad-ass hacker who’s at this kind of out of the way school that’s good but nowhere near what you’d expect from the caliber of the talent he had.
So, he had job offers. I was like, “Hey, dude, I don’t know how we’re going to make this happen but f*** those jobs, let’s do this.” I made the offer and he accepted it and came on board the business as a co-founder with a very significant equity stake. At the same time, I had a blog that I was pimping out on HackerNews. I had met some of the guys at VentureHacks and they had liked my stuff and showed me the ropes about how to game HackerNews when it was gameable by getting up voting rings and getting on the front page of HackerNews which I had no idea but at the time it’s this amazing community that gets the most quality eyeballs you can imagine.
I came out to the Valley to check it out. I was running it from New York but Silicon Valley had this appeal and one day, I checked in on FourSquare at University Cafe in downtown Palo Alto which is like the classic startup spot in PA. A guy who had seen my blog on HackerNews tweets at me and says, “Hey, I read your blog and I see you on HackerNews and I like it. Let’s meet up.” So we have dinner. I’m telling him I have some guys that are not committing and I’m like, “What the f***?” and he was like, I know a guy, and it turns out this guy had gone to Carnegie Mellon. And so, he’s like I know a guy you should talk to, so he introduced me to this guy, Ryan Waliany.
Waliany and I had beers. He too had seen my blog on HackerNews so I wasn’t just like some totally random, non-technical dude, being like, please be my code monkey. He was like, oh yeah, I had seen my s*** before. Actually, you’re not normal. Then, he introduced me to this guy, Matt Swanson, who was his best friend from the graduate program in robotics at Carnegie Mellon where he was doing machine learning and AI research. It so happens that Swanson had already been working on a side project to transcribe business meetings using automated speech recognition.
Matthew Wise: Interesting.
Matt Mireles: And so, it put us in touch, and it was like two peas in a pod. We just totally hit it off with like a two hour conversation on the phone. At the end of it, he was like, dude, this is the longest I’ve ever talked to a dude on the phone before who wasn’t my dad.
Matthew Wise: Nice.
Matt Mireles: He flies up to New York. We go out for the weekend, get like totally sloshed in the East Village, and he’s got another month left to graduate before he graduates with his master’s, And so, I was like, it’s on. Me, him, and Tyler, we were either going to move to the Valley directly or if we weren’t able to get money, we were going to move to Pittsburgh. So, we got some money in Pittsburgh and eventually we came up here to the Valley in San Francisco. I guess the thing is I was in New York, I saw a lot of people who were like me, who were non-technical looking for a technical co-founder. I had friends who had even raised money, the external trappings of success who were still looking for a technical co-founder, and then three months later they’re still looking.
I had been working with my buddy, Bjorn, and Tyler long enough to realize that this is actually really important, and these friends of mine who didn’t have a real team, it didn’t seem like they were really going to go anywhere unless they were able to go back [inaudible 08:19]. And so, I saw how a lot of other people had screwed up, and also I have a knack for promotion and getting name out there and hacking those sorts of systems.
I was a writer. As a writer, you create a psychological experience for readers and you tell a story. You kind of grab them by the throat and make them feel things, right? That’s how I write, at least, and so I understood that psychology, the psychology of consumption. I was able to leverage that, and the beauty of coming to a place like this is if you go into publishing, everybody there can write, but if you can to an engineering-driven world, people just do not have the same level of communication skills that they do in a more traditional media industry.
And so, I think I’m a good writer but where I was really able to stand out is the fact that most people who were mathematically-inclined were not good writers. So, I was able to make a name for myself way easier because there’s just so much less competition in that sphere.
Matthew Wise: That makes sense and so, before we close, I’d like to kind of recap and really hear from you, your vision for SpeakerText and how you want to change the world.
Matt Mireles: The global transcription industry is a $21 billion a year industry. It’s very old-school, dinosaur of an industry. It has not been webified. There are not companies with open APIs, transparent pricing, there’s a ton of things that aren’t out there. There are a bunch of mom and pop shops in Eastern Europe and outsourcing shops in India with just like the worst web interfaces you can imagine.
We’re a web company, and that’s part of our DNA. So, we believe in building clean user interface, great simple user experience, have open APIs that people can get to, and we want SpeakerText to power transcription to be the default transcription service on the web. Part of that is going into established industries that are already out there, like closed caption, legal, but also there’s this huge explosion in online video like this, and all that content is totally not discoverable by search engines. You can’t do this sort of targeted, contextual ads that you can do for the rest of the web.
The problem is that there is no sort of reliable way to turn it into text, and so we want to be that service that people tap into to transcribe their videos, really high-quality, human readable, machine-readable, in an on-demand way and dominate that across the web and just really make it the default service and part of the architecture of the Internet itself.
Matthew Wise: Excellent. Thank you for being with us today, Matt Mireles. We’ve enjoyed it. We hope you will come back as a guest on FounderLY, and for our audience for those who want to learn about SpeakerText, become a customer and use their service, you can visit them at www.speakertext.com. Thank you very much, Matt.
Matt Mireles: Thank you.