Daily Standup 06
– Elon Musk buys Twitter
– Beacon for free speech, open-source algorithms and public reaction
– Horizontal speech-to-text and NLP companies like Otter.ai, Rev.com and Happy Scribe
– Vertical speech-to-text and NLP companies like Chorus.ai, Gong, Kintsugi and Dovetail
– Developer-focused companies like Amazon, Google, Deepgram, AssemblyAI, Hugging Face and Symbl.ai
#toronto #portcredit #cofounder #saas #startup #seedfunding #startupcommunity #amazon #apple #google #founder #seedround #tiktok #dailystandup #naturallanguageprocessing #twitter #speechrecognition #stockmarket #earningsreport #psychedelics #otter #kintsugi #happyscribe #rev #chorusai #gong #dovetail #speechtotext #nlp #symbl #assemblyai #huggingface #assemblyai #symblai #deepgram #aws #elon #elonmusk
Hello hello Tyler Bryden here lots of things to talk about today. One thing is just like I guess a news story that you can’t miss out on. Elon Musk buys Twitter. This is the commentary. This is what’s people are buzzing about right now.
Honestly, this impacts me but doesn’t impact me. I tweet. One person likes it. I read things, I find it to be both interesting and quite a mess. I don’t like threads. I don’t like creating.
I don’t like why I need to separate all these things in the character limit. These are my complaints. Maybe these will be solved, but. Obviously, a big deal that a lot of people are interested in, and there’s lots of communities thriving on Twitter that make this important. And there are some issues that have come up around this.
Obviously, Elon Musk, a successful entrepreneur with like 5 companies worth over a billion dollars power user of Twitter meme Lord on Twitter. $44 billion lots of back and forth on what people thought. The outcome of this was going to be. And in the end that looks like it’s going through as a purchase and so. Interesting, I’m sure stories will come about like how slash why this happened and why it was approved.
After all this pushback and then sort of the the revolt. You know a starts which is, you know, people declaring that they’re leaving Twitter by tweeting on Twitter and and lots of people saying that this sort of drive for free speech. And I think you know there are there are valid points to this. What does this mean when someone says this? So the other last point that I I thought was interesting was Jeff Bezos.
Seems to be disturbed by this, even though there’s obviously this history of him with the Washington Post and and favorable content around him being published. So overall there is an impact culturally that will come from this. I like some of these ideas of open sourcing. No open sourcing algorithms. Why people are removed etcetera etcetera and you know Elon does seem to genuinely want to reinvigorate this company and so will be following this along as more stuff comes out.
I got a couple links here. You know about how Twitter employees are doing it, reacting, how it’s affecting stock, price, all that. All that is still just noise to me right now. But I will be following it like many of us here, so I’ll back out. The other piece that I wanted to talk about today, which is not necessarily related to any news stories, it’s just something that’s continuing to come up for us as as we move forward, and I think it’s just an industry.
Interesting thing, I would say, let’s say, looking at speech to text and NLP, but across any sort of space. If you are entering it and that is this idea of like different types, different types of companies and whether they’re horizontal, whether they’re vertical and or whether they’re or within that sort of nuances. Developer focused companies that then can you know, use APIs to build their own industry specific vertical specific or even horizontal applications of technology? And to me this has been very clear. In speech recognition and NLP space, because when you enter any space you think OK those.
It’s sort of like a little. I’ve used this before, but like it’s like a little puddle. Of of applications you think it’s tight, but then as you get deeper into it, it becomes wider, larger and larger, and that puddle becomes a lake and then ocean. You realize how deep this goes. How different job titles have different needs or problems that they’re facing and how something like transcription and natural language processing.
That seems like a single technology actually needs to be customized and tailored for different problems, different challenges, different business outcomes. And So what I wanted to do is take a look at a couple of companies that are. Tackling this horizontal approach, a couple that are tackling the vertical approach with some examples here to make it more concrete and then some of the ones that are the developer ones and so I’ll do this just in what I picked. I wouldn’t say random companies, companies that are leading this space and just maybe a little bit of commentary on it. As someone who is participating in this space.
Fascinated by speech recognition, transcription, NLP, all the stuff and then the different applications that are part of it. So one of the big ones. Most people who are, you know, interacting with zoom. Know of Otter and Otter was really when I was first building speak. I had a moment when I saw auto launch cause I’ve been building it before when it was like why am I even continuing this?
This seemed like they’ve got a lot of this figured out and really what they built was, you know, really accurate speech recognition program focused specifically on good transcription and then this allowed because of that horizontal approach. It allowed wide wide scale adoption. They built a nice intuitive Android iOS app. Lots of people using it and then from that it opens up. Of this playground, of where the most value creation is and what what they were catalyzed was by by meetings that were really obviously exceeding in, you know, increasingly generating because of COVID and they plugged into zoom and all of a sudden this transcription even whether, even though it was horizontal, had large applications for many people.
I think they’ve been very successful on executing with this and they’ve disrupted. A couple of things. Because of the pricing that they’ve done before this, I would say the leader and many ways are still a leader in the space. Looks like they changed their websites and some of their branding here is was Rev Rev was really good at at combining automated speech to text and then human sort of professional transcription to bring because. With the speech recognition with transcription, what oughter, for example had created it solved many problems, cheap automated speech to text that was accessible for many people it created more problems, which even though it was good accuracy.
It was not good enough and so you still needed to edit yourself, which could take 4 hours for like one hour, depending on how good the audio quality was and and that’s where something like Rev came in. You needed professional transcription transcription to then edit up those automatic transcripts for publishing for research. For machine learning, model training, and that’s where. These companies have succeeded in sort of this horizontal approach and it’s why transcription is becoming something that is so popular here and the the need the technical challenge that’s being solved there is. Hey I have audio file video file.
I need it to be transcribed accurately that transcends many different industries, verticals, use cases and and then allows these companies to grow their customer base and be more applicable to a general audience. And that has a lot of opportunities that it presents. Also creates challenges and I think that’s where we then jump into some of the more vertical approaches of this exact same kind of technology. And so I’ll take a look at one. This is one that my partner uses in her work.
Many people are familiar with the other sort of parallel to this would be gone, and these are speech recognition and then natural language processing applied specifically to sales. And So what? This unlocks this vertical focus unlocks a lot of. Value creation because when you go horizontal, what we’re seeing is that it allows you to get you know 50 to 60. If you’re lucky, maybe 80% of the value creation, but it will not let you get to the full.
Solutions or requirements to fulfill specific purposes and so for our sales person or a sales team. It’s great that you can go to a horizontal application, get something transcribed, and then you can get even with our system. Natural language processing and extract out all the keywords, extract out all the insights, but that’s not necessarily the output that they’re looking for. It’s almost the data is in many ways too raw for them. What you can see here, that chorus is done is specifically tailored their research to or their their system, their. Their analysis and then the output to two sales and so as an example, when you look at this you say I had 100 sales calls.
Here’s the ones that led to a sale, and it’s because it’s hooked into CRM. They have that data out of those 50 out of those hundred calls. Here’s the 50 that converted. Here’s the 50 that didn’t. What’s the difference between those?
And what was the deal size? And if we know all that information, then the insights that are coming out of that can be wrapped specifically around maximizing or optimizing the deal size and the success of those calls. And in many cases, these horizontal applications don’t have the depth. They don’t have the integrations, they don’t have the ability to do this, and so they fall flat, and so a company might say, start with something like Otter Rav, but then as they get more sophisticated, or as they see more needs or they get more advocacy from their sales people, they will then move on to something like chorus or something like Dom that is specifically made for sales teams and giving insights on how you can drive growth in sales with the integrations. Acquired with the inputs required, gone course, for example, looks at emails as well as video calls and audio calls where say Rev say Otter say even speak I in some cases actually do text so I can’t say that but they will look at just audio or just video and then transcribe it only not produce these insights that are really crucial for these these people to drive value and growth.
And then I just want to illustrate this in another another capacity. So kintsugi hello, amazing company specifically focused on voice biomarkers and mental health, and in a way that actually allowed them to transcend some of the even the the the the work that had been done in other spaces. So transcription obviously a big part, not inside extraction another part, but because they were looking at mental health and they were using specific objective biomarkers. They’re being used in mental health diagnosis. They then focused on PHQ 9 different markers like that and then they realized that they could actually do more acoustic analysis or look at only specific keywords or phrases and then create the output.
And you know, generally, their claim. Their piece here is that at a high level of accuracy, they are able to understand mental health and well being through only 30 seconds of voice and audio recording, and that isn’t. Powerful powerful thing because of that focus you can see investment. You can see growth and you can see adoption specifically in mental health in hospitals and universities and this vertical. Focus approach allows them to differentiate from more general purpose horizontal companies like say, Otter, Rev and and allow them to have a path of success and value creation that is unique and different.
And then I’ll just quickly jump. Another great company we love. Same sort of idea. I won’t spend too much time on this, but same sort of transcription and analysis that you can see here. They’ve allowed more manual analysis because they’re specifically focused on UI, product, research, voice of the customer and market, researchers, qualitative researchers, and so not only is the analysis different in the transcription needs, different the interaction with the software front end and the interface is also different.
And so by. Customizing and focusing on the specific needs that allows them to identify their problems within those use cases. Those end users, those ideal customer profiles that they have defined more narrowly than a horizontal company and then deliver on the value that is being created there. So really interesting insight into how you go vertical and even how we’re seeing a couple of vertical applications specifically in speech recognition and NLP. And then I’ll just jump into one last one which is.
This. This I mean, this proliferation of I would say developer focus companies and there are now many many companies today building API first companies that allow developers to tap into it build their own. Sort of technology stack and application there and this opens up. I would say both and again, another way that is horizontal but focused on developers and then as developers TAP into that program. Some companies and maybe better ways give they give the the the organizations the ability to train their own models, customize it to their specific industry and use case, and then grow from there. The capacity, the execution and we’ve seen success with that, but sometimes the actual challenge there.
Remains to be like you cannot. You don’t have the capacity or it will take too long to go to market to do that. So then all of a sudden you’re reliant on. Heavy development time, heavy technical research and all that, and so in many cases these more technical focused companies can actually sort of fall apart in terms of the value. Unless you have the team, the technology, the resources and the ability, and so that’s where some of these then all intersect.
And it also, I believe, creates a little bit of sort of blue ocean where you can actually create value that might not be fulfilled by either one of these sort of companies. Along this paradigm. That’s what we’ve done our best to. Focus here on its bki and we’ve seen that work in some cases, and we’ve seen that not work in other cases, but we really have. Explored and learned as we’ve dived deeper in this to see the sort of makeup of different companies and how they’re being applied.
And again, the sort of the opportunities, the challenges there, and I think that this journey that we’ve gone through and understood there are many parallels to that in other spaces. So there’s the consumer friendly applications. There are the API’s, the very technical ones, and then there seems to be this middle ground sometimes where the consumer isn’t powerful enough. The API’s and and and the adoption of that is too technical, and so there’s a middle where you interact where you get the the consumer. Friendliness and the intuitive interface.
With the deep technology stack and back end of of these bigger companies and then you customize it. You make it work and again that the vertical versus the horizontal application opens up opportunities. It opens up challenges. It’s exciting. It’s exciting to see how all this is being applied and hopefully you know.
As I went through this, it maybe gave you some own applications in your own space and industry. So that’s all for me today. I appreciate you joining this. I missed out on a couple of companies. Cappy scrap and hugging face that I would have loved to talk about more, but that was a lot for 15 minutes and I’m coming up right at the end, so thank you again for joining bye