Loading summary
Transcript1 lines
- [00:00]
A
Today on the podcast, I want to tell you about SmartCat. This is a company that has just received $43 million in funding for they have an AI powered translation platform. There's a bunch of really interesting things going on here. First off, I want to talk about some of the features, some of the things that they're actually doing, but also what different companies are saying about it. I think earlier this year, or maybe it was last year, G time flies. I believe it was, I believe it was Gizmodo that shut down like all of, I think their Spanish section of their site and they're essentially are just translating all the English content using AI tools. Now if you've used anything like Google Translate in the past, you know, it's not always good. I, I'm currently in living in France for a couple months and I message people in French all the time. I like, I speak French but like I'm always worried my grammar is bad. I'm always kind of like self conscious about it. So I put like my messages, even if I write them in French first through Google Translate, just to make sure that like all my, all my grammar is good and all the words I use are good or sometimes it's faster. I say it in English, it translates it in French to me. But what I do notice is like, I'll look at the translation and while it is like technically correct, I'm like, nobody says this phrase this way in France. And so like I change sometimes I'll change like the English input to be like it sounds like funny in English but I know like it translates correctly into French. In any case, this is not a problem. I see as much with AI generated translations because they're understanding more of like what's actually being said. So I believe that AI translations are actually better than like Google Translate or something like that. And apparently $43 million just got put into a company. So it seems like this is a thesis that is agreed upon. So I wanna tell you a little bit about SmartCap, but first I wanna tell you that I think the number one investment I ever made, whether I was like financial or time wise into myself and my businesses, was in starting a podcast. Obviously I'm biased, I'm telling you this from a podcast right now, but this is something that has taken my career to a completely different place. I've been able to network with, you know, founders of billion dollar companies that have raised hundreds of millions of dollars, had them on the podcast, talked about what they're doing. I've gotten amazing insights added them to my network. And all this is thanks to this podcast. So if you have ever thought about starting your own podcast, with which I highly recommend you do, I think it's one of the best ways you can elevate your personal brand. You can elevate your network and your business, get actual customers. I have a podcast course that I've just launched and there is a link in the description to that course. Um, it is probably one of the most useful. No, not probably. It definitely is the most useful podcast course you'll ever take. It covers how I went from 0 to 4 million downloads on my podcasts, all of the steps I did, everything about production, getting guests, growing the podcast organically, no ads for free, just organic growth. I was able to do that and everything in between. So go check out if you're interested in start a podcast link in the description. Let's talk about what is going on with SmartCap. So the first thing I want to talk about is some of the, I guess, skeptics that are kind of in this, in this space. If you go and there's recently an article on TechCrunch about SmartCap, and I just want to read you this quote from it because I thought it was so funny. It said, can AI ever fully replace translators? Not likely. AI translators translations tend to lack the lexical richness of their human translated counterparts, mainly because AI models make choices based on probability, not lived experience. Okay, so I get where they're coming from. Like, I don't think AI could probably replace my podcast because when I talk to you about like, podcasting or when I talk to you about, like, AI companies is like, I'm actively working on AI startups so I can bring in like, war stories of things I'm like, actually doing or seeing or consulting people on. And so, yes, I think lived experiences are not there, but I think that TechCrunch is completely wrong on this point that, like, AI translations are not as good as a real human, because a translation is taking what I'm actually saying. I'm already giving it the lived experience and all of the value, and it's just changing it into another language. So I actually think that AI translations are just as good as human translators, maybe better, because maybe there could be less bias. You get a human translator that like, does or doesn't know certain words and they like, sort of translate it differently than what you would have wanted. AI should know everything. It should know every, you know, how everything is being perceived and whatever. It's got way more knowledge. So I Actually think AI translators would do a better job in that case. So is AI better at everything? No, but in this case I do think it's better. So SmartCat, they were founded just back in 2016. So this is not a very old company, but it provides essentially does AI translation for enterprise. So this isn't really like direct to consumer like Google Translate. This is more going to big enterprises. This is how a lot of this software has to go today. Right now they said that they have over a thousand corporate customers and they, I think it's about like 20% of the Fortune 500 is currently using them. So overall, obviously impressive. They have local and international governments that use them and a bunch of large global companies as well. So their CEO is Ivan Smolnovkov, who is a co founder CEO. He was a psychiatrist at the Russian Academy of Sciences. Science, sciences. And he kind of specialized in fiber optic materials. So later he founded Abby ls, which was a language service company. This is essentially where this, his company here, SmartCat was incubated, debated. He left that back in 2016 and then he just focused exclusively on SmartCat. I think he, he probably saw that they were onto something and so he's diving into that. What's their business model? Right? Like how does this actually operate? Essentially they're trying to quote, unquote, modernize the traditional translation agency model. This makes sense, you know, before AI tools even like, and I believe, you know, Google Translate was even worse. I used to use, I used to do like a lot of SEO work for travel companies. That was kind of like my, I don't know, I. My background's in marketing, so it's kind of like my early, early workdays and Google Translate was not good enough. So we would, yeah, we would hire agencies, we'd hire people that spoke other languages to do the translations for us. So all in all, it was really expensive and it took a long time and they charged you per word. And everything about AI is cheaper and faster today. So I think they probably understood that right now they have as far as like their different services that they offer. With SmartCat, they have a bunch of different tools and apps that essentially are helping you to write. They have like a bunch of different areas that they're helping you write, which is like elearning courses, websites, files, software. And they do about 270 languages, which is honestly really, really impressive. That's huge. I mean, but to be fair, like ChatGPT does like all of this, right? So I'm curious to see like what the main Differences they have other than just me, I don't know doing all of this with Chat GPT. Now I think some of the bigger things that they can do is there's like actual software integrations that are automatic. And I think that there is value in that definitely. Right. It's not just a, it's not just a wrapper on top of ChatGPT if they have a tool that integrates with all of your files, for example, kind of like what Dropbox is doing where they can help you access or ask questions about your files that are in your Dropbox. I think there's a lot of value in that because it's kind of built into the process as you already do. You don't have to copy and paste it into ChatGPT. It's kind of like automatically there so that I see a lot of value in. So they have not trained their own models. They use a, what they call a matching engine which is essentially they're going to use third party models which they say are suited to the content and language. Now what this really means is like pretty much I'm developing something very similar with AI box where like when you, when you're trying to like generate a prompt for something we, we use, we use some of the, for example, we would use something like GPT4 or a different even cheaper, faster model to like look at the prompt originally determine what the person is trying to do if they're trying to generate an image or a video or whatever and then we send it to like a recommended model. So we have like a smart mode essentially. So if you don't want to have to pick sort through the models and pick one, we, we have a bunch of parameters, you can just give it your prompt and we can decide what type of content you're trying to produce and yeah, essentially spit it out that way. So they're, they're doing this, they're calling it a matching engine and maybe they have some, some home baked stuff for that. Um, to be honest, you, you can use Chat GPT to be that matching engine and maybe there, it's probably better if it's like custom made. But all in all that's, that's what that means. So I don't know, I, whenever I see like things that seem a little too buzzy or fluffy, I just try to explain exactly what it means and that's exactly how you could do it. So in any case they got their quote unquote matching engine, they're going to send it then to a third party model. So that means that there's no special model. They're just like, oh, okay. Anthropic's Claude is the best at Spanish, which I don't know if it is, but let's just say it was. Or more likely, because I think Anthropic and Chat GPT can do a lot of like the Latin languages very well. More likely, um, they'd be using it for like, they'd be using something like Quen for like Chinese from Andron. And you know, I doubt ChatGPT is going to be as good as Quinn, which was trained on a Chinese data set to be doing Mandarin. So stuff like that. I think there's like certain AI models that might be a little bit better and they're essentially just determining that. In any case, they say that they fine tune translation models when necessary, especially for companies with large data sets of really commonly used phrases. So that's interesting. They're adding some like extra stuff there. They have translation support. So obviously AI makes mistakes. They kind of offer access to a whole bunch of translators and copy editors to review it, which of course that's an extra fee, but I do think that's an interesting. And like, that's a useful add on, right? Like if you're a big company, you got a big budget, you get everything translated very quickly and cheap with AI, but you're like, oh yeah, we just want to make sure it's good. You send this over to actual human translators to go verify, look through and just see if there's anything that could be tweaked or changed. I think that there's a lot of, a lot of value in that, especially if you have a bigger budget for people, smaller budgets, you know, they just forget about it and, and just hope that it's good. But essentially AI or, you know, clients, they can choose from AI translation, human translation, or a combination of kind of using those two together. So they kind of have the both, which, to be honest, I think that, I think that AI can do everything just as well or better, to be honest. But there's some big companies that might not believe it. And so it looks like they don't want to cut off all those old companies and they want to keep people on to use both. So anyways, very, very interesting. I think this is a really, a really exciting round of funding. This is, this $43 million. This is the Series C. It was led by Left Lane Capital. So Total Smart cat has raised $70 million. They're gonna be using this money to essentially expand. They have about 200 people on their team right now in product development, marketing, sales, and whatnot. So they're gonna be growing all of that. The global market for machine translation solutions, apparently it's at about $978 million as of 2022, so supposed to grow about 13% every year till 2030. So there should be a lot of room to grow in this. I'm excited to see what this company does. I think it's overall a fairly good niche and industry, and I'll definitely be following it closely.