Loading summary
Transcript1 lines
- [00:00]
Alex Johnson
Meta has just made a massive acquisition of an AI startup called Waveform. Today on the podcast, I want to talk a little bit about the state of basically Meta acquiring these audio companies, where I think this is going in the future, what's happening to Waveform, how much they've raised. All of that is a really fascinating story as it feels like Meta is just on an absolute tear to go scorched earth on. Basically all AI researchers that exist on the planet is their target audience. So let's get into all of this. Before we do, I wanted to mention, if you want to try any of the AI models I talk about on the show, including all of the Llama models that I think are super fascinating, I would love for you to go check out my own startup, which is AI Box AI. I'll leave a link in the description. Basically give you access to the top 40 AI models that exist. Everything from Google, Anthropic, cohere, deepseek, llama, OpenAI, quen, all of that for just 20 bucks a month. So. So hopefully that saves you on a bunch of subscriptions you can consolidate. And if not, this is an amazing place to try out a whole bunch of models that you might not be subscribed to or might not have access to right now, including a whole bunch of image generators and audio generators like 11 labs that I personally had an account that was just paying 50 bucks a month for before I started doing this. So if you want to check it out, it's AI Box AI. I will leave a link in the description. All right, let's get into everything that Meta is doing. So Zuckerberg obviously is notorious right now as he's basically offering like, for some AI researchers, billion dollar pay packages at best over like four years. These are very serious amounts of money. Some people are saying, oh, it's just all stock. A ton of this is actually cash, hundreds of millions of dollars. And yet the top one we've heard of is over $1 billion to try to woo some of the top AI researchers and a bunch of them are taking him up on the offer. His super intelligence lab over at Meta is growing very fast and has an incredible roster team that they're putting together. Basically Zuckerberg knows if he gets behind on AI, this is not going to. It's not going to bode well for him and he feels like Llama has slipped a lot. So he's putting together a new superstar all star team. Basically the latest thing that they've been doing is actually acquiring a whole bunch of companies, sometimes it's really expensive to get. You know, you pay a billion dollars to get one of these researchers. You can actually get some pretty impressive, impressive researchers for cheaper prices. It feels like by just buying out their companies and bringing them on as part of the deal and having them work inside of Meta. One such deal, they just acquired Waveforms. This is an undisclosed sum of money that they acquired it for. But I think if you kind of look at some of the reporting that's been happening over on the information on the story, Waveforms has actually just raised $40 million. This is a company that was funded by Andreessen Horowitz, one of the ACVC. Right. Or sorry, yeah, anyways, Andreessen Horowitz, A16Z. Sorry, I forgot their name for a sec. But in any case, a round that they did valued them at $160 million pre money and they raised $40 million. So they've raised some serious money as $160 million valuation. A lot of people are like, what is this company? Who is it? It's only been around for eight months. So eight months later, $160 million valuation and they're getting acquired. How did they do it? Who is this company? And also I will say that this is Meta's second, second big audio acquisition they've made because they previously acquired a startup called Play AI and not that long ago, just back in the, you know, beginning of July, basically this was another one. So it seems like they're acquiring more. I think the reason is obvious. They're trying to build a great team to compete with OpenAI and everybody else that has these really natural language talking AI models. So they're, they're trying to buy a great team around that and great technology. Two of the startup's co founders, one of them was a former meta and OpenAI researcher, Alexis Kano. He was. They also have a former Google advertising strategist, Coralie Lemitri, and all of them have apparently joined Meta and gone over to the company at OpenAI Kano. I think this is kind of one of the big pulls that Meta was trying to grab. He actually co founded GPT4O's advanced voice mode and they're basically their neural networks. So the model that kind of took AI from talk and a little bit robotic and becoming very natural, he helped spearhead that. Then he goes and starts an audio, an AI audio company. And Meta's like, we need this guy, basically. And so they purchased the whole company. It's kind of interesting. TechCrunch was, was doing a bunch of digging on this, they reached out to Waveform and asked basically if their third co founder, chief technologist Cardi K, was going to join Meta as well as basically the outcome for there's 14 other people, according to LinkedIn, that work at the company and so we're kind of waiting to hear back from them. We've seen couple deals that just seem not super great for the whole staff where these companies get acqui hired, the co founders basically go on to the new company. Well, this happened with Google in a recent acquisition and basically everybody else just gets left high and dry and the company dries up. So hopefully there's a good outcome for all of the staff members, all of the employees of this company. But it is interesting. If you go over to Waveforms AI website right now, all you get now is 404. The page you're looking for doesn't exist or has been moved. So it looks like they have completely taken down their website at this point. But if you go over to LinkedIn, you can see a little bit more about them. Basically their LinkedIn page says that their mission is to solve the quote Speech Turing Test, which tries to measure if a listener can tell between humans or AI generated speech. Obviously this is a big challenge as you for basically every reason you don't want, like if you're trying to make a tool, you don't want people to be able to tell it's an AI. So when it's talking to you. OpenAI has done a phenomenal job on their natural speech that they've added into ChatGPT and we've seen a couple other players, Google doing some, doing some good work and 11 labs doing some good work there. But this would be huge if they really are able to, I guess, trick a user or make them unable to determine if it's truly AI or not. I think this is basically what the problem they're trying to solve in Meta needs. This for a couple different reasons. One of the first things that I think is going to be huge in Meta is the Meta Ray Ban sunglasses. Okay, well, obviously just the Meta AI app, which is a direct competitor to ChatGPT and every other AI company, it has voice mode and so they want this to be natural and as good as it can get because right now it's not incredible. So I think that just makes perfect sense. In addition, there's a bunch of like real use cases that people are actively using. And I think of the Meta Ray Bans as one where you basically look through the Ray Bans and you can ask Meta AI about stuff that you're looking at. You know, I'm looking at a bike, how do I fix the chain, what's broken on this, you know, this engine that I'm looking at. And Meta will talk to you in your ear. Now if it sounds like a robot talking to you, it's the same problem as Siri and Amazon Alexa before they kind of fixed it where it just sounds so robotic and fake. You just, you don't really want to use it. And so this I think is obviously one place where Meta could make their product a lot better with this technology. And you know, basically everything they're doing, if they have good voice, you kind of need that as a feature today. Another thing that Waveform was doing is they were developing something called quote, emotional general intelligence, which they were basically saying focused on understanding individual self awareness and management. So they're doing some really interesting research beyond just the technology. And I think that, you know, they have a great team, people that have worked on some of the really complicated problems in AI and audio today. So I think this is a no brainer for Meta to go and acquire. I'm not surprised by this. I'm really curious of basically what's going on with the actual deal and what was kind of pounded out as they did that deal. But overall, really excited for the whole team and for the founders. Great work to everybody involved. If it sold for anything like the recent valuation of 160 million, which I don't see why it wouldn't, probably more than that, then yeah, seems like a great outcome for everyone involved. Thank you so much for tuning into the podcast today. If you learned anything new, I would appreciate it if you could leave a rating review wherever you get your podcasts or just if you appreciate me making these podcasts, I would love a rating interview. You could go hit the about tab on Spotify and that's where you can leave a rating. And on Apple it is, you know, feel free to drop a comment, hit some stars and on YouTube. So subscribe. I will catch you all in the next episode. Make sure to go check out AI box AI if you want to try all of the AI models in one place, I will see you in the next episode.