Unveiling Fusion Intelligence

0:00Miriam Marati has just raised$2 billion for her AI startup thinking machine labs, or as everyone on Twitter is calling it, Thinky. We'll see if that name actually sticks. But this is really impressive. They've revealed a little bit about what they're actually going to be doing with the money, who they got the money from. We're going to break all of that down in the podcast today. This is a huge deal. This is one of the top funded AI startups at this point with this new$2 billion. And it has a$12 billion valuation before ever launching a product. So this is really quite impressive. She went on Twitter recently and was tweeting about this.

0:38We're going to dive into all of that. Before we do, I wanted to mention, if you ever wanted to try out the latest AI models that I talk about on the podcast here, there's new ones coming out every single week and sometimes every day. I'd love for you to go check out AIbox.ai. That's my own startup. We just launched our beta playground platform where you can go try the top 40 models. That's text, image, audio, all in the same chat. So you can just toggle between any model, switch between them. You can also ask one question and you can get it to regenerate the response with tons of different AI models.

1:07So if you want to test out if Gemini or Claude or ChatGPT answers your particular question the best, you can do all of that. We have benchmark data and all sorts of interesting stuff. It's 20 bucks a month and you get access to all of them. So you don't have to have subscriptions to every platform, including 11 labs for audio, tons of cool image generators you may not have tried yet. It's all on there. So if you want to check it out, AIbox.ai, there is a link in the description. All right, let's get into what Miriam Moratti is doing. So what's interesting here, Miriam Moratti, of course, is famous for coming from OpenAI.

1:39She was one of the founding team members. Now, those founding team members have this definitely kind of this golden aura about them where they're able to go spin out of OpenAI. We've had a lot of them leave OpenAI and go spin up their own companies and raise huge amounts of money. And Miriam Moratti, she is, you know, a very serious player. She was the interim CEO for, I think, one or two days at OpenAI. So she's got a great track record. And she was actually in charge. I believe she was the CTO at one point. She was in charge of a lot of the most impressive products that came out of OpenAI, including, I believe, their image generator.

2:11So what's interesting here, she's spinning out this whole new thing, and it has been very secretive. If you go to their website, there is very little information that we've essentially been able to find on there. But she took to Twitter and gave us a little bit of an insight with this new round of funding. This is what she said. She said, quote, Thinking Machine Lab exists to empower humanity through advancing collaborative general intelligence. We're building multimodal AI that works with how you naturally interact with the world through conversation, through sight, through the messy way we collaborate.

2:43We're excited that in the next couple of months, we'll be able to share our first product. OK, already, I just got to stop there for a second. At first, the description of like advancing collaborative general intelligence, like no one knows what the heck that means. But I think that like for their investors and everyone else, it's like the it's that signal. And that's probably what they're raising the money for. They said multimodal AI models, right? So we're probably talking like it sounds like definitely they have an audio model. Definitely. They have a text generation model. The two billion dollars they've raised, they're definitely gonna be able to train some impressive models.

3:13And we know they have a partnership with Google that they made Google Cloud in the past to do a lot of this. So they have the partners in place. They have the money. They're going to be training their own models. And I'd be curious to see if this also gets into video and image generation model just based off of her background. But regardless, I think the keyword here, the reason that she's making money in the way she's differentiating, differentiating herself is the word collaborative. Right. She literally puts this in the title, advancing collaborative general intelligence. What the heck is collaborative?

3:42And then later on, she said that she says through the messy way we collaborate. So I think this is going to be a tool that's probably for teams, probably for enterprise or maybe for regular users, too. But it's going to have a collaborative feature. So you're using it. Someone else can use it. You share a link and you're all messaging inside of the same thread, which is honestly a very cool idea. Then the cool thing is they're actually getting ready to release a product. Now, this is usually tied to these fundraising rounds. They just raised$2 billion. I don't think they get to raise that$2 billion after I believe they already raised a billion.

4:15And I don't think they'd be able to like, hey, we raised a billion. Now we want to raise another$2 billion for something cool, which we don't really know what it is. They've obviously developed their product. And they said that they're actually going to be releasing that soon, which is probably why they got the funding. They probably showed a demo to the investors. So they said, we're excited that in the next couple of months, we'll be able to share our first product, which will include a significant open source component and be useful for researchers and startups developing custom models. So this is cool.

4:43It seems like they're taking a unique angle. It's going to be open source, whatever their tool is. And it's going to be for startups that are building their own custom models to be able to use their tools to better do this. She said soon we'll also share our best science to help the research community better understand frontier AI systems. So it sounds like they have some breakthroughs and discoveries they'll also be able to share, which is, you know, always exciting. In addition to the open source tools that they can give, the research behind how they actually made those tools are sometimes even more important.

5:12You know, we've seen companies like OpenAI or others like release a model and they'll release the weights, but no one really knows how they got that model. And so it's a little bit murky. And so it's cool when we actually get, you know, the science behind it. She talked a little bit about who they raise money from, which was A16Z, who actually led this. So, you know, the number one venture capital company in America is leading this. They had participation from NVIDIA, who always just puts in money for any new AI startup that looks serious because they know they're going to have to buy NVIDIA GPUs.

5:45So it's kind of like, you know, buying their own book or whatever. But I mean, great play by NVIDIA. No, no hate there. Acel, another top VC, ServiceNow, Cisco, AMD, Jane Street, and a bunch of others. She said, we're always looking for extraordinary talent. And then she pretty much has like a call out to, you know, if anyone wants to go work at her company. It's kind of interesting when you see these like big announcement. They're like, here's a big announcement. Here's how much we raise. Here's kind of what we're doing. If anyone wants to work with us, let us know. Right. It's kind of like always a recruiting thing, too.

6:16So this is interesting. The responses on X have been kind of funny. Dylan Patel said, please honor the nickname everyone uses. Thank you. someone has a picture of or someone said due to inflation one million dollar seed rounds are now one billion dollars and someone said when API key someone said looking forward to think he's first product all right well it sounds like it's thinky that's the name of the company going forward but overall really exciting and they have you know obviously the top VCs are putting billions of dollars into it I don't think this is vapor where I think they have a serious product that they're they're getting ready to release, which is really exciting.

6:56And of course, with her background, we know that this is going to be something that is pretty exciting. But this is kind of crazy, because we're seeing these this this larger and larger seed rounds. This is, I think, one of the biggest seed rounds that we've ever seen in Silicon Valley. A billion or$2 billion for their seed round. We had just in June, a couple different news outlets were saying that they were getting close to raising$2 billion with a$10 billion valuation. But now it's up to$12 billion. So the valuation evidently has gone up, which I think is really interesting. We're not 100 % sure if it's going to be an open source model that they're releasing, they kind of alluded to this sort of open tool.

7:39And so like, we know that it's cool when these models are completely open source. But we also know that some people are sort of sneaky, like a la meta, when they released their open model, but it wasn't open source, like a lot of people just say open model where basically they release a model that anyone can download and use. So it's like open sort of, but they don't always like share all of the weights or they don't always, you know, give people license. They're like, yeah, it's like open, but if you want to use it for anything commercially, you can't. And so then it's like, well, other than running a model on your computer by yourself, it's not super useful to actually build anything with.

8:13So we don't know what exactly that's going to look like as far as if it's going to be completely open source or not. but um since miria started this whole company she has a really incredible team that she's been able to pull together a bunch of former open ai colleagues that she had i have joined her we have john shulman brett zoff luke metz um and so we know right now she's really trying to staff up and get a bunch of people she specifically says she's looking for people that have quote that are quote building successful ai driven products from the ground up that's what that's what their website says.

8:49And we also know that this isn't the only way that people have that they've, you know, that they've tried to get money. They actually had an offer allegedly from Meta to acquire the entire company. That never, according to insiders, that never actually went to an offer that Meta made. But we know right at this point, Meta is offering these$100 million deals to a whole bunch of researchers. She has a ton of the top researchers at her company. So I think Meta would love to acquire them. Probably the tech they have would be super cool. Meta is trying to build out their kind of their general super intelligence team.

9:19So it's going to be interesting to see to see what she's making. Some people are saying based off of the team they have and the amount of money that she's actually been able to raise. This is, you know, it's a quote unquote, legitimate threat to leading AI models. That's what TechCrunch says, like open AI anthropic deep mind. I don't know. I mean, I would love for there to be more competition in the field. I'm not going to, I'm not going to say before seeing any sort of product or any sort of deliverable, what what the competition rate is, I think everyone has been shocked by a company like, for example, XAI that was able to come up with Grok and built spin up a bunch of, you know, data centers and AI tools very quickly.

9:57And you know, their bet their latest benchmarks had it as the number one AI model. So and then they launched like within a year. So I think it's possible to do to leapfrog the entire industry in like one specific domain, right? Grok did it with just kind of text and PhD research responses. It's not the best doesn't have anything from video or very the best audio or anything really great for image. So I think we're we're going to see more and more in the in the future, what this company is able to do. And if Thinking Lab, Thinking Machine Labs is able to come up with something impressive after being out for just a year, I think it'd be cool to see other people make the same play.

10:33I think we'll see more VCs back these companies. Now, of course, it's not like Miriam or Addy came out of the blue. She's obviously one of the top brass at OpenAI, the co founder. So that's why she's able to get I I think the resources team and build something this impressive. So it'd be interesting to see if anyone can replicate that. But in any case, this is absolutely, it's going to be an incredible uphill battle to watch as she brings her company to the forefront and competes potentially head on with a lot of the other big players. Her former colleague, Sam Altman. We're excited to see what happens.

11:06I'll keep you up to date on all of that. And whenever her model launches, if it has, you know, the ability to have API access, we will be posting this onto AI Box, my own platform. So if you want to check out the platform and get all the models we currently have and all the future ones we'll be adding as they're released, go check out AI Box.ai. As always, there's a link in the description. Make sure to leave a rating and review for the podcast if you learned anything new and if you enjoyed the episode today. Thanks so much for tuning in and I will catch you next time.