← The Last AI
The Last AI
Examining Unexpected Acquisition Logic in Cursor Acquires Koala: Saving Employees
00:12:37
Loading summary
Transcript1 lines
- [00:00]
A
Cursor, the AI coding tool, has acquired a new enterprise startup called Koala. And they're doing this basically to try to compete better with GitHub Copilot. But I think that there is so much more to this story. I think the entire industry is going through a really interesting moment right now with these kind of aqua hires and half hires and buying the top talent from companies without buying the company. There's, it's, it's kind of crazy. We're seeing it in a big way with another AI coding company, Windsurf, so that basically the leadership team got Acqui hired by Google, leaving the entire rest of the company to kind of fend for themselves. They were all kind of depressed. There wasn't the big exit they hoped for, but they were actually then purchased by I believe Cognizant, which is the creator of Devin, the AI coding tool. Anyways, all of these AI coding tools, I'm going to get into why so many of them exists. But specifically I want to talk about what just happened with Koala here because, because this is more interesting than I think any of those deals and it's got more, more drama maybe, but really, really juicy. Very, very interesting. Based off of what's going on at Quality. Before we get into that, if you want to try any of the top AI models, I would love for you to try out AI Box AI. This is my own startup. We currently are in beta and we have a platform that allows you to test out the top 40 AI models with all for all for one subscription price. So for 20 bucks a month you get access to all of the anthropic Google Cohere, Deep seek, meta, Microsoft, Nvidia, OpenAI, Quen Xai. You get all a bunch of really interesting image models and audio models and you can chat with all of them in the same chat thread. So you don't have to have you know, 10 tabs open if you want to try that. Use these different models all in the same thread. You can talk with them. You can just switch your model while mid conversation. You can switch to a new model that might be better at doing certain tasks and you can also compare results from different models so, so you can ask a question and get it to get and you can get, you know, four different models to respond to it to see is Google better? Is OpenAI better? Is anthropic better? Is deep seek better answering this particular type of question. And you also can then, you know, based off of that same conversation, talk to an image model or talk to an audio model and create audio files all within the same platform for one price. So if you want to try it out, I'll leave a link in the description. It's AI box AI. It's currently in beta. We're getting some great reviews, but also some really good feedback and some cool features that we're adding based off of that. So if you see anything that you think, man, I really wish there's an AI company that did X, Y and Z thing, let us know. We're actively building, adding a whole bunch of cool new features. So yeah, we'd be, we'd be thrilled to get your feedback. All right, let's go into what is happening with Koala. So first let's talk about the fate of Koala. Koala recently announced in a blog post that they were going out of business. They have a blog post called so Long and Thanks for all the Eucalyptus. And basically they said that effective September this year, they're going to be winding down their entire company. They said that they are going to be processing refunds for any unused subscriptions for their customers. Okay, so this is kind of interesting. We understand what's happening to their customers. They're winding down the company and they're refunding, you know, customers that maybe paid for the annual, you know, the annual package or whatever. But like, what is Koala? Where did it come from? What is it doing? This is what I found to be super, super interesting. Basically. Basically, Koala raised their Series A. They just raised a Series A five months ago and. Five months and, and by the way, by Series A, like, is this a small company, is a big company, they raised $15 million. So to me, $50 million is pretty good. I feel like when you raise a Series A, basic startup wisdom is that you want your, like you, I guess we're in a hyper growth phase right now, but you like technically want to do a raise and make sure that it's good for 12 to, to 12 months if you're risky, 24 months if you're not. Like, basically that's what you want it to be good for. And if you need more money in the, you know, as you're starting to run out, go do another raise or try to get revenue up, whatever, right? Like that's basically what what you want. They spent the $15 million in five months and or are closing down or announced that they're closing down in the five months. So to me, that's kind of crazy because I'm like, did they just do this raise as like a cash grab? Like, did they know they're going to go bankrupt in, in five months? That's not a very long time. That's not even six months after they did the race. So, I mean, yeah, I don't know. To me, I will throw shade at that. I think that's kind of very not cool to their investors. They have HubSpot Ventures, recall capital of four who are all in this round. It was led by crv. Yeah, I don't know, I think that's not cool. But in any case, regardless of that, they are getting bought out. Right? That should be exciting. So they were four years old, they had about 30 employees. Um, if you go look at their LinkedIn, you could see that. Um, and they had a bunch of big clients, right? They had Vercel, Stat, Sig, Retool, that were all clients of theirs. Okay, so no one from, by the way, no one from Koala or Cursor is like commenting on this story. But you could just go figure out a lot of this data publicly available and you can understand. And then some sources from inside the company are essentially leaking this because probably, you know, I feel like we get better leaks when this story is a little bit more heinous. Like in my opinion, five months after raising 50, $15 million, you go, go completely out of business. I'm like, ugh, someone probably doesn't feel good. They're just going to leak, leak. What's happening, basically spill all the tea inside the company. Okay, so that's what was happening with Koala, what's happening with Cursor. So Cursor, obviously big AI tool, it's growing really fast. It's competing with Microsoft Anthropic. Anthropic now has Claude Code, who personally we. That's what we use over at AI Box to develop everything. And it's like insane. I know Curse is probably a great competitor, not the one that we, we are typically using. And by the way, Koala wasn't like a random company, I guess it's, you know, had a bunch of customers that I mentioned, but their co founders came from Meta. They had advisors like Jack Altman on there. So like it should have been a legit company. It wasn't like necessarily an AI company, quote unquote. It's, you know, basically from what it was doing. But Cursor is going and hiring them. And basically what they're doing is they're not taking the whole company. They're just grabbing their top talent pretty much from the company. So basically the deal is that Cursor is going to bring on a bunch of quality top engineers to build out a dedicated enterprise readiness team. But the rest of the Koala team are not going to be joining any sphere and they don't plan on integrating any of the startups core CRM products. So it's like a CRM, right? So basically they're like, okay, we'll buy you guys which going out of business anyway. So basically we're acqui hiring your top engineers and getting them like to work on our stuff as soon as possible and the rest of the company can just go bankrupt. We're not using anything that it's doing and everyone's getting let go basically. So it's kind of like a weird deal. I mean basically it's like I think there was some talent in the company and this is sort of their like golden. I know it's like their parachute out. Although I don't know why they wouldn't have just gotten jobs there. Maybe they wanted to make it feel like an acquisition. Maybe as part of the acquisition some money was getting repaid to investors so investors were pushing like hey, don't go get new jobs, please let us get an acquisition out of this thing so that we could get some sort of payback. Maybe this. The CEO felt bad about his five months of raise to bankruptcy. But in any case this is basically what's happening, which is really interesting. So the reason I wanted to talk about this on that podcast today is because I think this is a trend we are seeing in so many different AI companies. We are not seeing these clean cut acquisitions that we used to see in the past. We're seeing these weird acqui hires. Maybe they take the company. And this by the way isn't the first time that Cursor has purchased a company like this and sort of just like leashed the talent out. Nor basically all of these top AI, all these top companies. Google has done this with a number of companies. Everyone, everyone seems to be doing this. They just like kind of buy the company, take the top talent, let the company die. Maybe the company was going to, was kind of doing that to itself. So this is one thing that I did think was pretty interesting from this. Basically Cursor also did this to a company called Resourcely. And I think they're pretty much hoping that Qual and Resourcely's talent are going to help them make their company much better. They have something which is like an IDE which or basically copilot works as an AI powered extension to existing IDEs or integrated development environments. So you could you think about like VS Code, Jetbrains, all of that, they plug into that and they're hoping that by acquiring companies like Koala and Resourcely, this is another one they also did this to. They're basically going to boost their own product, but they don't want to take on any of the companies. Now. Cursor is expanding very fast. This is a huge company. They recently announced that they reached $500 million in annual reincurring revenue. This is just back in June. So very recently they have, I think they say that they now work with more than half of Fortune 500 companies. Nvidia, Uber, Adobe are all kind of are all using Cursor. I mean, it is a great product. It's one of the ones that I hear talked about the most in this kind of field. But it's also competing directly with Microsoft, with GitHub, Copilot, directly with Google. And so it's. It definitely has a lot, some pretty stiff competition here and of course directly with Anthropic, who basically it's a very complicated situation because they rely on Anthropic. It's the number one model used for coders. So they like Cursor needs Anthropic, but also Anthropic has cloud code, which is a direct competitor to Cursor. So it's kind of this interesting position where a lot of people have said, oh, Claude, like Anthropic just made Claude code because basically their biggest vendor, their biggest user was Cursor. And they're like, oh, we could do, we could do that. Like they, you know, and so this is kind of interesting. But in any case, all of these are definitely different types of tools. Employers tend to view the product similarly. All the AI tools are improving productivity of the software engineers, but they might have different, like, pros and cons. When you're looking at the tools from Anthropic, Google, Cursor, Cognition, all of them are kind of slightly different, but I don't know, they all are helping developers. So why. This brings up the final question of the podcast, which is why are there so many of these tools for AI developers? Like, like, why is this the number one thing that I see news articles on? I think really it's because this is the biggest area that has reached product market fit right now. Developers are seeing so much value from these tools. And our company included over at AI Box, like Claude Code has completely changed the scale at which we can build out new features and products. And that's why on the pod, like normally when I build, you know, when I work on software, things I'm normally not telling people, like, hey, go try my product. If you think of any cool new features, like, let me know. Because we just get. I get so inundated with requests. But at this point with cloud code, we actually can execute on a huge portion of feature requests that we get. We're just so much faster basically with. With what we're able to pump out there. And so a lot of developers are seeing this exact same thing. If you go on Twitter, you can see people, you know, we were spending thousands of dollars with Anthropics credits with Claude Code before we realized that there was something called cloud Code max, which is 200amonth per user, and it lets you basically get unlimited. They sort of have some rate limits, but in any case, really phenomenal value that developers are getting. And because of this, there is a ton of competition. You could spend thousands of dollars. And I think a lot of companies want every single one of their developers to be using one of these tools because you get such a multiplier on what they're actually able to produce. So overall, this is a very interesting part of the industry. I think that there is going to be a lot of growth here. I'll be curious to see if one of these companies in particular continues to run away with the market the way Claude Code has, or if Cursor is able to kind of catch up. If GitHub Copilot is able to cement itself. It feels like cloud code is in the lead, But Cursor and GitHub Copilot are not far behind. So, yeah, I'll definitely keep you up to date on how this, this market plays out because it's going to be a very fierce battle. Like we're talking the biggest companies, Google, Microsoft, competing directly with these startups. And so, yeah, it's definitely going to be some stiff competition. Hey, if you enjoyed the podcast today, if you enjoyed the episode, and if you learned anything new, I would love for you to leave a review of the podcast. It helps the show basically get shown to new people, so it helps out a ton. And also make sure to go check out AI box. AI if you haven't already. We're in beta and we are currently letting people try the top five 40 AI models for just $20 a month, so you don't have to have subscriptions to every single AI platform out there. Thanks so much for tuning in and I will catch you in the next episode.