Rabbit co-founder and CEO Jesse Lyu isn’t afraid of death… the death of the company, at least. He told TechCrunch that the company is a startup whose fortunes can be swayed by the whims of multibillion-dollar rivals — but that’s no reason to give up and go home.
Appearing on stage at StrictlyVC LA, Lyu explained his rather philosophical approach to the threat of Google, Microsoft or Apple crushing them. (Quotes have been slightly edited for clarity.)
Rabbit’s r1, the pocket AI assistant that attracted a lot of publicity after its debut at CES, is certainly an original proposition. Half the size of a phone, the device functions strictly as a voice-enabled assistant, but is capable of remotely controlling your apps and performing complex actions, as well as answering questions and continuing a chat like ChatGPT. He described the two parts as “intention” and “action.”
“I had this vision many years ago, actually 10 years ago, but the technology wasn’t ready. This is the first time in history that a device like this is actually possible,” Lyu said.
He explained that he had been intrigued by LLMs’ abilities to understand language and intent, and that with the apparent flexibility of transformer-based systems, it was only natural to try to get them to perform actions as well.
“We immediately tried to use hyper-prompts to make this language model do things, and the result was pretty miserable,” he recalls. “There is a demo from another company to use an LLM to go to MrBeast’s latest YouTube video and leave a comment. Yes, in theory, language models can do this. But you would have to literally watch your screen doing this step by step. And it takes about two to three minutes to complete such a task. We just don’t think this can translate into a good end-user experience.”
Their solution is the “big action model,” which is trained on hours and hours of real users interacting with popular apps: “Spotify, Uber, Expedia, DoorDash, you name it. We have the top 800 apps with the highest frequency. Then we built this the neural symbolic network and we ask this AI, which we now call a large action model, to control these clips, but frame by frame. The idea is that symbolically, the AI will eventually be smart enough to extract all the buttons , all the elements, and then we can basically create a logic for automation.”
The r1 rabbit in use. Hand model: Chris Velazco of the Washington Post. Image Credits: Devin Coldewey / TechCrunch
The language part is still running on third-party LLM services like Perplexity, which seems to be making an effort to capitalize on Rabbit’s success by offering a year of free service on top of whatever r1 provides. I suggested that API costs and other considerations could pose a risk to the startup’s solvency.
“First of all, we’re not losing money selling the r1, which is a very, very, very important achievement, especially for the new startup in generation 1. We’re not going to go broke selling more units. I give all the credit to my equipment team of amazing guys who were able to basically negotiate the parts and BOM [bill of material] cost,” he said. “We’re really close to 100,000 orders. Two days before Keynote II he told my team it would be great to sell 500 units on the first day. But we sold 18,000.”
As for a subscription, Lyu just doesn’t see it working, especially when the device’s thesis is cheap and simple. Although he mentioned that users will be able to train and sell their own models for specific applications later, Rabbit would get that, but he cautioned that this is a long-term plan with no details yet.
Finally, when faced with the fact that the world’s biggest, richest companies are spending billions to advance AI, Lyu gave an almost Zen perspective on the prospect of crushing under the heel of Google, Microsoft or Apple (the whose CEO Tim Cook just said it will “break new ground” in artificial intelligence this year).
Image Credits: TechCrunch
“I have no illusions, to believe that we are not a startup. We are a startup,” he said. “I mean, the first lesson I ever learned from Y Combinator two years ago is that 99% of startups are going to die. If your mindset as an entrepreneur is, “Oh, I have a genius idea and I can guarantee this will work, no matter what all these big tech companies are trying…” I mean, you’re delusional. There is no such thing. The reality is that a startup is a game of survival and you’re better off spending your time focusing on your own stuff.”
“They’re going to do what they’re going to do, and I’m going to do what I’m going to do, right? There must be some founders who, when they heard that Apple was making Apple Cars, stopped, right? They just canceled. And now what? I think it’s good to have that level of competition that will help us grow faster or die faster, which is the nature of startups. It’s either – I don’t know yet. But I’m trying my best – like I said, it’s a game of survival.”
You can watch the entire panel below.
