when Instacart Caper AI, obtained from Lindon Gao, Caper AI, was obtained for $ 350 million in 2021, and it achieved a great victory for the founder in a competitive space led by Amazon Through GO stores free of exit. Caper used sensors, computer vision and other artificial intelligence technologies to detect items in shopping carts for customers so that they can avoid treasure secretary lines.

Six months ago, Gao left his role in InstaCart and Caper to face a new challenge – this time in robots. His new company, Dyna Robotics, came out of Stealth on Tuesday with a seed tour worth 23.5 million dollars, led by CRV and First Round Capital, to build more affordable and easy -to -spread robots for bricks and brick companies. Robots aim to deal with tasks ranging from danger to boring and dirty, including cutting foods, loading dishes, toilets, folding toilets and cleaning toilets.

Gao Dyna Robotics, who said it was estimated at about $ 100 million in the last financing round, with one of his institutions in Caper, Engineer York Yang, as well as Jason MA, Google Deepmind alum. What was the main author UrikaWide reading paper on training robots with human -like ingenuity.

Robots that are one task experts

With Caper, Gao said it helped grocery chains like Shoprite, Kruger Aldi, as well as independent groceries, grows their business. Now, by starting robots, he hopes to do the same with a new set of customers: restaurants, groceries and dry shops.

Most companies in the “AI’s material” space-that is, artificial intelligence of independent systems in the real world such as robots and self-driving cars-works on either AI models for general purposes (such as material intelligence and confirmation) or human robot devices (such as AI and Agility Robotics). However, Dyna Robotics is going a different path, as it builds simple devices in the form of a pair of fixed automatic weapons, supported by the Amnesty International Model trained to carry out one task or a set of tasks. Gao said as much as he realizes that DYNA is the only non -human robotic company that tries to put Robot Ai models that have been seized on specific data data sets.

This narrow focus maintains low costs. Robots from some of the high -value start -up companies cost hundreds of thousands of dollars, even if they are absolutely available. Dyna is expected to cost tens of thousands of dollars when they are for sale. There are no fixed dates yet when robots will appear for the first time, but Qaw said that it is likely to be in the next few months. He said that his robots in experimental production are currently “but do not live completely yet.”

The goal is to automate many tasks that many people do not want to do. He added: “This is a very high value for companies of all kinds,” especially since the robots of many of these types of tasks are not present. For example, traditional machine learning struggles with an unpredictable nature of functions such as folding fabric. But AI models can be trained today to deal with them-especially as Dyna Robotics focuses on collecting comprehensive data for specific tasks, instead of collecting the vast, realistic, realistic and costly world data through a wide range of procedures.

Robots for general purposes will take some time

This is where there is research in Euka. While the tasks he explored in the paper-a robot hand was put to rotate a pen or a robot dog to manipulate yoga-not coherent, Gao said to refresh the same idea: Establishing an Amnesty International Model at the expert level of odors that could enter production very quickly. He said: “I think he is sharing very similar feelings like me regarding robots, which is that reaching robots for general purposes will not happen at the speed that we were hoping.” However, Gao, Yang and MA are still working to achieve a final goal is to develop robots that work at the Public Prosecution for Public Purposes. Dyna robots are mastered one task at one time, which allows Amnesty International models to learn and improve production environments.

Of course, the robot industry has become only more crowded: as of March 2024 More than 1500 Robotics company emerging Globally. For many, small to medium -sized business customers that robots are a better investment than humans may continue to sell a difficult sale.

However, Gao repeated that few companies are currently able to expand their scope of work quickly, as Dyna Robotics plans to do it. In addition, there is a shortage of employment in the types of jobs that Dyna robots treat, such as preparing food, so he said that the convincing customers of the need is not difficult.

He said that the biggest challenge is to make artificial intelligence robot models work reliably and efficiently in a production environment in the real world. “At the present time, the speed of the basic models is about 10-30 % of efficiency at the human level, and we are conducting a lot of research to bring us closer to the speed of the human level,” he said.

Gao said that the company, based in Redwood, California, in the heart of Silicon Valley, has already 30 employees. As a founder for the second time, he said he knows how to build products faster than before. “We have a very basic philosophy that good engineering will still win,” he said.

However, the start of Dyna robots is much more difficult than his Caper experience, as GAO admitted. He said: “The first time you do not have luggage.” “But now I have a kind of expectations and a busy record. I also want to prove to myself that I am not surprising.”

This story was originally shown on Fortune.com

https://fortune.com/img-assets/wp-content/uploads/2025/03/lindon-e1742751407787.jpg?resize=1200,600
2025-03-25 13:00:00