Several robotics companies are working to build foundational models that can perform one of the most difficult problems in AI: tasks for which they have not been explicitly trained. Their approaches range from studying millions of web videos and running computer simulations to collecting behavioral data from humans wearing gloves with built-in sensors.
Enigma, the institute scheduled to emerge from stealth on Monday, is taking a radically different approach. Rather than focusing purely on model functionality, the less than one-year-old startup wants to study how humans interact with robots, and hopes these interactions will lead to intuitive interfaces and perhaps a different kind of robot brain.
To fund its mission, Enigma raised $70 million in a seed round led by Index Ventures and Ribbit Capital with participation from Sarah Guo of Conviction Partners.
To test how humans want to communicate with machines, Enigma is launching a large-scale experiment that will allow anyone around the world to interact online with more than 100 of its own AI robots. The robots, located in hangars in Israel and California, can perform tasks such as painting with paintbrushes, fighting each other with swords, and performing simple chemical experiments such as picking up flasks and mixing them with liquids. Enigma claims to have developed both the robotic arm and its underlying model completely from scratch.
The youngest employee in Microsoft history, Jonathan Jacobi (pictured right), was hired by Wiz founder Assaf Rappaport during his tenure there and co-founded Enigma with his longtime friend Gal Nib (pictured left). The two met while participating in a hacking contest as teenagers and became close friends while serving together in Israel’s elite Unit 8200, where they conducted cybersecurity research.
When Jacobi and Niv set out to launch a startup together last year, they decided to apply their technical expertise to AI for robotics. Although they lacked direct experience in this area, it was an area they believed had the most exciting unsolved problems in technology.
They assembled a team of people Jacobi describes as their “smartest friends” from the Israeli tech ecosystem and community. These include graduates of top AI labs, winners of the Mathematics Olympiad, and several who have decided to drop out of their PhD programs.
“While many robotics industry players are joining the next wave of embodied intelligence, Jonathan and Gal are outsiders; they are not roboticists. That gives them more room for creativity,” said Shardul Shah, partner at Index Ventures. “An insider might start with remote control and dexterity capabilities, but Enigma is starting from a completely different place: ‘What is the ultimate experience?'”
Jacobi told TechCrunch that Enigma aims to completely ease human-robot interaction.
“If you have to do the dishes and you spend 15 minutes telling a robot where to put everything, you reach a point where you say, ‘Forget it, I’ll do it myself,'” Jacobi says. “Everyone is at that stage now, even with the most high-performance models.”
Jacobi believes that eventually operating a robot should be as intuitive as adjusting the volume knob in your car. He argues that users would be frustrated if, instead of turning a dial, they had to adjust the volume by a set percentage without knowing whether the result was too loud or too soft.
Enigma hopes that the data collected from the online experiment will reveal an interface that is the robotic equivalent of a car’s volume knob. The startup’s public tests will evaluate different ways people can communicate with its robots.
“We’re going to learn a lot about what is the right way to interact with robots,” Jacobi said. “Do you want to just have a conversation via text or audio? Do you want to show examples as a video? Or do you want to tap, drag, and drop?”
Jacobi admits that the Enigma experiment is very open-ended. The hope is that by collecting real-world data on human-robot interactions, the startup will be able to discover not only better interfaces, but also better ways to train the underlying AI models.
The company may ultimately figure out how humans prefer to communicate with robots. But for now, the business use case itself remains a mystery. Jacobi declined to share specific use cases for Enigma’s AI, but said the startup is already partnering with companies in healthcare, logistics, and entertainment.
If you buy through links in our articles, we may earn a small commission. This does not affect editorial independence.
