Another robotics company has entered the "embodied intelligence" field, one of the hardest challenges in AI - how to teach models to do tasks that no one has ever specifically taught them. The approaches vary widely: some analyze millions of online videos, others run computer simulations, and some embed sensors in gloves to collect human movements. Enigma, which emerged from hiding on Monday, took a completely different path: instead of focusing solely on model capabilities, it is studying how humans and machines interact, hoping to develop intuitive interfaces and even a different kind of machine brain from these interactions.

To fund this journey, Enigma raised a $71 million seed round led by Index Ventures and Ribbit Capital, with Sarah Guo from Conviction Partners also investing. This startup, less than a year old, aims to first understand how humans want to communicate with machines.

image.png

To bring this question into the real world, Enigma is launching a large-scale online experiment: anyone around the world can go online and interact with its 100 or so self-developed AI robots. These robots reside in garages located in Israel and California, capable of painting with brushes, fighting with swords, and lifting beakers to mix liquids for simple chemical experiments. Enigma claims that from the robotic arms to the underlying models, everything is built from scratch.

The company's two co-founders are like childhood friends. Jonathan Jacoby was once called the youngest employee at Microsoft - he was personally recruited by Wiz founder Assaf Rappaport when he was young; his friendship with Guy Nave began when they competed in hacking competitions as teenagers, later working together in Israel's elite 8200 unit on cybersecurity research, becoming close friends. When the two decided to start a business last year, they deliberately chose the field of robotics - a domain they were not experts in, but one that is known for containing the most fascinating unsolved problems in the tech world. They gathered a group of people from the Israeli tech circle whom Jacoby called "the smartest friends" - people from top AI labs, math competition gold medalists, and even a few who had been convinced to drop out of their doctoral programs.

Shardul Shah, a partner at Index Ventures, was particularly impressed by this "outsider" spirit: "Insiders may start with teleoperation or dexterity, but Enigma asks a very different question - what should the ultimate experience be?" In Jacoby's view, the interaction between humans and robots should be effortless. He gave a heart-wrenching example: if washing dishes required spending 15 minutes explaining where each item should go, people would just say "forget it, I'll do it myself" - and right now, even the most advanced models are still stuck on this point. He believes that dealing with robots should eventually be as intuitive as turning the volume knob on a car: no one likes to adjust the volume with percentages, yet they don't know whether the result is too loud or too quiet.