How will Q.ai, acquired by Apple for $2 billion, revolutionize human-computer interaction in the AI era?

Recently, Apple acquired Q.ai for nearly $2 billion. This is Apple's second - largest acquisition to date, second only to the $3 billion acquisition of Beats in 2014.

Q.ai uses AI to enable devices to recognize human whispers in noisy environments. It also has a technology to detect subtle movements of facial muscles. Both of these technologies have the potential to revolutionize the interaction methods of smart wearable devices.

Before being acquired by Apple, Q.ai received seed - round financing led by Google Ventures (GV) and Aleph in 2022, and Series A financing participated in by GV, Kleiner Perkins, and Spark Capital in 2023. The total financing amount was in the tens of millions of dollars. Matter Venture Partners and Corner Ventures were also on its list of investors.

Perception - based interaction allows AI hardware to finally "blend" into our daily lives

After the acquisition, Q.ai's core team, including Aviad Maizels (CEO) and co - founders Yonatan Wexler and Avi Barliya, will join Apple.

This is the second time Aviad Maizels has sold a company to Apple. In 2013, PrimeSense, a 3D sensing technology company he co - founded, was acquired by Apple. The technology of this company was initially applied to Xbox Kinect and was later transformed by Apple into the core technology supporting iPhone FaceID.

A group photo of some members of Q.ai's core team with Nabeel Hyatt, an investor from Spark Capital. Source: Nabeel Hyatt

Among Q.ai's core team, Yonatan Wexler is the recipient of the Marr Prize (the highest honor in the field of computer vision). His achievement, OrCam MyEye 2 (which helps visually impaired people "see" text and recognize objects), was named "Best Invention of the Year" by Time Magazine. Avi Barliya is a serial entrepreneur who has participated in the development of everything from autonomous driving systems to the SpaceIL lunar probe.

Interestingly, a few weeks ago, a company called Lyte received $107 million in early - stage financing. Lyte was co - founded by Alexander Shpunt, Arman Hajati, and Yuval Gerson. Alexander Shpunt co - founded PrimeSense and served as its CTO.

Lyte develops an end - to - end visual perception system that combines hardware and software, providing a perception layer that connects the intelligent layer for physical AI, enabling direct communication between the "eyes" and the "brain" of robots.

What Q.ai does seems to be in line with the thinking of the PrimeSense era.

Whether it's Xbox Kinect or iPhone FaceID, they are essentially perception - based interactions.

Kinect enables somatosensory/gesture interaction through visual perception, and FaceID enables interactions and functions such as unlocking and payment authentication through visual perception.

Q.ai has currently announced two technologies. One uses AI to enable devices to recognize human whispers in noisy environments and optimize audio performance in complex environments. The other is to detect and recognize subtle movements of facial muscles to detect lip movements or voices, identify users, and evaluate their physiological indicators such as emotions, heart rate, and respiratory rate. They are both perception technologies.

How can these two perception technologies be applied to interaction? We speculate that it may be a combination of whisper recognition and lip - reading recognition. Currently, the interaction methods of AI - native hardware are mainly based on buttons + voice. However, compared with traditional pure touch - screens, voice interaction requires clear pronunciation, which may cause "social embarrassment" in some situations and make users hesitant when interacting with devices.

Apple has always attached great importance to interaction and user interfaces and is a master at creating "invisible interfaces." They hope that technology will become a natural extension of users themselves.

After joining Apple, Q.ai's technology is expected to reach billions of users and fundamentally reshape the interaction methods between humans and machines and between humans. This makes people wonder what it will be like when AI hardware finally "blends" into our daily lives.

The evolution of the interaction interface essentially lies in increasing the "bandwidth" of human - machine interaction

The development and popularization of personal computing devices are jointly promoted by factors such as computing power, intelligence (functions), volume, and interaction interfaces.

Under Moore's Law, the computing power of devices per unit volume and cost continues to increase. This has driven the continuous miniaturization of personal computing devices from PCs to mobile phones. Moreover, with the support of computing power, the functions of personal computing devices have become more and more diverse.

In terms of the interaction interface, its evolution is a process of continuously reducing resistance and lowering the threshold.

Initially, the interaction interface of personal computers was the command - line interface (CLI), and at that time, its user group was concentrated among developers and computer enthusiasts.

Later, the GUI and mouse - keyboard were introduced to PCs, making precise text editing and table making possible. This enabled professionals in various industries to use PCs, and with the further growth of users, it promoted the popularization of the Internet.

The emergence of the iPhone brought a multi - touch interaction method, which further increased the popularity of personal computing devices, and super - apps with billions of users began to appear.

Why does the reduction of resistance in the interaction interface lead to a lower threshold? Essentially, it increases the "bandwidth" of human - machine interaction. From the command - line interface to the GUI and then to multi - touch, people think less and operate more intuitively.

In the AI era, smart wearable devices have emerged, and the relationship between humans and machines has begun to change. In the past, humans precisely controlled machines, but now machines will learn to understand humans.

What is the most common way of interacting between humans and smart wearable devices at present? It is multi - modal interaction. Multi - modal interaction includes both voice and some gestures. In short, it is perception - based interaction.

This perception - based interaction actively and directly captures and analyzes human thinking activities. It is expected to establish the highest - bandwidth human - machine interface to date and has the potential to bridge the gap between human thinking and machine interaction.

At present, for smart wearable devices, AI models are constantly developing. Thanks to the foundation laid by smartphones, a path has been initially found in terms of volume (usually related to battery life and edge - side computing power). If the interaction methods converge in the future to form a default interaction method like multi - touch, the energy generated by the intersection of AI and the physical world will be huge. The popularity of AI smart hardware in the future is likely not to be lower than that of smartphones.

This article is from the WeChat official account "Alpha Commune" (ID: alphastartups). Author: Discover Extraordinary Entrepreneurs. It is published by 36Kr with permission.