A secretive AI lab founded by serial entrepreneur Brett Adcock unveiled new insights into what it claims is an innovative blend of model-building and hardware design poised to transform human interaction with intelligent software.
The company announced plans to create multi-modal end-to-end models, alongside their hardware and interfaces, to offer a “seamless end-to-end personal intelligence product.” This system will maintain a persistent memory of your life, with capabilities to listen, see, and engage with the world in real time.
Though specifics remain under wraps, Hark’s aspirations mirror Silicon Valley’s quest for the standout application to market AI as a coveted consumer product, rather than a feature awkwardly integrated into existing digital platforms.
“In my opinion, today’s AI models fall short in intelligence, appearing quite dumb as the devices for accessing them are fundamentally pre-AI,” Adcock mentioned in a January memo shared with TechCrunch. “We’re heading towards a future resembling sci-fi characters like Jarvis or Her, with systems that anticipate, adapt, and genuinely care for their users.”
Details are scarce, but Hark highlights Director of Design Abidur Chowdhury as a vital addition. Previously an industrial designer at Apple, credited with leading the team behind the iPhone Air and other models, Chowdhury left last fall after meeting with Adcock and endorsing his vision for modernizing human automation.
In an exclusive TechCrunch interview, Chowdhury refrained from revealing Hark’s roadmap but hinted at the anticipated first release of the company’s AI models this summer. Commenting on different approaches to AI integration, the designer offered some insights.
Chowdhury remarked on the ongoing use of outdated devices in a changing world, emphasizing untapped potential if intelligence formed the base layer of our interactions rather than being treated as an app or website.
Chowdhury pointed out the awkwardness of routine tasks like completing forms, sharing information across devices, booking travel, or planning home renovations.
“We firmly believe that the myriad small tasks that accumulate into significant burdens today can be automated from our lives,” Chowdhury stated.
While Chowdhury acknowledges Hark’s certainty about its goals, he cannot yet clarify how users will experience them. He implies that wearables similar to Meta’s Glasses seem improbable.
“I’m skeptical about many current wearable AI platforms,” Chowdhury stated. “I don’t believe it’s suitable to insert a layer between humanity and the interfaces we use globally. I share similar apprehensions with pins or other items involving cameras.”
Initially viewing generative AI as a fleeting trend, Chowdhury eventually recognized its potential to transform his work. The name Hark signifies attention, aligning with the company’s mission.
“User experience traditionally involves discovering the simplest solution for everyone,” he told TechCrunch. “Future user experience will focus on identifying the suitable option for each individual. I believe that is achievable, but it requires considerable effort.”
Hark’s emphasis on elegance and user simplicity echoes Apple’s design successes and naturally invokes thoughts of Jony Ive, the renowned former Apple designer now working on AI-driven hardware at OpenAI. A comparison that Hark’s spokesperson chose not to comment on.
A similar synergy can be observed between Elon Musk’s xAI advancements and Tesla’s quest for autonomous vehicles and humanoid robots.
A comparable corporate synergy exists between Adcock’s humanoid robotics venture Figure and the new AI labs. Hark’s models are already training on Figure’s robots, though the purpose remains unclear. An insider suggests no merger plans between the two firms.
Hark employs 45 engineers and designers, including former Meta AI researchers and former Apple and Tesla designers, all working on the same campus hosting Adcock’s other companies. Hark plans to utilize a newly established cluster of thousands of NVIDIA GPUs in April.
Now, Hark, supported by $100 million of Adcock’s personal seed capital, enters the race for talent as global giants strive to develop a format that integrates deep learning models into everyday life, at a time when dissatisfaction with current digital life models is peaking.
“It feels like there’s potential for something better, a sentiment I haven’t experienced since the iPhone emerged,” Chowdhury said.