Two Harvard juniors, AnhPhu Nguyen and Caine Ardayfio, have captured global attention with their controversial I-Xray project—a facial recognition system that uses Meta Ray-Ban smart glasses to instantly identify people and retrieve their personal information from the internet. The demonstration, which went viral this month, showcased how the students integrated AI-powered facial recognition into consumer AR glasses, raising significant privacy concerns worldwide.
However, the I-Xray project represents just one chapter in these young innovators’ ambitious journey. The duo founded Harvard’s AR/VR club and have been building cutting-edge projects since their sophomore year. “We lived in the science and engineering building for like an entire summer and just built random projects,” Ardayfio explained. Their creations include a flamethrower, a finger-controlled electric skateboard, and a four-foot robotic tentacle capable of moving through the air.
Through their AR/VR club, they gained access to Meta’s smart glasses and developed another AI application: augmented reality glasses that can fact-check statements in real-time during arguments. Nguyen emphasized his project-focused approach to education, stating he prioritizes hands-on innovation over traditional academic metrics like GPA.
Now, the students are setting their sights on industrial applications of AI, particularly in manufacturing, construction, and industrial technology sectors. They believe recent advancements in large language models (LLMs) like OpenAI’s ChatGPT and Anthropic’s Claude have unlocked unprecedented opportunities for innovation.
Ardayfio provided a compelling example of AI’s transformative potential in construction: autonomous robots equipped with LLM reasoning capabilities can now make contextual decisions independently. Previously, if a construction robot encountered an obstacle like a person blocking its path, engineers would need to hardcode every possible response. With LLM integration, the robot can reason: “I’m going to wait for this person for a couple seconds until they move, and then if they don’t, then I’ll do this.” This reasoning capability ensures more natural, adaptive responses without extensive programming.
“That’s actually, like, very incredible, and a capability that has only existed in the past six months, essentially,” Ardayfio noted, highlighting how rapidly AI technology is evolving and creating new possibilities across industries.
Key Quotes
We lived in the science and engineering building for like an entire summer and just built random projects
Caine Ardayfio described the intensive hands-on approach he and Nguyen took during their sophomore year at Harvard, demonstrating their commitment to practical innovation over traditional academic pursuits.
I’m going to wait for this person for a couple seconds until they move, and then if they don’t, then I’ll do this, and you can have a very strong likelihood that whatever it does is pretty reasonable
Ardayfio explained how LLM-powered construction robots can now reason through obstacles independently, illustrating the transformative potential of integrating large language models into autonomous industrial systems.
That’s actually, like, very incredible, and a capability that has only existed in the past six months, essentially
Ardayfio emphasized the unprecedented speed of AI advancement, noting that contextual reasoning capabilities in autonomous systems have only recently become possible, highlighting the rapid evolution of LLM technology.
I’ve always been prioritizing projects a bit more than, like, my GPA because I don’t really want to get into grad school right off the bat
AnhPhu Nguyen explained his educational philosophy, choosing hands-on AI and hardware projects over traditional academic metrics, reflecting a broader shift toward practical experience in tech innovation.
Our Take
What’s particularly striking about this story is how it encapsulates both the promise and peril of accessible AI technology. The I-Xray facial recognition project serves as a wake-up call about privacy in an age where powerful AI can be integrated into everyday consumer devices. Yet the students’ pivot toward industrial applications reveals the more constructive potential of these same technologies.
The integration of LLM reasoning into robotics represents a genuine paradigm shift. Traditional robotics required exhaustive programming for every scenario; AI-powered systems can now adapt and reason through novel situations. This could be the breakthrough that finally makes autonomous construction and manufacturing robots practical at scale. The fact that these innovations are coming from undergraduate students, not just well-funded labs, suggests we’re entering an era where AI development is increasingly decentralized and rapid. Industries that fail to recognize and adapt to this pace of change risk obsolescence.
Why This Matters
This story illustrates the democratization of AI development and how young innovators are pushing boundaries in unexpected ways. The I-Xray project demonstrates both the remarkable accessibility of AI tools and the urgent need for privacy safeguards as facial recognition technology becomes more portable and powerful.
More significantly, the students’ pivot toward industrial AI applications signals a broader trend in the AI industry. While consumer-facing AI tools like ChatGPT dominate headlines, the integration of large language models into robotics and autonomous systems represents the next frontier of AI innovation. The ability for machines to reason contextually rather than follow rigid programming could revolutionize manufacturing, construction, and logistics.
Their work also highlights how recent LLM breakthroughs are enabling capabilities that were impossible just months ago. This rapid pace of innovation means businesses across sectors must adapt quickly or risk falling behind. For the construction and manufacturing industries—traditionally slower to adopt new technologies—AI-powered autonomous systems could address labor shortages, improve safety, and dramatically increase efficiency. The fact that college students can prototype these solutions underscores how accessible and transformative AI technology has become.
Recommended Reading
For those interested in learning more about artificial intelligence, machine learning, and effective AI communication, here are some excellent resources: