i recently went to see my friends’ band perform at a small, sweaty venue, where everyone knew each other’s names. it wasn’t a coachella—just the kind of place where you could feel the heartbeat of the atmosphere and catch the lead singer’s eye mid-verse.

yet, almost every single person had their phone out recording the performance. a sea of glowing rectangles between me and the music, each one capturing what i was seeing with my two eyes from slightly different angles (oh, and the occasional person who didn't realize they weren't even recording). i get it: let's support our friends, let's “preserve” our favorite songs. but watching everyone experience the concert through their camera app felt sad. ten years ago, this wouldn't have been the case. clearly, our mindset behind & pace of capturing media is changing.

i was left wondering: fundamentally, why do we capture media?

from the beginning of time, we’ve captured media to externalize our memory. whether it be through cave paintings at lascaux preserving hunting scenes from thousands of years ago or daguerreotypes requiring eight-hour exposures to fix a single moment in silver, we’ve captured media to hold onto moments that would otherwise simply fade into the ephemeral nature of human experience. it's about preserving not just the visual/auditory data, but the emotional context and meaning of our experiences. we record to share, to remember, to prove we were there, and increasingly, to augment our own limited biological memory with digital permanence.

but what if the very act of deliberate capture—pulling out a phone, framing a shot, hitting record—is fundamentally at odds with actually experiencing the moment?

when cameras disappear

the landscape of capture technology is evolving—or rather, societal attitudes have finally caught up. what previously were failures—google’s glass and snapchat’s spectacles—are now inspiring a new generation of ambient capture devices.

meta x ray-ban’s smart glasses represent the first commercially successful attempt at mainstream wearable capture, selling over 2m units in just a few months. unlike their predecessors, these glasses prioritize design over features—they look normal, even trendy, rather than like sci-fi gadgets strapped to your face. the key insight: people won't wear a computer on their face unless it looks "normal"—or better yet, looks good.

but the real revolution isn't in the hardware—it's in the ai-first approach to capture. modern smart glasses aren't just cameras; they're context-aware assistants equipped with microphones and cameras that understand and react to your environment. meta's ray-ban glasses use meta ai to handle voice commands, while upcoming devices will leverage even more sophisticated ai for real-time translation, object identification, and contextual awareness.

meanwhile, a wave of startups are pushing even further into the ambient territory:

• previously, humane's ai pin clips to your clothing as a "screenless, seamless" ai assistant that uses cameras and sensors for contextual computing interactions, responding to gestures and voice without requiring you to look at a screen

• bee ai creates an "always-on personal ai memory" through a tiny wearable that listens to everything you say and hear for a week, then acts as your second brain for querying conversations and generating insights

• omi goes even further, combining always-listening microphones with experimental brain-computer interfaces to detect when you're addressing the device versus talking to someone else.

• friend takes a different spin by trying to capture media to target emotional companionship rather than productivity.

• xpanceo is developing a sci-fi ar smart contact lenses that would overlay digital information directly onto your field of vision.

• even our beloved cluely hinted at this future in their launch video, although this was more of a method of information retrieval (to help roy’s rizz) as opposed to just necessarily capture.

there’s a clear trend to move from deliberate capture to passive sensing. instead of deciding when to record, the ai decides what's worth preserving based on context, emotion, and importance.

the ambient paradigm

the way that i see it: the current wave of media/perception capture is not just a hardware upgrade—it's a paradigm shift interweaving social acceptability with technical capability.

devices like meta's ray-ban smart glasses selling millions signal that we've entered a new chapter where society is finally ready for more integrated ways to record and augment their lives. the winning products in this space will therefore be those that make technology disappear: delivering magical functionality (memory enhancement, real-time knowledge, effortless sharing) in a form factor that feels natural. the entire premise of capture will become more about perceiving and remembering rather than just taking pictures or videos.

ambient capture means your devices are always sensing the world as you go about your day, ready to intelligently preserve what matters. this is passive and ai-mediated: the devices won't dump hours of raw footage on you. they will leverage ai to decide when to record and what to surface to you later. your smart glasses might automatically snapshot a joyful candid moment with your family, or record the exact 30 seconds of an important instruction someone gives you, all without you explicitly hitting a button.

critically, this future shifts from capturing media to capturing perception and experience. traditional cameras capture images or video that you later look at. next-gen devices capture richer contextual data—what you looked at, where you were, who you met, even your biological signals—capturing all the necessary biodata to log an experience.

stepping into memory

given all the biodata to produce an experience, a logical follow-up is: will there be a way to relive memories? can i record something and then step into it and re-experience it?

the trajectory of ambient capture suggests we're moving toward exactly this—experiential playback rather than just media consumption. as devices capture not just video and audio but spatial data, biometric information, and contextual metadata, we're building the foundation for immersive memory reconstruction.

imagine querying your ai assistant: "take me back to that conversation with rahul at the conference," and instead of getting a transcript or video clip, you're transported into a vr recreation of that moment—complete with spatial audio, the ambient lighting, even your own emotional state reconstructed from biometric data captured at the time. with google’s recent project beam, this seems more and more possible by the day.

clearly this isn't just about better cameras or more storage—it's about creating a digital extension of human memory that doesn't just preserve information but preserves the qualitative experience of being present in a moment.

the invisible interface

this is where the concept of ambient capture is birthed. the future of media capture is one where you might not even think about "using a camera"—you'll simply live your life, and your personal devices and ai systems will seamlessly handle the rest, from preserving precious moments to letting you know when to relive it (and how to).

in the ambient future, the boundary between experiencing and recording blurs: our devices will increasingly share in our perception, augmenting human memory and creativity in ways that are both exciting and scary.

but perhaps the most radical aspect of ambient capture isn't technological—it's behavioral. when capture becomes as effortless as seeing, we might finally be liberated from the sore of seeing phone screens at concerts or from the anxiety of getting the perfect flicks for your instagram-obsessed friends.

we might, paradoxically, become more present by letting our devices handle the remembering.

it seems that there's a growing plague of inauthenticity within people (which is not to be confused with inauthentic people). so many people are chasing after what other people find interesting or valuable—i mean have one stroll on tech twitter or on linkedin, and you'll see countless examples of people building things solely to get some sort of recognition for their build. or the infamous "cs projects to put on resume" pipeline for students. and i want to clarify here: i don't think posting about your build is bad by any means; quite the opposite, it's really rewarding to see. but, when you're solely building for the purpose of posting or some sort of extrinsic reward, i think this mindset can be inimical.

i describe this as a plague because it's super easy to fall trap into this mindset—trust me, i did too. you see all these brilliant people building cool things, and they're always going viral too. it's in human nature to try to—or at least want to—emulate that same process. then, you become a random node in this network producing "hot" meaningless builds, that may not even interest you. solely for the purpose of external validation. the more nodes in the network, the easier it is for someone else to be infected (to continue the analogy).

this past weekend, i participated in this private hackathon sponsored by eric schmidt, the ex-ceo of google, and sebastian thrun. i was with some of the top minds from the top universities in the world, and what surprised me the most wasn't really their accolades: it was their genuine curiosity and deep interest in tackling the difficult problems we were presented with.

the robustness of thought, their commitment to thoroughness, and the fun they had while being challenged with difficult questions is not something that can be artificially replicated. you can't pretend to be interested. either you were, and you built something amazing in the 6 hours that we had to hack, or you weren't, and you didn't. it wasn't about who was the most technical. it was just simply about how interested you were in the problem at hand. these individuals approached problems with a unique perspective, driven by a desire to understand and innovate, rather than to achieve external validation. in this sense, they were being authentic.

i realized that it's this authenticity that i was missing. i realized i've been navigating the cs world with a broken compass. for the longest time, i've avoided looking at some of the low-level, fundamental projects that help one learn, driven by this belief that it wasn't necessary. i was told to build these complex projects, that i really didn't care too much about, just so that i can put it on my resume. i always thought it was about being the best at programming. but, i learned that it's really not: i strongly believe that while that style of education might help, if you really just chase what you're interested in—what you're truly excited about—then you'll be fine. better than fine.

i want to dedicate the rest of my summer to just exploring the things i really am interested in, and putting less of an importance on the things that may be extrinsically motivated. for me, this translated into the following research goal: understanding how this world works and how everything fits together in it—from a philosophical perspective but also from a technical perspective.

hopefully, that's the cure to the plague.