We spent a long time considering whether to buy Looki.
Price and specifications were not the reason. We hesitated because we understood that
this is not a mature consumer electronics product.
Looki's slogan is 'Own Your Story.'
With 1080p video, a 30-gram body, and no screen, the hardware is not remarkable by conventional standards. The concerns surrounding it are also legitimate: covert recording, privacy, social boundaries, and whether the product delivers enough value.
It is exceptionally light, however.
Judged only as hardware, we probably would not have purchased it.
What kept drawing us back to Looki was a different question:we have become increasingly dissatisfied with the way AI is used today.
We are not AI skeptics.
Over the past two years, AI has become an everyday tool at GeekOnUp for writing, research, coding, and product planning. In many situations, it is materially more efficient than a person working alone.
The longer we use it, however, the more noticeable a certain disconnect becomes.
The problem is not that AI lacks intelligence. It may be that the intelligence is unusually broad.
It can answer almost any question.
Reframe the question and the response often improves.
But every conversation requires me to stop what I am doing.
I must compress what is happening in my life into a sentence.
I must decide how to describe it, how to ask, and how to remove ambiguity.
Before the AI can help, I must organize the world on its behalf.
That gradually reveals a contradiction:a system with access to enormous knowledge has no idea what I am currently experiencing.
It does not know whether I am tired, how I have felt today, or whether a decision is emotional rather than logical.
That disconnect is what made an initially unappealing product like Looki interesting to us.
The product itself was less compelling than the problem it attempts to address.
01
I wore Looki for several days
There was no ceremony in beginning the experiment.
I did not choose a special time or plan a controlled scenario. I attached it to my chest and continued through several ordinary days: going out, walking, eating, commuting, and sitting down to work.
At first, the experience felt slightly strange.
The device is nearly weightless, but that makes one fact more noticeable:you are being recorded.
Even without a screen or notification sound, the awareness that a device is present is difficult to ignore during the first few hours.
That awareness did not last long.
Once I returned to normal activities, life itself took over my attention. By midday I had almost forgotten the device. I had pressed its button before leaving and had not interacted with it again.
Image quality still needs to improve.
The more thought-provoking moment came afterward.
When I returned to the phone and reviewed the day that AI had organized automatically, the feeling was difficult to describe.
It was not the shock of a spectacular demonstration. It was a quieter, persistent shift in perspective:an ordinary day could genuinely be reviewed in this form.
This was different from browsing fragmented photos. The day had been reorganized:
commuting, working, and unstructured time were separated,
periods of movement and rest were identified,
and moments I had already forgotten had been retained on my behalf.
I asked several ordinary questions:
'What did I do today?'
'What did I eat for lunch?'
'Have I been working late too often recently?'
The questions were simple, but the answers were produced differently.
Instead of inventing a summary, the system could return to specific images and explain
which observed moments supported its conclusion.
That is when I realized that Looki offers more than improved recording.It creates a passive sense of confirmation grounded in evidence.
For a technology team that builds systems and platforms around traceability and reuse, the pattern felt familiar.Looki behaves more like an automatically generated life log than a camera.
Life is preserved without deliberate documentation, and the result issearchable, understandable, and reviewable.That changes the nature of the record.
A video captured and edited by Looki
The experience is not perfect.
Image quality is limited, and moving scenes do not remain exceptionally stable.
The AI can also sound too confident when interpreting complex emotions or ambiguous situations.
Some features need a much longer history of data before they become meaningful.
Those limitations did not make us reject the concept immediately.
Instead, they clarified that this is not a product whose valuecan be determined after one use.Its value, if it develops, will emerge over time.
That resembles our experience building software and systems.
The most valuable capabilities are rarely one-time deliverables. They improve through continuous use, correction, and development.
02
After Hands-On Experience
After wearing it for several days, I began to see that
the underlying issue extends beyond one product. It is common to the current form of AI.
Today's leading models are exceptionally good at language.
Life does not occur primarily as language.
For most of the day, we are not expressing ourselves. We are walking, observing, hesitating, weighing options, or doing nothing in particular.
Almost none of that context enters the AI system.
The result is a disconnect:
AI knows a great deal about the world,
but very little about the situation a person is actually in.
That is why its advice can be logically correct without feeling relevant.
The same issue appears when we introduce AI collaboration and agentic systems inside organizations.Without continuous context and an accumulated view of real state, even an advanced model remains at the response layer.
From that perspective, Looki is an experiment in direction rather than a final product answer.
As a camera, it has many weaknesses.
As a sensory interface for AI, it becomes more interesting.
The central question is not whether every image looks good, but whether AI can maintain a continuous view of a person's day.
Looki is not trying primarily to improve recording.
It is trying to record without interrupting life.
Before using it, I was not convinced by that distinction.
Afterward, I at least understood why the product was designed this way.
03
Taking a Step Back
Following the idea further takes the discussion beyond Looki.
If AI is to enter its next phase, it may need to participate in life itself.
It must not only summarize or simulate life, but acquire context from living alongside it.
Positive feedback from an international Looki user
An AI-generated comic created by Looki
This is why spatial intelligence is receiving increasing attention.
AI cannot fully understand the world if it has never meaningfully experienced it.
In that sense, AGI will not result from one model release.
It is a path that requires time and continuous contact with the physical world.
That is why GeekOnUp continues to purchase and use emerging AI software and hardware, then document each firsthand experience in detail.
For us, this is both technology research and a product-minded way of working.
04
Conclusion
I do not know whether Looki will succeed,
or whether this form will become the lasting answer.
I am increasingly certain, however, that truly general AI may first need to learn how to participate in an ordinary person's day.
The future interface to AI may not be an app that someone chooses to open.
It may be a persistent presence that does not constantly demand attention.
It may be quiet and visually unobtrusive,
yet understand how the day has actually unfolded.
Follow the GeekOnUp Technology Team。
for ongoing field experienceacross AI, product, and engineering,including practical lessons frommobile apps, mini programs, and web systemsbuilt for real operating environments.
