Equity

Mina Fahmi on wearables that extend you, not replace you

Mina Fahmi· Co-founder and CEO at Sandbar
·~29 min·English·TechCrunch
MultimodalAI CompanyBusiness Strategy
TL;DR

Sandbar CEO Mina Fahmi argues the wave of always-on AI wearables gets the relationship backwards: his Stream voice ring is push-to-talk by design, because a device only earns a place on your body when it feels like a controllable extension of you rather than a passive recorder listening in.

01Core Mental Model

Self-Extension, Not a Companion

Fahmi's north star is not an AI that acts like a friend or coworker but one that feels like a part of you, technology that augments the person instead of replacing the people in their life.

I was excited about human computer interaction because I believed it was a way to augment us, not to replace us.

— Mina Fahmi, Equity
Key Insight
Most consumer AI is sold as company, a friend, a coworker, a stand-in for people. Fahmi is quietly rejecting that entire category: if the goal is to extend you, the product's job is to disappear into your intent, which is a very different design brief from being good company.
02The Problem

Nothing Keeps Up With the Speed of Thought

The gap Stream targets is the lag between having a thought and capturing it, because unlocking a phone and opening an app is too slow for people who live on the go or think out loud.

most products today cannot keep up with the speed of thought

— Mina Fahmi, Equity
Key Insight
The framing turns a convenience feature into an accessibility one. Fahmi points to people with dyslexia and to working parents, people for whom reading-and-writing tools are a tax, as his most active testers, hinting the real market is not gadget lovers but anyone the keyboard has failed.
03The Design Principle

You Drive It, It Doesn't Watch You

Every major computer interface, from the command line to the mouse to multi-touch, was driven by deliberate human input, and Stream continues that lineage instead of passively recording everything around you.

you see a lot of devices and software that just records everything. But the downside of that is you don't know what the person actually thinks or wants.

— Mina Fahmi, Equity
Key Insight
There is a technical claim buried here: passive capture forces the system to infer intent, which is exactly where AI is least reliable. Push-to-talk sidesteps the hardest problem in ambient computing by asking the human to mark what matters.
04The New Primitive

The Mouse for Voice

Fahmi frames the ring as doing for voice what the mouse did for the graphical interface, a dedicated input that turns an awkward-to-invoke capability into something effortless.

it does for voice what the mouse did for the graphical user interface.

— Mina Fahmi, Equity
Key Insight
The analogy is a platform claim in disguise. If voice is a new interface layer the way the GUI was, then whoever owns its default input device owns a platform, which is why a startup that has raised about $36M is spending on a custom ring rather than just an app.
05Why Wearables Stall

A Wearable Has to Pass in Public

Fahmi's read on why so many AI wearables stalled is that whether people are comfortable being seen wearing one may matter as much as what the device can actually do.

social acceptability is perhaps the most important attribute of a wearable.

— Mina Fahmi, Equity
Key Insight
It reframes the recent wave of pins, pendants, and always-on recorders that struggled to find users: they leaned on raw capability, but the device still has to sit on a human body in social space. Fahmi's tell is that Stream is push-to-talk partly for aesthetics, since a continuously-recording device can read as hostile to everyone nearby.
06The Build Bet

Hardware Is Hard, So They Chose the Hard Path

Prototyping hardware has gotten dramatically easier, but taking a custom device to market at quality and volume is still brutal, and Sandbar chose full custom over reskinning an off-the-shelf ring.

we wanted this to be the best possible expression of the device that it could be

— Mina Fahmi, Equity
Key Insight
Note who is now in the prototyping loop: Fahmi says agents can design much of the electronics and housing, collapsing the idea-to-prototype gap. That is why he expects an explosion of hardware attempts, and why durable advantage shifts to the unglamorous part, manufacturing at quality and margin.
07The AI Stack

Many Models, No Monolith

A single interaction with Stream runs many open and closed models at once, because Fahmi expects neither open nor closed models to win the whole future outright.

I think it's very unlikely that we hit a future where it's just closed source models or it's just open source models.

— Mina Fahmi, Equity
Key Insight
This is a quiet counter to the personal-superintelligence vision of one model that is your whole world. Fahmi expects each side to keep an edge — closed models on raw compute, open ones on cost and control — so the durable design is orchestration across many, not a bet on one. He extends the same instinct to your data, wanting your notes to stay portable to the tools you already use rather than locked inside one app.
08The Future Is a Choice

An Always-On Future Isn't Inevitable

Fahmi rejects the premise that constant ambient recording is where wearables must end up, arguing a privacy-optional future is something people opt into rather than have decided for them.

we get to choose the future we want to be part of.

— Mina Fahmi, Equity
Key Insight
When proponents of always-on say social norms will simply adapt, Fahmi's rebuttal is a question: why would they, and who is asking for this? His concrete answer is design and data minimization — push-to-talk, a mic tuned to only a close whisper, encryption in transit and at rest, and zero retention with the outside processors he relies on — so the default leans toward privacy rather than capture.