Two bets on your user's attention, and you have to pick one

Two new products shipped the same week, and they point in opposite directions. Halliday put a green screen in your glasses to feed you more during a meeting. Light built a flip phone with no screen on the outside so nothing reaches you until you open it. Both are betting on where your user's attention wants to go. Let me catch you up on what they see, because the split is real and it lands on your roadmap.

The screen that follows your eyes

Halliday's Gen 2 glasses drop a monochrome green display into your field of view. No camera, on purpose, which helps privacy and gets you about 12 hours of battery. The pitch is work. Turn on "Meeting Flow" before your 2:30 staff meeting and the glasses transcribe, surface facts, and let you rewind 30 seconds if you zoned out. They cost $599 and ship in September.

The interesting part is not the transcript. It's the tracking. Victoria Song got a hands-on and flagged three features: Thread Tracker, Decision Confirmation, and Commitment Check. One shows how a conversation moved from budget to hires to timelines. Another confirms what got decided. The third checks whether action items got assigned, to whom, by when. That's a real product idea: the meeting managing itself while it happens.

When more input is the feature

Halliday's bet is that your user wants more coming in, as long as it's the right stuff at the right moment. COO Carter Hou told Lifehacker the goal is to "improve the efficiency of their work life." He talks about the glasses helping at a doctor visit, a networking session, even a traffic stop. An agent that listens to everything and tells you what to do next.

Song is honest about the risk. She's had "little success with AI wearables successfully intuiting what I actually need surfaced in a conversation." That's the whole game. If the AI surfaces the wrong fact at the wrong second, you've added noise to a meeting instead of removing it. The value lives entirely in the quality of what it chooses to show. Miss on that and you've built a distraction with a display.

The phone that makes you open it

Light went the other way. The Light Flip has a 2.8-inch screen inside and nothing on the outside. Cofounder Kaiwei Tang made that call so you have to deliberately flip it open. "You close it down and it goes away," he says. There's a small notification light on the front, and that's it. It's $299, or $39 a month with service, and it ships in 2027.

What drove this was feedback Tang didn't expect. Young users told him they didn't want a smartphone, and didn't even want something that looked like one. Tang, who helped build the original Motorola Razr, heard Gen Z describe their old flip phones in negative terms but stick with them and feel proud. So he gave them fast custom software in a body no one mistakes for an iPhone.

Same word, opposite meaning

Nokia is trying to have it both ways, and it shows. HMD put a voice-controlled AI button in the center of the D-pad on four retro feature phones. Press it to make a call, set an alarm, turn on the torch, without hunting through keys. Dezeen readers were not kind. One called them "dumb-idea phones." The idea of a simple phone that also carries an assistant reads as confused to the people who want simple.

Tang is blunt about which way the wind blows. "No one wants an AI button on their phone," he says of his interviews with young people. More users, he adds, ask him to guarantee there's no AI anywhere in the Light universe. Same technology, same week, two products treating "AI on a device" as either the selling point or the thing to promise you left out.

The deep cut

Both companies picked one job and refused the other. Halliday dropped the camera and the app store to be a meeting tool. Light dropped the touchscreen, the outer display, and the AI to be a phone you put down. Nokia tried to bolt an assistant onto simplicity and got mocked for it. The lesson for your team is not which side to join. It's that the products people trusted this cycle each named their user's one real job and cut everything that fought it. If your feature list is a hedge between focus and input, your user feels the hedge. Pick the job. Cut the rest. Light even built a "Flip Your Life" newsletter to walk buyers through the change, because they know a clear promise still needs support to stick.

Three questions for your team

• What is the one job our product does when the user is at their most distracted, and does every feature on the roadmap serve that job or hedge against it?
• If we surface AI help in the middle of a task, can we prove it shows the right thing at the right moment, or are we shipping Song's problem of the wrong fact at the wrong second?
• Are we adding an "AI button" because users asked for it, or because we assume they want it, and have we actually asked them the way Tang asked his users?