A new social app called Bond is taking direct aim at one of the internet’s most entrenched behaviors: doomscrolling. Instead of competing to keep you glued to an endless feed, Bond’s founders say their goal is to “use AI to get you off the couch and back into the real world.” The iOS and Android app, which is now rolling out publicly, strips away familiar engagement hooks like infinite scroll and algorithmic timelines and replaces them with an AI system designed to turn your own memories into prompts for real‑world experiences.
A social network without a feed
At first glance, Bond looks like yet another visual‑first social platform, built around profiles, circular avatars and rich media posts. Look closer, and a core ingredient is missing: there is no scrolling feed.
“Bond has no feed and no infinite scroll,” co‑founder and CEO Dino Becirovic says. Instead of a central, constantly refreshing stream, the app is organized around “memories” photos, short videos and audio clips users upload to document what they are doing and what they care about over time. Friends can tap through someone’s profile to see these stories, which surface for 24 hours and then move into a private archive the user can revisit later.
The absence of a feed is deliberate. Becirovic argues that traditional timelines are “the engine of doomscrolling,” tuned to reward outrage and novelty rather than wellbeing. “Most platforms today are engineered to maximize time on site,” he says. “We’re deliberately optimizing for the opposite: how quickly can we get you inspired and back into your life?”
Turning memories into suggestions
Bond’s distinctive twist is what happens behind the scenes once you start posting. Those memories are not just a scrapbook; they are training data for the app’s AI. Over time, the system analyses what you share, where you go, what you eat, which bands you see, what hobbies you mention to build a picture of your tastes and habits.
“Experiences stored within Bond become fodder for its AI system,” Becirovic explains. The platform’s models use that history to propose specific, real‑world ideas tailored to you. The examples he gives feel intentionally grounded in everyday life. “If you have been posting about how much you like pho and how you have not had it in a while, Bond might recommend a nearby Vietnamese restaurant getting good reviews,” he says. “If you are into heavy metal, it might tell you Iron Maiden is playing in your city next week.”
Where conventional social algorithms learn what content will keep you scrolling and then feed it back to you, Bond’s engine is trying to learn what will get you to close the app and walk out the door. The company describes the platform as “tailored to encourage users to step away from the app and engage in real‑world activities,” a direct response to the culture of “bed rot” and late‑night “dooming” that has grown up around endless feeds.
Designed as an antidote to doomscrolling
Despite its ambitions, Bond does not throw away every piece of social media’s visual language. Profiles are still arranged in clusters, memories still appear in tappable circles, and the interface borrows the short‑lived, story‑like format most users already understand. The difference lies in what you can’t do.
There is no central river of content to fall into while bored. You cannot pull to refresh and watch the app refill with new posts. There is no bottomless queue of algorithmically chosen clips lined up to keep you watching for “just one more minute” that turns into an hour. Instead, each session is structured around a small number of intentional actions: add a memory, check in on a few people you care about, see what the AI suggests you go and do next.
Becirovic characterises Bond as “an AI‑powered solution to screen addiction,” a deliberate attempt to flip the logic that underpins much of today’s social web. “We don’t need another app that keeps you glued to your phone,” he says. “We need one that nudges you to put it down.”
His critique echoes a growing body of research linking infinite scroll and engagement‑optimised feeds to anxiety, disrupted sleep and compulsive use, especially among younger users. Where that research often frames AI as part of the problem powering ever more precise engagement engines, Bond is trying to demonstrate that the same technology can be pointed in the opposite direction.
Memory chat and control over what AI knows
Because Bond’s AI model is built on highly personal material, the question of who controls that “memory” is central to the product. Each user has access to a dedicated Memory section inside the app, which shows what the system has stored about their past posts and inferred preferences. On top of that, there is a conversational interface called “Memory chat” that lets you interrogate and edit the AI’s understanding of you in plain language.
“Memory chat” functions like a private assistant for your archive. You can ask it what it thinks you like to do, tell it to forget certain events or places, or wipe entire categories of data that you no longer want the system to use. You can also, in Becirovic’s words, “delete your entire profile if you choose,” clearing out both your memories and the model built from them.
“Our view is that your memories are yours, not some ad network’s,” he says. At launch, Bond stores user data in its own infrastructure and does not yet offer end‑to‑end encryption, but Becirovic calls stronger protections “a priority for us in the near‑future” and says the team is exploring more robust cryptographic guarantees for the archive.
The ability to see and edit what an AI has learned from you, rather than treating its internal model as a black box, aligns with broader efforts in the field to make systems more explainable and user‑directed. In Bond’s case, it is also a trust play: if the app is asking for your life’s highlight reel, it needs to show clearly what it is doing with it.
Rethinking engagement and business models
Bond is entering the market at a moment when big social platforms are under political and regulatory pressure over their impact on mental health. Lawmakers have hauled executives into hearings, researchers have published evidence of harmful usage patterns and parents have become more vocal about the costs of constant connectivity. In that context, a startup promising to help users “kick your doomscrolling habit” stands out.
Whether that promise is compatible with a sustainable business is the open question. Traditional social networks make money by selling ads against attention; their economic incentives are tightly coupled to keeping people looking at screens for as long as possible. Bond’s team insists they are exploring a different path.
“We’re trying to prove you can build a sustainable social product that doesn’t monetize addiction,” Becirovic says. Instead of charging brands simply for impressions, the company is experimenting with models “aligned with real‑world value” for example, partnerships with venues, events and local businesses that users actually decide to visit after receiving a recommendation.
Even the way the team talks about metrics hints at a different set of priorities. The company is less interested in daily time spent and more in whether suggestions translate into action. “The question for us is not ‘How long did you scroll?’” Becirovic says. “It’s ‘Did we help you do something you’re glad you did?’”
A quiet test of what users really want
For now, Bond is available to anyone who wants to download it, and early adopters are starting to discover what social media feels like when the feed is removed from the equation. The interface is familiar enough that most people can navigate it instantly, but the absence of that endlessly refreshing stream is noticeable. You check a few profiles, add a memory, maybe tap on a suggestion and then, if the app works as intended, you put your phone away and go outside.
Changing deeply ingrained habits will not be easy. Doomscrolling is as much about psychology stress, boredom, avoidance as it is about interface design. Still, AI systems that learn from our digital traces have already been used to detect and manage other risky behaviors, from substance use to gambling. Bond is betting that similar techniques can be repurposed to interrupt the impulse to scroll and replace it with something more constructive.
As Becirovic frames it, “If AI can predict which video will keep you up at 2 a.m., it should also be able to predict which plan will get you out of bed tomorrow.” Whether millions of users, and an industry built on capturing attention, are ready to embrace a social app that measures success in minutes spent away from it may be the platform’s ultimate test.
Comments