AI is killing authentic dating - with Daniel Cheaib founder of Feels App

Dating apps weren't supposed to be slot machines. They were supposed to help people fall in love. But somewhere between swipe mechanics and engagement algorithms, the entire industry got hijacked by addiction models borrowed from social media. Daniel Cheaib, founder of Feels App, argues that fixing dating requires treating these platforms like what they've become: digital drugs that need regulation.

The core problem isn't technical, it's psychological. Modern dating apps fail because they've created impossible user expectations while simultaneously making the experience addictive rather than effective. "People come and they expect that you find them that you'll find their partner for life directly. Like, in the matter of, like, a few days. And if you haven't and if you have even been premium on the app and you didn't find them, the person of their life, they'd be pissed and they say, like, this app doesn't work," Cheaib explains. Users approach dating apps like Amazon, expecting instant gratification and perfect matching algorithms. When romance doesn't arrive on demand, they blame the platform rather than adjusting their timeline.

This expectation mismatch creates a vicious cycle. Apps respond by making the experience more engaging to keep users active longer, which means borrowing dopamine-driven design patterns from social media. The result is platforms optimized for retention rather than relationships. Cheaib draws the parallel explicitly: "I really think it's like a drug. And imagine it's like a drug, where the people who have these networks, they don't do it necessarily purposely to get people addicted. But as a consequence of having so much data of what they like and what they don't like, they are able to get you to to stay longer and longer and longer."

The generation using these apps most intensively, Gen Z, represents a fundamental shift in how people form connections. Unlike millennials who used online platforms primarily for dating, digital natives have been making meaningful connections online since their teens. "We have a generation that is purely digital native that made connections online with new people already in their teenage hood, sometimes with people that are closer for them than people they have as friends in real life," Cheaib notes. This creates both opportunity and danger. These users are comfortable forming deep relationships through screens, but they're also more susceptible to the addictive mechanics embedded in the apps.

Cheaib's solution isn't just building a better dating app, it's reimagining the entire category with drug-like restrictions. He advocates for treating social media consumption like controlled substances: "I think it's just it should be treated the same way as a drug because it is a drug. You can use devices. You can use screens. You can play games within the limited like, a real limit. But consuming content through the day like we do on TikTok or on reels, that's a no go."

This perspective comes from someone who nearly lost everything building in this space. Cheaib and his cofounders went two years without paying themselves, with team members returning to Google just to survive. The company hit rock bottom before an investor called three weeks later, changing everything. That near-death experience shaped his conviction that dating apps need fundamental reform, not incremental improvements.

The implications extend beyond romance. If the most intimate human connections are being mediated by addiction-optimized algorithms, we're conducting a mass experiment on an entire generation's ability to form relationships. The early results suggest we're failing.

Watch for dating apps that explicitly limit usage rather than maximizing it. Look for platforms that set realistic expectations about timing and outcomes. Most importantly, recognize that the current model, infinite scrolling through potential partners while dopamine systems keep you engaged, isn't broken by accident. It's working exactly as designed, just not for the outcomes users actually want.