Algorithms play a decisive role on social networks, determining which posts users see. They ultimately help major platforms like Meta or X to monetise attention.
As Steve Muller, a technical adviser at Bee Secure, explained, platforms consider factors such as how popular a post is, how prominent its author may be, and how relevant the content appears to each individual user.
In recent years, another challenge has emerged: the growing flood of AI-generated content, particularly deepfakes. These are images, texts, or videos manufactured by generative artificial intelligence to appear convincingly real.
Because detection tools are still imperfect, experts argue that users need to shift their mindset. Rather than obsessing over whether an image is authentic, Muller suggested that people should instead focus on interpreting intent, asking themselves what the poster is trying to communicate or influence.
It is precisely these concerns that inspired entrepreneur Christos Floros to build an alternative to the dominant foreign social media platforms. His Luxembourg-based start-up launched the first version of its platform, Monnett, in mid-October. The core idea is simple but radically different: users should be able to configure their own algorithm.
Floros explained that Monnett operates like a personalised control panel, comparable to a television settings menu, where each user can decide exactly what appears in their feed.
He explained that people can choose to see only posts from friends, only from accounts they follow, or restrict content geographically, whether to Luxembourg, Luxembourg City, or even specific regions. In his words, the aim is to give individuals the maximum possible control over what they consume online.
The ability to tailor algorithms is not only useful for private users but also for organisations seeking greater transparency and autonomy in how content is prioritised. The young start-up, currently an eight-person team, has ambitious goals: it intends to compete with the largest players in global tech.
Floros noted that this ambition is paired with a willingness to learn from American entrepreneurial culture, particularly the value of launching early and refining the product based on real user feedback. That, he said, is why Monnett was released before being fully complete: the team wanted to understand users’ actual priorities from the outset.
Monnett has already attracted around 10,000 registrations. A revised, improved version of the platform is expected in January. The start-up hopes to exceed one million users by the end of 2026.
