In the process of building the Luna dating app, we realized that the algorithm we were trying to build was a much bigger project than we originally thought. For years, dating apps have continued to modify and improve an inherently flawed theory of matching. The only innovations in this landscape have been changes to how you engage with the app, either reducing you to a photo, artificially limiting how often you can message, or limiting who is allowed to message first. Each of these techniques underscores a gaping problem in the current dating app marketplace: they assume you are the problem, rather than the app.
Here at Luna, we know that’s not true. In fact, it would be arrogant to assume users don’t have the capacity to make smart dating decisions. The issue has always been the limitations of the matching algorithms themselves. Building yet another matching algorithm would be relatively straightforward, but the issue was that they don’t work.
This is also true in lots of sectors, from the job-hunting world to connecting users with services. Despite automated resume-matching tools and online job search engines, most jobs are still found the old fashioned way: through networking and personal connections. Investors, whether traditional or crypto, still rely on old-school insights and recommendations when choosing who to back.
In each of these sectors, matching apps might work if people were only as complex as a deck of cards. We realized that all these use cases share one missing component. They require something more sophisticated than a pure matching algorithm, something that mimics the sensitivity and insight of human connection, complete with its own matching syntax and with a capacity to learn and self-correct.
Introducing the Luna Connection Engine. The connection engine will power the Luna dating app, but will also power a robust ecosystem of matching apps meant to replicate our natural human capacity for finding meaning and synergy in each other.
This engine is unlike anything that currently exists. It will utilize a generative grammar that can match subjects and objects using multiple different verb functions. By creating topic categories for various different subjects/objects, by defining verb functionality, and by allowing for variations in subject-verb-object syntax, you have a machine that operates the way the human mind thinks. When you set someone up on a blind date, it’s as simple as thinking “Jordan might like Taylor.” When you recommend someone for a job, it’s “Taylor could work for Sarah.” After any human-initiated connection, you process and integrate feedback: turns out Jordan hated Taylor but Sarah loved Taylor. With a generative grammar model, you have the capacity to differentiate between these two different match cases while contributing useful information to both circumstances.
This generative grammar also has potential applications beyond peer-to-peer connecting. Eventually, a strong connection engine will be able to compile information from restaurants and tourist attractions to tell you what you’ll like based on what your peers might have previously recommended. With a generative grammar model, you have an infinite number of possible subject/object classes, and an infinite number of potential verb functions, meaning it can power any circumstance in which self-learning and highly-adjustable matching is of value.
The Engine’s potential is huge. The Luna dating app won’t need to be limited by the old language of dating: “Who fits my initial criteria?” It can be as varied as “Who has the same values as me” to “Who might be a good artistic collaborator” to “Who can show me around this city?” We also want to continue our explorations into creating connection solutions for conferences and corporate entities. We hope to eventually expand into creating enterprise solutions to connect resources to those who need them most.
The problem facing us today is a false scarcity caused by an overabundance. The job market is saturated with qualified employees, dating apps are filled with interesting people, there are resources being wasted that never find those who need them. The Connection Engine is not about merely finding people; it’s about finding the right people, reporting their compatibility, and using the information to make better connections in the future.