Thesis
Why we are building Thesio.
Networking has a volume problem.
Students send hundreds of cold messages to land a handful of replies. The few coffee chats they get rarely go anywhere. Analysts and associates sit on the other side of the same loop, buried under low-fit inbound from people they cannot meaningfully evaluate. Both sides are exhausted, and the platforms that own the space are built for volume, not fit.
Every other category online has been rebuilt around prediction. Music knows what you'll like before you do. Streaming knows which show to put in front of you. Even dating, the messiest signal problem of them all, has moved past keyword filters into something closer to inference. Networking has not. It's still a directory. You search, you scroll, you guess.
The right professional connections are predictable, and the signal to predict them already exists.
What makes two people click is the long tail of niche signal: the books they return to, the side projects they sink time into, the ideas they cannot stop thinking about, the moment in their career they are actually in. Job titles and schools are anchors, not answers. When the right signals align between two people, the probability of a real connection is high. When they don't, no amount of mutual industry will fix it.
This signal is everywhere. Every connected account, every behavioral trace, every piece of taste a user is willing to share is a high-density input. One user with the right inputs produces more matching signal than a hundred resumes ever could. The platforms built for volume treat this signal as exhaust. Thesio reads it.
This is not a filtering problem. It is a prediction problem. Filters narrow a directory. Prediction reads the signal and points at the answer. The shift sounds small. It is not. It is the difference between a search engine and a recommendation engine, and it is the same gap every other category crossed a decade ago.
Thesio is a prediction engine for professional connection.
It reads signal layers across every profile: industry, role, motivation, behavioral patterns, timing and network proximity probability. It then translates them into potential matches. It surfaces the people who most likely form genuine and beneficial connections.
We are building Thesio because the most important relationships in my own career happened by accident, and I refuse to accept that the infrastructure to make them findable does not exist. Every prior challenger to LinkedIn has died on the same hill: trying to rebuild the directory. Bumble Bizz, Shapr, Lunchclub, Polywork. The graveyard is well-marked. Thesio does not compete on the directory. It competes on the prediction layer the directory cannot produce.
Thesio will not become a directory. It will not optimize for volume of connections, time spent in feed, or messages sent. It will optimize for the only metric that matters in this category: did the connection it predicted actually become real, and did it shape something.
The matching engine is consent-driven by design. Every signal Thesio uses is approved by the user it belongs to. Nothing is inferred silently. Nothing is shared without sign-off. Trust is not a privacy disclosure at the bottom of the page, it is the substrate the prediction sits on.
This is what we are building. This is why now. This is the bet.