I’m exploring three bets on the future of human connection. Each one tackles a different stage of building relationships, and they all align with my mission of reducing the barrier to human connection
1. AI Wingman/Companion (Ember)
Problem: we cross paths with so many people that can transform our lives. Whether it’s a co-founder, a best friend, or a life partner. People want to meet other cool people around them, but that takes a lot of effort and knowledge on who to talk to. What if the whole process was effortless? Because you have an AI friend that really knows you and can proactively recommend you people to talk to as you go about your day.
Progress so far:
- Built Ember (heyember.me) with 100+ users at UCalgary and UWaterloo (uw.heyember.me)
- Current features:
- knows exactly the kind of person you’re looking to meet (e.g. finding someone from your class who also loves to play basketball after lectures)
- Knows all the clubs around campus and which one you should join
- find the perfect course to take based on your major and what you’ve already taken (e.g. Ember suggesting you a class you would enjoy or a GPA booster)
- still seeing if there’s a need for this
- Users chat and call Ember (it’s just a phone num like Boardy), and I’ve livestreamed the build process here. Next:
- Sharpen marketing to show how Ember fits naturally into daily life. Offline (UofT project) is a good example of simple, clear positioning.
- Explore on-device AI models for privacy.
- Discover how people are using Ember and see if they’re finding other ways to get value from it.
2. Bonding Tools (Personalized Linktree / Card Game)
Problem: First impressions rarely uncover shared ambitions or vulnerabilities, which is what usually enables the transformation from a casual encounter to a lasting bond.
Ideas I’m testing:
- Card game: Inspired by Fluster and Tim Ferriss’s “Coyote,” I’m considering a social game that surfaces deeper conversations. I have a mentor from Florida who built and launched several previous card games. He is interested in working together on it.
- Personalized linktree: A profile that only shows shared overlaps with others (countries visited, favorite music, TV shows, podcasts, movies, books, bucket list, goals, etc…). Powered by graph databases + GenAI, it not only reveals similarities between two people but also across your wider network. Demo video. Next:
- Build a paper prototype of the card game and play it with people I’m not too close to but I’d like to be
- Build a very simplified version of this app through a course that I’m taking (so killing two birds with one stone)
3. Autonomous Personal CRM but it’s not a CRM: Exo
Problem: Nobody wants to manage a CRM, but everyone wants the benefits of remembering key details from all conversations like:
- Where you met someone?
- When did you meet them?
- In what situation would you want to reach out again?
- What are some things that you can bond over next time you talk?
Progress so far:
- Built a prototype at Hack the North: 7degree.
- Tool records conversations only with consent (we came up with a cool way to achieve that during the hackathon), and extracts key info into a private knowledge base accessible by both users (uses the same graph database I shared in my demo video, which is what is used for Ember)
- You can go on a call or chat with an AI (let’s call it Exo) to help you reflect on these conversations. What did you learn about them? What did you learn about yourself? Which is something I explored last year with https://getexo.vercel.app/. I explain how I use it in here
- These insights will also be stored in our CRM automatically on top of what’s already there from the conversation itself
- There can also be a page for ourselves that shows things we learned about ourselves over time as we continue chatting with people and with the AI.
Differentiation: There are similar tools of both extremes. One of them focuses on virtual meeting note-takers like Granola and Circleback. Two of many start-ups building in this space.
On the other hand, there are wearables like omi.me, friend.com, or even the meta glasses, which is in some way only an AI companion but with no other value added. But if we add Ember’s ability to connect people, and/or Exo’s ability to help you effortlessly manage and deepen your relationships into something like omi.me… There’s something in there worth exploring for sure
I will be executing on this idea with 6-8 other people who are interested in building this with me (long story)
How These 3 Connect
- Ember = An AI companion that knows you and helps you find your people.
- Linktree/Card game = helps you bond with them.
- Autonomous CRM = Also kinda of an AI companion that helps you remember, reflect on, maintain, and grow those relationships.
Together, they fit into the broader Orbit ecosystem: a set of tools designed to reduce loneliness and lower the barrier to real human connection.
I also make content on overcoming social anxiety since knowledge and being your own therapist is another key thing to fighting (that was how I overcame it).
The Chalant Society is also another thing I’m doing. It’s a university club striving to help people build social confidence through rejection therapy challenges while having fun & making friends around campus.
FYI, I am aware that sometimes knowledge is the solution to these problems and not technology. I’m also aware that sometimes the way society works might affect what kind of startup is actually possible in this realm. And I’m extremely aware of tarpit ideas as I mention in this LinkedIn post
I feel like I’m sitting in a goldmine of insights, and the only thing that’s stopping me is time and resources to validate them and bring them to life. Would be nice to also have a co-founder who’s just as ambitious as me in solving the good old problem of loneliness, and more specifically reducing the barrier to real human connection
Even investors are starting to validate this space. Amber Atherton (Patron Fund) is actively backing companies in two of the three directions I’m exploring. That gives me conviction I’m not chasing tarpits.