Building the Strongest possible Social Graph.
Dabble privately stacks a social network protocol to any system.
Companies can now participate in and add to a collective social graph.
DOORDAsh
Would restaurant ratings matter if I knew my best friend ordered from that thai place 12 times last month ?
AirBNB
Would trust ever be an issue, when the only people staying are friends of friends?
Mastercard
What happens when you pool social graph data from Doordash and Airbnb, to create Group Credit and Shared Loyalty ?
Each layer richens the collective social graph.
So how does one actually start ?
How do you accurately measure social graph distances to begin with ?
You bootstrap it off a whole new social network : Dabble
Owning the Viewing Experience
Superscrolling through Time
Log View, Location Sort
Facial Recognition
Fancy Log View, Location Search
Image Upload Speed : 100k images in 200 milliseconds
Beating Apple, google and MEta at SOTA FAcial Recognition
Elo Based Consensus Polling.
The first application of the Social Network Protocol (SNP).
-
Distributed facial recognition with retroactive elo score (individual model ratings) updates to improve iteratively.
-
We instantly sync metadata using image hash sets, accounting for compression across all platforms.
-
Anonymous polling using descriptive vectors, encrypted using shared matrices.
Owning the Sharing Experience
Ambient SHaring
Photo sharing made pre-cognitive using Facial Recognition.
Automate sharing with friends and family.
Predictive Sharing
Preemptive sharing, optimised with location data.
Share speeds on Airdrop : 20+ seconds for 100mb, with high failure rates.
Share speeds on Dabble : Zero Seconds, Zero Failure.
We pre-send hidden, encrypted, ephemeral files before the send command, and delete them in arbitrary time.
Heartbeat Protocol
Privacy preserving protocol for Predictive Sharing.
Owning the Capturing Experience
Pods
Creating the infrastructure for hyper casual, hyper contextual, short form video.
Messaging
Implementing the interface to collate multi-threaded conversations.
Owning the Interaction Layer
Dabbles
A 1-few system directed to your ‘n’ Friends.
Stamps
A 1-some system disseminating to your ‘n+1’ Friends.
Stacking
The ability to fluidly convert 1-some systems to 1-few systems and back.
The one thing Dabble does better than anything else in the world, is its ability to convert n+1 friends to n friends, butchering all social costs.
And the only way to do it is by making 1-few and 1-some systems truly interchangeable.
Never do We want anyone to think twice before openly sharing something they want.
Team :
Shivang Raisurana
I design things.
Aditya Jain
IOI Coach of India, 2x ICPC World Finalist, International Grandmaster.
Adit Srivastava
IIT Bombay, International Quant, DaVinci
People building Dabble include 9 Quants from Jane Street and Optiver, including 3 MIT CS, and 7 IIT Bombay CS.
We’re actively looking for more. Currently at Stanford.