Don’t miss it: Q&A livestream tomorrow (Thursday 23rd) 11:30am PT with Stephen, Seth, and JP.
The last week has been eventful for all of us. We launched, and got a ton of media coverage, Youtube videos, lots of people joining Discord (400+), and lots of people registering to get the data (300+). Welcome everyone!
Over the weekend, two wise and generous souls offered to contribute $50k each to the prize pool. When we announced their pledges on Twitter Monday morning, we were not prepared for what happened next.
Over the course of the day, more than a dozen other people reached out to offer their support for the prize pool. By the time the dust had settled, the grand prize had grown from $150,000 to $700,000, and the overall prize pool stood at ONE MILLION DOLLARS.
Nat Friedman@natfriedman
Nat Friedman @natfriedman
🥰 Thanks to @levie for his $25k pledge! Clearly a man passionate about data storage and recovery. Grand prize: $655k, total pool $755k. Stay tuned – we're not done yet.... https://t.co/5RrK9eDRX4
12:49 AM · Mar 21, 2023
44 Reposts · 282 Likes
We are in awe of this support. Sincere thanks to all our generous new sponsors: Joseph Jacks, John Collison, Patrick Collison, Tobi Lutke, Guillermo Rauch, Arthur Breitman, Matt Huang, Julia DeWahl, Dan Romero, Anonymous, Bastian Lehmann, Aaron Levie, Ivan Zhao, Stephanie Sher, Amjad Masad, and Conor White-Sullivan.
The competition itself has been off to a good start too:
Almost 100 contestants have submitted on Kaggle.
Great discussions on Discord, about papyrus fibers, X-ray imaging techniques, what signals ML models are picking up on, and crowdsourcing segmentation.
Daniel, Brent, and Nat had a fun discussion, in which they dive into the details of the competition.
Contestants are already combing through the data, with _kai_ discovering interesting artifacts in the surface volume data, which might point to better ways to sample the surface volumes.
Francois Chollet (famous for creating Keras and contributing to TensorFlow) created an awesome starter notebook on Kaggle.
Lots more good stuff on Kaggle, such as relating ink detection to other domains, and a fun animation/notebook of layers.
Cineris posted an early version of a UI tool for data processing on slices.
And as mentioned at the top, tomorrow (Thursday 23rd) 11:30am we’ll have a Q&A livestream with Stephen, Seth, and JP. Be sure to join us for some heated debates about volcanoes, scrolls, particle accelerators, and more!
If you haven’t already, be sure to stay in touch through all the channels: Discord, Kaggle, Twitter, and of course this Substack.