I made a raft out of longleaf pine hope its strong enough to carry me to the Karman line

Dear Y-Combinator type CEOs, how have you leveraged personality, ethnographic or demographic data to select for founders who are most likely to succeed in their respective markets?

Are you overfitting today? Are you relying too heavily on deepfakes, patterns and yesterday's data? Do you need to improve your generalization ability and robustness to out-of-distribution data? #introductions

I am searching for the future of work. I am looking for a company that isn't afraid to act like a startup when testing new ideas. I am looking for a leader who is obsessed with his customers. #introductions

The availability of huge data sets makes it harder to study private info without violating privacy. To ensure privacy while training with distributed data use Federated Learning. Your raw data never leaves the device, data stays local. Think of it as a decentralized approach to machine learning when multiple devices or servers are being used to collaboratively train a model.

My machine is trying to learn hard things. I wonder if my fans will survive this round of heavy batch processing. Stay tuned!

I am very proud to be a friend of the #SFPL library. Whats your ISBN-10 tho?

The best copyeditors read the phages.

I made it to the party. #introductions