wey_lynn_com_nebula_linkedin
To: Wey (Siwei) Gu & Lynn Bender
Dear Lynn and Wey,
I hope this message finds you in good spirits. As recent additions to my LinkedIn network, I’ve admired your contributions and wondered if we could embark on an exciting collaboration.
Motivation:
While preparing to share a recent article on RAM-based GNNs , I realized its relevance to my previous posts. The current LinkedIn design, however, doesn't facilitate easy referencing of such connected content.
Context:
LinkedIn, in its current state, encourages fast-paced, buzzword-heavy content, often neglecting the depth and interconnectedness of knowledge. Rather than building on previous insights or referencing established knowledge (standing on the shoulders of giants), content risks becoming ephemeral and redundant.
Solution:
It's time for platforms like LinkedIn, and potentially others like Facebook, Instagram, or Twitter, to evolve. Imagine a version-controlled system, akin to Git, manifesting as a knowledge graph on LinkedIn.
Proof:
By harnessing graph theory, we can utilize hashtags as edges and topic embeddings as nodes. The post's content can serve as neighboring nodes, creating a knowledge-rich network. This approach facilitates better connectivity between topics and in-depth discussions.
Approach:
I've created a video demonstration using the Nebula Graph, which ingests my LinkedIn posts, rendering a visual knowledge network. The underlying engine powers this visualization.
Collaboration Request:
I am eager to explore this idea further with your expertise. I have a Python class ready that extracts user post content from the official LinkedIn API. However, I lack the partner status required to query member post content. Your association and insights could bridge this gap.
The prospect of reimagining LinkedIn's UI and algorithm with your collaboration excites me. Together, we can bring depth, interconnectivity, and a renewed sense of purpose to content sharing.
I await your thoughts.
Warm regards,
Kilian
Video: https://youtu.be/2qz3npFvTts
