cyyeh

↗ HN profile [ 42.0 ms ] full profile
Karma
0
Created
()
Submissions
0
  1. One kind of agent is enough for a team: all of one kind, or any mix of Claude Code, Codex, Grok, Cursor Agent, Antigravity, GitHub Copilot CLI, Muse Code and five open-source coding agents on your own model backends.…

  2. micrograd(from Andrej Karpathy) implemented in Rust with Python bindings via PyO3 using Cursor agent mode

  3. VulcanSQL and LLMs (twitter.com)
  4. I just thought that it would be an effective learning method to learn how to effectively train neural networks using case study approaches. Are there any resources like that? Thanks!