Yixiong Hao

Projects

Writing

Resources

Ideas

Resume

A dump of links that I think could be useful to value-interest aligned friends.

Last updated June 2025

General

Atlas Fellowship – Resource Gallery

https://airtable.com/app53PsYpHxJW61l3/shrQSYXSW9z96y5WE/tbls2B9CeeFfavLXo - List of opportunities in high impact fields

https://docs.google.com/spreadsheets/d/1JTNfDeRXZKR5hEF3WYIBpi2Wu8QAeghFwJvJ53DrJWo/edit#gid=0 - for whatever you’re building.

AI alignment ඞ

Opportunities and Resources - List of opportunities to participate in entry level AI safety research!

Opportunities to work on alignment research

https://aisafetyfundamentals.com/ - introductory course

ARENA: technical upskilling for alignment researchers

https://arena3-chapter0-fundamentals.streamlit.app/[0.0]_Prerequisites#pytorch - ARENA ch0

https://arena3-chapter1-transformer-interp.streamlit.app/ - ARENA ch1

https://arena3-chapter2-rl.streamlit.app/ - ARENA ch2

https://arena-ch3-training-at-scale.streamlit.app/ - ARENA ch3

https://www.youtube.com/c/robertmilesai - conceptual arguments for why care about AI safety, from the goat that got me into alignment

https://arc.net/folder/D0472A20-9C20-4D3F-B145-D2865C0A9FEE - Sutskever’s list of relevant ML papers in the current paradigm

Tech

My personal evaluation of the LLM based products out there:

• ChatGPT o3 & Gemini 2.5 Pro: best for non trivial general queries
• Claude family: by far the most interesting artifact for studying LLM behaviors.
• Perplexity: best for fast, fact based search. Way better than google search or their AI mode thing.