Hot take: what if Google Scholar reported two new metrics: (1) median citations per paper and (2) *percent* of papers with 100+ citations? I computed these metrics for some ~200 senior AI researchers: see https://
bit.ly/google-scholar
-metrics
…. The top researchers by median citations per paper
@_jasonwei
-
Google Scholar new metrics for AI researchers citation analysis
By
–
-
Hardcore AI Followers Challenge: Understanding a 9/9 Reference
By
–
only hardcore ai followers will get 9/9
-
Jason Wei finds the movie Her convincing
By
–
I definitely can, I found the movie "Her" pretty convincing 🙂
-
Character.ai Users Spend Over 2 Hours Daily
By
–
"Active users spend on average over 2 hours daily interacting with our AI." (
https://
blog.character.ai/character-ai/) -
AI Products Strategy: Beyond Productivity Assistants
By
–
I've been thinking about AI products recently, and while the major labs (OpenAI, GDM, Anthropic) are competing on productivity assistants, character is clearly going in a different direction. Thinking from first principles, human beings have a few fundamental desires: not being
-
Instruction Tuning Risks When Models Lack Training Data Knowledge
By
–
Glad to see this work on alignment; an interesting takeaway for me here is that instruction tuning on data that the model doesn't know can be really, really bad. Basically when you finetune on data the model doesn't know, in addition to teaching it that particular training
-
Les vraies compétences clés selon Jason Wei
By
–
A somewhat amusing personal revelation: a friend recently asked me what were the best skills I had. I think he expected me to say something like "prompt engineering", "writing papers", or "making evals." But my answer surprised him a bit: I said prioritization and communication.
-
Emergent Abilities in LLMs: Responding to LeCun’s Critique
By
–
Yann LeCun is obviously a legend but I found this tweet to be quite misinformed. The whole point of "emergent abilities" such as few-shot prompting and chain-of-thought prompting, is that we clearly *did not* explicitly train or fine-tune them into the model. These abilities
-
List of Names with Food-Based Puns and Wordplay
By
–
sam almond
demis wassabi
greg broccoliman
oreo vinyals
guac le
baguette zoph
liam feta
ilya sushikever
yoshua bento -
Jason Wei Opens AMA Document on AI Research Questions
By
–
In the past weeks I received many questions (from undergrads especially) about AI research, so I'm putting together a "Ask Me Anything" doc. Add any questions to the doc, I'll answer all of them: https://
docs.google.com/document/d/1QR
EmdzLwJ0CR3kdFeenJbBowT1IFFREd46y10tW6pog/edit?usp=sharing
… Yes, I'll actually answer them all, because writing answers