Elle Yuan Wang, Ph.D. (@ellewy) profile picture

Elle Yuan Wang, Ph.D.

@ellewy

Tech Start-Up Advisor & Investor | AI & Learning Scientist | Previously @Columbia @Columbia_Tech @MTV

elleyuanwang.com
Joined March 2009
  • Tweets1.5K
  • Following263
  • Followers652
102Media
Photo by Elle Yuan Wang, Ph.D. (@ellewy)Photo by Elle Yuan Wang, Ph.D. (@ellewy)Photo by Elle Yuan Wang, Ph.D. (@ellewy)Photo by Elle Yuan Wang, Ph.D. (@ellewy)Photo by Elle Yuan Wang, Ph.D. (@ellewy)Photo by Elle Yuan Wang, Ph.D. (@ellewy)Photo by Elle Yuan Wang, Ph.D. (@ellewy)Photo by Elle Yuan Wang, Ph.D. (@ellewy)Photo by Elle Yuan Wang, Ph.D. (@ellewy)Photo by Elle Yuan Wang, Ph.D. (@ellewy)
I reviewed for NSF SBIR panels. The same fixable writing mistakes sank strong proposals over and over. Grantproof — a free Claude skill that flags them, quoted from your draft: github.com/ellewangyuan/gran…
Image from Elle Yuan Wang, Ph.D. (@ellewy)
1
123
more replies
Narrow by design — not a full editor. One class of writing-style mistakes that consistently cost proposals. Not content, not structure, not research design. npx skills add ellewangyuan/grantproof --global github.com/ellewangyuan/gran…
GitHub - ellewangyuan/grantproof: Paste your federal R&D grant draft, get back the writing-pattern...

GitHub - ellewangyuan/grantproof: Paste your federal R&D grant draft, get back the writing-pattern...

Paste your federal R&D grant draft, get back the writing-pattern fixes reviewers expect. A free Claude Skill. - ellewangyuan/grantproof

github.com
1
16
First public version. Want a heads-up on future skills + grant tools? I send occasional updates: elleyuanwang.com/sbir-pitch
Dr. Elle Yuan Wang - AI & Education Expert

Dr. Elle Yuan Wang - AI & Education Expert

Leading expert in AI and education. Columbia PhD, AI & Learning Science Advisor.

elleyuanwang.com
8
If you have used #chatgpt for a while, a fun question to ask it “based what you know about me, what is my MBTI profile?” The result can be revealing as much as amusing. :-)
57
Elle Yuan Wang, Ph.D. retweeted
Quincy Larson (@ossia) profile picture
Quincy Larson
@ossia26 Feb 2025
If you want to start a charity of your own, I sat down with @ellewy to answer her questions about the practical realities of running a charity. We also talked about the future of education and the role LLMs are playing. 90 minute watch / listen. youtube.com/watch?v=5ETnXlm1…
Image from Quincy Larson (@ossia)
2
6
20
2,751
“Mimicking the herd invites regression to the mean (average performance).” - #CharlieMunger
1
162
I do see more coders transition to Claude these days.
Image from Elle Yuan Wang, Ph.D. (@ellewy)
Andrew Ng (@AndrewYNg) profile pictureAndrew Ng
@AndrewYNg6 Jul 2024
Shoutout to the team that built artificialanalysis.ai/ . Really neat site that benchmarks the speed of different LLM API providers to help developers pick which models to use. This nicely complements the LMSYS Chatbot Arena, Hugging Face open LLM leaderboards and Stanford's HELM that focus more on the quality of the outputs. I hope benchmarks like this encourage more providers to work on fast token generation, which is critical for agentic workflows!
AI Model & API Providers Analysis | Artificial Analysis

AI Model & API Providers Analysis | Artificial Analysis

Comparison and analysis of AI models and API hosting providers. Independent benchmarks across key performance metrics including quality, price, output speed & latency.

artificialanalysis.ai
323
Completed #rwri 19! Lots to digest, reflect, and apply. Concepts helpful to understand tail risks and complex systems: - Entropy - Cellular Automaton - Pareto distribution It's not about precision into risk estimation. Rather "how NOT to get surprised by loss". - @nntaleb
1
16
2,590
Attending #RWRI seminars is a privilege, made possible by the latent privilege of being able to attend from '8 AM to 12 PM on weekdays'. #Gratitude
4
424
Interesting since I first picked up “Antifragility” by @nntaleb via the “Phoenix” public library 4 years ago, of all places I could get a book. Didn’t realize the connection until the 2nd day of #rwri #rwri19
5
1,211
Started reading “A Man for All Markets” by @EdwardOThorp waking up at 2:30 am at age 11 to work on top of school… in awe
Image from Elle Yuan Wang, Ph.D. (@ellewy)
1
299
Human beings ingest tons of training data every day. Some of us use unsupervised models; some use supervised ones. #ai #ThoughtForTheDay
1
214
Impressive!
Quincy Larson (@ossia) profile pictureQuincy Larson
@ossia18 Dec 2023
On Wednesday I'm going to announce several new FREE interactive learning resources that the freeCodeCamp community has been building all year. 🎄🏕️
1
295
Sounds like great ideas for #GPTs 🤔
Adam Grant (@AdamMGrant) profile pictureAdam Grant
@AdamMGrant9 Dec 2023
How to keep an open mind: 1. Think like a scientist: treat your opinions as hypotheses and decisions as experiments 2. Embrace confident humility: argue like you’re right, listen like you’re wrong 3. Build a challenge network: seek out people who sharpen your reasoning
Image from Adam Grant (@AdamMGrant)
1
248
One of those “so simple but so difficult” hacks
Naval (@naval) profile pictureNaval
@naval30 Nov 2023
Living within walking distance of your friends completely transforms your quality of life.
136
Interesting cutoff for optimal correctness. Does #GPT4 “working memory” turn into long term memory?
Ethan Mollick (@emollick) profile pictureEthan Mollick
@emollick9 Nov 2023
You can now load the equivalent of 205 single spaced pages into the “working memory” of GPT-4. If you load less than 117 pages, the system can flawlessly recall details (at least in this test), after that, it has better recall of the very start and end of the text than middle.
136
Custom #gpts are like personal and shareable prompt libraries?
1
1
150
- The best coding prompts are written by expert coders, not non-coders. - The best essay prompts are written by expert writers, not lousy writers. - The user’s expertise level matters. #chatgpt #PromptEngineering
116
Load more