Ten years ago in 2015 we published a paper called End-to-End Memory Networks (arxiv.org/abs/1503.08895). Looking back, this paper had many of the ingredients of current LLMs. Our model was the first language model that completely replaced RNN with attention. It had dot-product

At what point do we acknowledge that this is a higher administration problem? Devote enough resources to the first year courses and make the culture stern that disincentivizes cheating.

The great thing about chairing a conference is that either: 1. You can handle/fix all the little details that bother you for other conferences; or 2. You can learn how hard it is to chair and get every little detail right. (Probably both)

Massive gains with Sonnet 4 on SWE-agent: Single-attempt pass@1 rises to 69% on SWE-bench Verified! Sonnet 4 iterates longer (making it slightly more expensive) but almost never gets stuck. Localization ability appears unchanged, but quality of edits improves.

Congratulations to SACSS and Sudha Archarya on 25 years of incredible service! The gala was filled with so much joy and so many great friends. Love Light Laughter Song: Bhoopali from Soul Call #sacssny #southasian #helpinghands #gala #charity #nyc #lovelightlaughter

Claude 4 just refactored my entire codebase in one call. 25 tool invocations. 3,000+ new lines. 12 brand new files. It modularized everything. Broke up monoliths. Cleaned up spaghetti. None of it worked. But boy was it beautiful.

Yann LeCun is not so optimistic about LLMs. 😟 "We are not going to get to human level AI by just scaling up MLMs. This is just not going to happen. There's no way. Okay, absolutely no way. And whatever you can hear from some of my uh more adventurous colleagues, it's not going

🤯 Our new Google DeepMind paper reveals a vulnerability in the AI supply chain. Our paper, "Cascading Adversarial Bias," shows how tiny, malicious changes to a large "teacher" language model can create amplified biases in smaller "student" models after distillation.

super cool work led by Sara Ghazanfari, smthg i found really interesting is the extent to which reasoning traces from synthetic videos transfer to real videos....

Our flight was delayed because of turtles on the runway at JFK. I looked it up and seems to happen every June and July as turtles cross the runway to lay eggs.

To be perfectly honest, when I taught undergrads for 4 years at umich, one of the finest universities in the country, I used to often think: these kids “don't actually reason at all. They just memorize patterns really well.”

Worked my way up through NYT, WaPO, New Yorker, Atlantic etc. and finally broke into IndiaToday 😎 indiatoday.in/technology/new…

In the broader discussion of AI revolutionizing healthcare, Associate Professor Rumi Chunara discussed the essential regulatory framework governing these technologies with Bloomberg Originals #NYUTandonMade buff.ly/qlhwyGg