R2L Lab will be at #ICLR2026! Check out Arthur Chen's grounded adaptation of agents post deployment, and Sitao Cheng's study of compositional generalization through RL! I'll also be there, please email to meet.
arxiv.org/abs/2511.04847
arxiv.org/abs/2512.01970
Also - I remember back in the day when we spent days writing harnesses for text2sql. Now you can just ask for "a text2sql agent" and AI will make the harness in < 1 minute. What a crazy time to be a PhD student - you can prototype so many ideals so much faster!
I spent the last week going ham on coding agents. I feel like they are much more powerful than using web apps because CLI agents are much more customizable and can spawn subprocesses.
They are better - just look at the Gemini privacy statements for Google Workspace vs. individuals
Easier to anthropomorphize a “Claude” than a corporation
I will be at the IVADO agents workshop followed by COLM in Montreal. Please email me if you would like to meet. R2L Lab will be looking for 1-2 graduate students in 2026!
Now that school is starting for lots of folks, it's time for a new release of Speech and Language Processing! Jim and I added all sorts of material for the August 2025 release! With slides to match! Check it out here: web.stanford.edu/~jurafsky/sl...
We would like to extend our warmest thanks to the organizing committee @sivareddyg.bsky.social, Dawn Song, @yoshuabengio.bsky.social, Alane Suhr, @chrisjpal.bsky.social, @yusu-agent.bsky.social and @vzhong.bsky.social.
Do you have a PhD (or equivalent) or will have one in the coming months (i.e. 2-3 months away from graduating)? Do you want to help build open-ended agents that help humans do humans things better, rather than replace them? We're hiring 1-2 Research Scientists! Check the 🧵👇
I think in retrospect ARR might have been a bad idea. Am I hallucinating or is there an exodus of junior NLP researchers from *ACL into ML conferences and COLM?
Are you an established professor in AI open to moving to the University of Waterloo? Please DM me directly!
Flying out of DCA for the first time and I gotta say - this has ought to be one of the worst airports in North America… so cramped people are literally sitting on the ground or standing shoulder to shoulder
This is how digital feudalism works.
It offers the illusion of empowerment. Likes become validation. Followers feel like influence. Monetisation looks like success.
But the rules are hidden. The platform decides what rises and what vanishes.
Is Best-of-N really the best we can do for language model inference?
New paper (appearing at ICML) led by the amazing Audrey Huang (ahahaudrey.bsky.social) with Adam Block, Qinghua Liu, Nan Jiang, and Akshay Krishnamurthy (akshaykr.bsky.social).
1/11
Incredible lineup of speakers at #CUA #WORKSHOP @ #ICML2025!
Speakers part 3/3:
Qingyun Wu (Penn State University) @qingyunwu.bsky.social
Zhiyong Wu (Shanghai AI Lab)
Victor Zhong (University of Waterloo, Vector Institute) @vzhong.bsky.social
Teaching for the first time: I finished the last official lecture of the class and got a surprise round of applause! Feels so great!
A bit of a mess around the conflict of COLM with the ARR (and to lesser degree ICML) reviews release. We feel this is creating a lot of pressure and uncertainty. So, we are pushing our deadlines:
Abstracts due March 22 AoE (+48hr)
Full papers due March 28 AoE (+24hr)
Plz RT 🙏
We've kept pushing our Open-R1 project, an open initiative to replicate and extend the techniques behind DeepSeek-R1
And even we were mind-blown by the results we got with this latest model we're releasing: ⚡️OlympicCoder
[1/3]
Jesus…
#InternationalWomensDay is a time to promote women’s equity by celebrating women’s achievements, raising awareness about discrimination and taking action to forge gender parity. Meet some of the exceptional women in Science who are helping to #AccelerateAction: bit.ly/3XCAft9 @uwaterloo.ca
Canada being the best at hockey and the best at everything and you can never defeat us and CANADA! CANADA! CANADA!
Report: CANADA! CANADA! CANADA! CANADA! CANADA! CANADA!
The democratization and commoditization of foundation models is happening and will continue to happen - regulation will not slow nor stop it. Secrecy, propriety moats, and trade embargoes may succeed briefly in AI, but they will fail. And when they do, they will likely fail hard.
In the photo is Cheriton School of Computer Science Professor Martin Karsten in the Davis Centre.
📣 Big news for sustainability in tech! 🔋
CS Professor Martin Karsten and his collaborators Peter Cai and Joe Damato have developed a small modification to the Linux kernel — now part of Linux 6.13 — that could cut energy use in data centres by up to 30%.
cs.uwaterloo.ca/news/cherito...
Squid games but NeurIPS acceptance
#AI hallucinations are a problem; #UWAllen Ph.D. student @akariasai.bsky.social may have the answer. She was named a @techreviewjp.bsky.social Innovator Under 35 for her work to make #LLMs more transparent and useful—without making stuff up. #IU35 #AIforGood
news.cs.washington.edu/2025/01/07/w...
Coincidentally when I did research during undergrad, my advisor was also a baby faced fresh graduate. Students used to wander into the lab and ask him where the professor was.
Sadly, we have lost a brilliant researcher and colleague, Felix Hill. Please see this note, where I have tried to compile some of his writings: docs.google.com/document/d/1...
On the way back, I was asked if I “go to Waterloo” by undergraduate student. I hesitantly replied yes and eventually had to explain I was a new prof. Student refused to believe me - said I looked like an undergrad 😂