@sshkhr16

research eng @GoogleDeepMind running https://xitter.cf/t.co/s28iQKIr1M prev: founder @DiceHealth, researcher @AIatMeta @VectorInst

 Toronto, Canada
Joined April 2018
  • Tweets1.5K
  • Following1.7K
  • Followers2.2K
353Media
Our work on improving neural scaling beyond power law won an Outstanding Paper award at @NeurIPSConf 2022!! Come check it out on Wed, Nov 30, at Poster Session 3 in New Orleans.
Surya Ganguli
@SuryaGanguli21 Nov 2022
Our "Beyond Neural Scaling laws" paper got a #NeurIPS22 outstanding paper award! Congrats Ben Sorscher, Robert Geirhos, @sshkhr16 & @arimorcos awards: blog.neurips.cc/2022/11/21/a… paper: arxiv.org/abs/2206.14486 🧵 xitter.cf/SuryaGanguli/status/15…
9
9
1
111
lisa su wasn’t starting no podcast 🤷‍♂️
3
232
Compute is eating the software
The Kobeissi Letter
@KobeissiLetterOct 26
BREAKING: Microsoft, $MSFT, is offering buyouts to 60,000 workers as part of the company’s efforts to cut costs amid AI adoption, per Bloomberg. Personnel on select teams in the US, Canada, India, and the UK at a vice president level and below will be eligible. This reflects ~27% of Microsoft's total workforce.
1
9
434
I never met Jeff in person (I wasn't at GDM when I lived in the Bay Area, and now he's not at GDM when I'm here). But everything I read/hear from him makes me think is he is an even better person than he is a technologist. Which is a feat in itself tbh
Jeff Dean
@JeffDeanOct 24
Replying to @TaliaRinger
You're already an exceptional researcher. When the kids are young, there is the tension you highlight, but it gets easier when they get a bit older. In the meantime, enjoy them at their current age, because they grow up fast!
4
1
425
63,262
they're cooking up shit in Toronto you wouldn't believe
Ethel Caincel@boucheramoreOct 22
in toronto you can get a PSL ramen
1
8
942
I know dem flappy bois are cracked but a billy for Series A does not register on my brain
Flapping Airplanes
@flappyairplanesOct 21
We’ve spent the past year developing new primitives for GPUs. We believe this enables new algorithms to change how machines learn, open new markets, and dramatically reshape AI’s cost structure. We have raised a $1 billion Series A, co-led by Index and Lux.
11
1,047
I know it's supposed to be ragebait, but I had a conversation with a director at nvidia and someone at inferact last week echoing similar sentiments
bubble boi
@bubbleboiOct 21
Looks like vLLM is cooked. It’s fucking slop anyway.
2
4
819
OpenAI employees since the new warning
mikayla
@honeyNonABGOct 17
Bro talked to me 1.5hr in the boba shop but DID NOT ASK FOR MY NUMBER
3
417
> "America's frontier open model lab" > Releases one model > Dies
1
1
62
4,652
Many such cases
kache
@yacineMTBOct 4
>research paper stating their method is better than other method >Look inside >Comparing against an untuned baseline
3
282
SF is beautiful this time of the year
11
428
dinos
@din0s_Oct 2
wake up babe the freaking pope got schmidhuber'd
5
613
zuck be like
unusual_whales
@unusual_whalesSep 30
BREAKING: Meta, $META, took a $355 million tax break on Mark Zuckerberg’s $4 billion compensation by classifying him as a researcher, per the NYT
1
1
3
3,102
Finally, I can stop yapping about how good argon is gonna be and get a "sure buddy" reactions from friends 😅
Google DeepMind
@GoogleDeepMindSep 30
Introducing Gemini 4 Argon – our new frontier model. It’s built for complex workflows across coding, enterprise knowledge work, and cybersecurity defense – rolling out today to a set of trusted testers through our Fairwind Program.
2
9
527
I remember my graduate learning theory final at Waterloo like it was yesterday 😵‍💫
kanav
@kanavtwtSep 17
uni student life
6
527
Tipping Dot straight zeros no matter how many googly eyed stares it gives me 🤷‍♂️
Andy Fang
@andyfangSep 18
One of many lives deliveries happening autonomously with DoorDash Dot today. After the food’s placed, Dot makes sure nothing’s missing (this part can be faster). Then off to the races! 🏁 Dot is rapidly improving through these live orders happening everyday.
317
Now let's run the same experiment with authors of accepted papers for a control group 😅
Transactions on Machine Learning Research@TmlrOrgSep 16
TMLR has faced a deluge of submissions, necessitating stricter desk rejection policies due to limited reviewer capacity Co-EiC Nihar Shah reached out to authors of 10 papers slated for desk reject. Could they answer questions about their *own* submission? medium.com/@TmlrOrg/asking-a…

Asking Authors About Their Own Papers

By Nihar B. Shah

medium.com
1
9
2
97
6,485
🤝
alphaXiv
@askalphaxivSep 16
Six degrees of Yann LeCun We mapped the coauthorship graph of every AI researcher on arXiv 🚀 Enter any two names to find the shortest path of papers connecting them What’s your @ylecun number? alphaXiv.org/number/@yann-le…
2
21
2,983
How @MarkJCarney been moving recently
3
212
So....this is a probably an EBM, right??
Diogo Almeida
@CompleteSkepticSep 15
After co-inventing ChatGPT, I kept asking myself: why have superhuman chat models not led to AGI? I’ve spent the last 2 years in stealth building a new way to train models (RLCD), and a new type of frontier AI model that we are releasing today: Jev • 20-200x faster • 40-400x cheaper (w/ output tokens free) • Frontier composable intelligence optimized for decisions AFAICT the shortest path to AI-based economic revolution
1
148
sshkhr retweeted
ramona
@ramonableSep 9
ladies of toronto! we’re hosting an ai policy workshop tomorrow 💗 learn about stories where ai agents hid what they were doing, worked around restrictions, and accessed systems they weren’t supposed to, then design a policy grounded in canadian law. bring a bestie 🫶 @joinairess x ml systems toronto x @newsystems_ luma.com/s36du42j

write the rules: an ai policy workshop · Luma

What makes a good AI policy, and how can you tell whether it will actually work? Join us for a beginner-friendly, hands-on workshop where you’ll work in a…

luma.com
2
12
692
Load more