as ac of ICLR (and other conferences), I’d rather reviewers opt in to disclose AI reviews and the models used (no penalty on that!), or claim human-reviewed if that’s true, so I can better weigh each review’s significance and merit in final recommendation stage!
364
CY Harry Yang retweeted
BREAKING: A mathematician at Anthropic claims to have used the company's AI to solve the Riemann hypothesis, the most notorious open problem in all of math
sciam.com/article/anthropic-…
When Figures Disagree with Text: Evaluating Autonomous Visual Evidence V... youtu.be/295BnHtEKK0?si=sOtl… via @YouTube
164
Today has been a distressing day for many of us.
I've seen Jensen a couple of times, at conferences and talks. Almost 10 years ago we all surrounded him at CVPR, back when he wasn't a superstar yet but could articulate the latest research from 10 academic papers better than most other oral talks.
When I interviewed at Nvidia, the word I heard most was Jensen. Jensen this, Jensen that. I felt the strong respect, even worship, for him inside the company. I don't remember people mentioning Zuck or Satya when I interviewed elsewhere.
Then he gradually rose to the top: the kingmaker, one of the most significant figures in the world. I saw Madison at conferences. The whole family seemed so nice and easygoing.
Sometimes I didn't agree with Jensen. On AI risk. On job displacement. He seemed overly optimistic to me. But I never assumed bad intentions. Brains are just wired differently. I still viewed him as a North Star and even followed all the restaurants he’s been to.
Then came the fallout. I'm not a conspiracy theorist, but this time things don't add up. Just like I know I will never mess around with my reports or clients, Jensen should know that much better than me. And all of this comes out at once.
Everyone can have different opinions about marriage, fidelity, and love. I strongly believe Lori is a wonderful lady, and I feel sorry for her. But why would Jensen risk everything? The empire he built, his family's careers, his daughter's, his son's, his son-in-law's, the family business. I don't think he saw this coming. In Emily's interview he was still confident and wanted to work as long as he could, without thinking about successor. Then in a week comes the revelation of multiple scandals, even though they all happened a long time ago. Even the ultimate fools will feel there is much more to this story.
Of course, Jensen will still be far wealthier and better off than almost all of us, and it's entirely possible we'll find a darker side of him as things unfold. But right now I see the most dramatic, sudden fall of one of the greatest figures of our time, and it's hard to comprehend.
Money is nothing without power. Leading the most successful company in the world, being revered and greeted by presidents and heads of state, being chased and worshiped by fans in every corner of the globe, being the Taylor Swift of tech: those are things money cannot buy. If it were me, I would feel deeply suffering no matter how much wealth I have left.
5
7
1,050
Recent news reminded me of a thought experiment I had:
Imagine you’re one of the greatest scientists of your generation. You deeply love your wife, with whom you’ve built a beautiful life.
Now imagine two parallel universes:
Universe 1: Your wife passes away when you’re both in your 80s. You then meet a woman in her 20s who is remarkably similar to your wife when she was young. She’s admired you for years, seeing you as her North Star, and eventually falls deeply in love with you. You marry, beginning an extraordinary second love story. She lovingly cares for you for the rest of your life, well past 100.
Universe 2: You and your wife both live well into your 100s, growing old together and caring for each other until the very end. You still get to meet the 20s girl who loves you deeply, but you cannot be with her due to loyalty to your wife as well as the irreparable reputational damage as a renowned scientist.
You get to choose which universe you live in.
Which would you choose?
285
Introducing VidDub: video dubbing in just 2 denoising steps.
The key idea: teach the lips to follow the new speech while keeping the person and scene consistent, then distill the model into two steps.
What the demos:
law1223.github.io/viddub-pro…
2
1
216
what to do when you are bored in meetings?
my answer: ask Claude to make games.
Over a two hour meeting I asked it to make something similar to the game Spell Brigade, and it came out quite nicely:
hyang.org/spell-legion/
what a time to be a gamer 😃
1
271
Glad to give a talk at UNIST today, hosted by Prof. Taehwan Kim: "From Video Generation to World Models."
Check out the slides: hyang.org/ppt/From_Video_Gen…
1
208
listened to this while driving.
fun anecdote.
Dario reveals the people who called his GPT-2 fears "pseudoscience" in 2019 are now all applying for jobs at AI labs
"I personally was super freaked out by [GPT-2]. I looked at it and I said, my God, you just scale this thing up another factor of a million and it'll be better than humans at everything. This is the defining economic and scientific moment of the modern age. We just don't know it yet. This is going to determine the fate of nations and economies and the future of humanity."
"Maybe 10% of the people around me had exactly the same reaction, and it was often the smartest ones. 90% of the people around me, because GPT-2's capabilities in any particular area were not that impressive, it could translate languages, summarize a document and write a few lines of code here and there. It was the broadness and the generality."
"But people would look at it and say, in any particular area, that doesn't seem so smart. And so many people just shrugged and they just dismissed it."
"I just look at it and I say it's the same thing over and over again. The limitations of the models, six months from now, the things that seemed hard, they're going to be easy. 12 months from now, the things that seemed impossible, the models are going to be doing them regularly."
"It's everything we saw with GPT-2; we're seeing it over and over again in wider and wider swaths of the world. And the world is gradually finally being convinced."
Daniela Amodei: "I remember at the time there were some very well-respected computer scientists and AI and ML leaders who basically bullied some of the researchers who worked on [GPT-2] on Twitter and were like, this is stupid. We don't know why you guys are even talking, this model is so incapable."
"It was like you were sort of a quack if you believed this was a really significant moment."
Dario: "That's all been memory-holed. All the researchers who were like, oh, this is pseudoscience, it doesn't make sense, all of them are applying for jobs at the LLM companies."
"We never say it out loud to any particular one, but we look at them and they're like, we remember what you were saying in 2019. You have no credibility on this matter."
1
326











