Steven Wong @StevenRCEN

Well Done Tianjin China Joined September 2012

Tweets

543
Followers

82
Following

287
Likes

1K

Steven Wong @StevenRCEN

3 months ago

@YongQuan 不是噪声开局我不是很认可

0 0 86 14K 0

View Details

Steven Wong @StevenRCEN

3 months ago

@Shipwreckrodney @lrlilolam 吊在哪里？

0 0 1 104 0

View Details

Steven Wong @StevenRCEN

9 months ago

@nitzukai @flarup Not even creative, that’s generative

0 0 0 21 0

View Details

Steven Wong @StevenRCEN

11 months ago

@jerrylimee @BasicAppleGuy Finally, after those years of transition starting with superficial, fake, recorded keynotes, no human is needed.

0 0 1 95 0

View Details

Steven Wong @StevenRCEN

11 months ago

@_devJNS bool ==

0 0 0 8 0

View Details

Jeff Dean @JeffDean

2 years ago

We're using a ReLU to set tariffs?

Zygi @nonagonono

2 years ago

@orthonormalist oops, fixed. Now we get that nice straight line

31 256 2K 1.3M 350

82 426 5K 441K 371

View Details

Steven Wong @StevenRCEN

2 years ago

@Inwal3310 喷不了这个是真想玩

0 0 2 15 0

View Details

fenx @haxfenx

2 years ago

做了一款为位图添加矢量描边的 Figma 插件 figma.com/community/plug…

1 6 39 5K 23

View Details

Steven Wong @StevenRCEN

2 years ago

@WuDefa22 @hatadanna 你说的对，但是新版系统的下载页面还没打开说是

0 0 0 12 0

View Details

Absent Insight @AbsentInsight

2 years ago

@elonmusk Funny you post a wojak meme. I made this one just today lol

26 28 690 38K 43

View Details

Steven Wong @StevenRCEN

2 years ago

@WildCat_zh @JWang_Paradies RSS 不是屎，但是剩下的没啥问题

0 0 1 519 0

View Details

Steven Wong @StevenRCEN

2 years ago

@RobertDaleSmith @TheJackForge Yup, especially for now, if I were to save a few keystrokes, I’ll go for AI code generation. 😓

0 0 1 102 0

View Details

OpenBuild @OpenBuildxyz

2 years ago

5 28 21 5K 0

0 0 0 22 0

View Details

🚀 Shenzhen, get ready! 🚀 The big day is here—ETH Shenzhen kicks off TODAY! 🌟 We’re beyond excited to welcome all of our incredible speakers, and we can’t wait to hear from: @0xGantoL from @morphl2 @coldStar1993 from @meet_mina @btwiuse from @gear_techs @AlongHudson from @chainlink and many more! 🙌 Who are you most excited to see? Tell us by retweeting and tagging us @OpenBuildxyz! 🎁 But wait, there’s more! We’ve got some amazing prizes up for grabs. Simply follow us @OpenBuildxyz, retweet this post, and tag us to enter our on-site raffle! Show us your retweet proof at the event for your chance to win BIG! 🏆 Please check the event’s official website: ethshenzhen.org Don’t miss out on the action—let’s make today unforgettable together! 🎉

5 28 21 5K 0

View Details

Macklemore @macklemore

3 years ago

HIND’S HALL. Once it’s up on streaming all proceeds to UNRWA.

15K 173K 410K 39.4M 33K

View Details

Steven Wong @StevenRCEN

3 years ago

First match! Thx teammates! #Splatoon3 #NintendoSwitch

0 0 1 69 0

View Details

Steven Wong @StevenRCEN

3 years ago

@arcinternet @rishimody Shrug for presets like unit conversion and weather forecast stuff. But I'd like to see a better command bar extension framework come true, like what chromium's @ history, with better design & functionality like Boosts.

0 0 0 48 0

View Details

Jim Fan @DrJimFan

3 years ago

I see some vocal objections: "Sora is not learning physics, it's just manipulating pixels in 2D". I respectfully disagree with this reductionist view. It's similar to saying "GPT-4 doesn't learn coding, it's just sampling strings". Well, what transformers do is just manipulating a sequence of integers (token IDs). What neural networks do is just manipulating floating numbers. That's not the right argument. Sora's soft physics simulation is an *emergent property* as you scale up text2video training massively. - GPT-4 must learn some form of syntax, semantics, and data structures internally in order to generate executable Python code. GPT-4 does not store Python syntax trees explicitly. - Very similarly, Sora must learn some *implicit* forms of text-to-3D, 3D transformations, ray-traced rendering, and physical rules in order to model the video pixels as accurately as possible. It has to learn concepts of a game engine to satisfy the objective. - If we don't consider interactions, UE5 is a (very sophisticated) process that generates video pixels. Sora is also a process that generates video pixels, but based on end-to-end transformers. They are on the same level of abstraction. - The difference is that UE5 is hand-crafted and precise, but Sora is purely learned through data and "intuitive". Will Sora replace game engine devs? Absolutely not. Its emergent physics understanding is fragile and far from perfect. It still heavily hallucinates things that are incompatible with our physical common sense. It does not yet have a good grasp of object interactions - see the uncanny mistake in the video below. Sora is the GPT-3 moment. Back in 2020, GPT-3 was a pretty bad model that required heavy prompt engineering and babysitting. But it was the first compelling demonstration of in-context learning as an emergent property. Don't fixate on the imperfections of GPT-3. Think about extrapolations to GPT-4 in the near future.