Jun (Richard) Wang

[DeepReach AI @DeepReach_AI]

How will AI learn about the physical world? We don’t believe the answer is one giant data source.

It will be thousands of entrepreneurs building local data businesses around the world.

That’s why we built @DeepReach_AI — a platform where entrepreneurs build local data businesses for Physical AI.

Today, we’re excited to launch on @ycombinator
Join our network today!

https://t.co/x6TUl1rPhc

#yc #robotics #startup #AI #physicalAI #data #BusinessSuccess

[Jun (Richard) Wang @AI_richard]

@hijkzzz just curious, why using vllm backend, not sglang? thanks.

[Jun (Richard) Wang @AI_richard]

@suchenzang 小子*😂

[Jun (Richard) Wang @AI_richard]

@suchenzang There are many reasons people can’t contribute. Many times it’s not something they can control easily. Especially when there’s an establishment

[Jun (Richard) Wang @AI_richard]

@elonmusk @agihouse_org Agree with what’s written here for the most part. The problem is without a frontier model learning is hard, agent is crippled..

[Jun (Richard) Wang @AI_richard]

@yunta_tsai 哈哈，他是个大魔头，杀人如麻。。

[Mark Chen @markchen90]

I joined @OpenAI as a resident.

First, get the fundamentals down. If there's one subject you need to know inside and out, it's linear algebra. Read and understand a classic textbook like Bishop's Pattern Recognition and Machine Learning.

Then, take on an ambitious project. I was deeply inspired by AlphaGo, and in preparation for my residency interviews, I replicated DQN. There's something magical about training a model better than you at something you care about (even if it's only Atari games).

Finally, there's no replacement for learning from re-implementing papers. I started my research career in generative modeling and still vividly remember replicating the results from ResNet, Attention is All You Need, the VAE paper, the PixelCNN paper, etc. No cheating - match the plots and tables in these papers *exactly*. You'll learn so many little tricks and common pitfalls in the process.

I'll be looking out for your friend's application soon!

[Jun (Richard) Wang @AI_richard]

@ruizhang_nlp @YusenZhangNLP Congrats

[Jun (Richard) Wang @AI_richard]

@suchenzang

[Jun (Richard) Wang @AI_richard]

@simonw What should we do 😂

[Jun (Richard) Wang @AI_richard]

Insightful findings!

[Damien Ferbach @damien_ferbach]

It's very difficult to improve the *exponent* in scaling laws for loss vs compute, especially by changing the optimizer!
Our new paper shows that scaling momentum correctly can *provably* improve the scaling exponent on a theoretical model. Empirically, it works on LSTMs too!

[Jun (Richard) Wang @AI_richard]

@zacharynado Hard to scale for RL training?

[Jun (Richard) Wang @AI_richard]

@ben_athi Way to go!

[Jun (Richard) Wang @AI_richard]

@jluan Robotic AI is the next big thing ? 😄

[Jun (Richard) Wang @AI_richard]

@vwxyzjn @CyouSakura if you do leetcode, you'll be familiar with this kind of tricks :)

[Jun (Richard) Wang @AI_richard]

@miramurati Congrats!!

[Jun (Richard) Wang @AI_richard]

@johnschulman2 Congrats!!

[Jun (Richard) Wang @AI_richard]

@lifan__yuan what do you mean by `decomposition for multiplication`?

[Jun (Richard) Wang @AI_richard]

@PattyMurray Standard time please

[Jun (Richard) Wang @AI_richard]

@jluan @pabbeel 🔥🔥🔥