Charles Qi @charles_rqi Twitter profile

Last Seen Profiles

@SenStadelman

@DAlathel

@FWISDAthletics

@moTik

@Francis_Lam

@konoko_no_tani

@vantezinho

@gatosegatinho

@zg_newengland

@SomosAlticeDo

@Sg6i4J5hP18j0wr

@zalyshenikvity

@allyouareismeme

@jandakembangstw

@SenjuCryptoClub

@freshprxncej

@Gun_Batman

@iconicbronze

@RutWaterPark

@____yutauta

@ezeqariasmusic

@MrsCooperYHS

@M_Alialhaweri

@yappauruseiwaya

@nejattastan

@SidekickBrick

@AliciaContursi

@roland_tweets

@SidekickBrick

@AinarsZabers

@0Knowledge_net

@upsolveAI

@Joguin_

@holyhuppert

@112Deventer_nl

@KTTunstall

Charles Qi

@charles_rqi

2 months

Day One at Tesla. It’s already mind blowing 🤯

166

176

5K

Charles Qi

@charles_rqi

2 months

Career Update: Today, I bid farewell to Waymo, marking the end of a chapter in my career. Joining Waymo in late 2019, I entered a world where Level 4 (L4) robotaxi services were a concept rather than a reality. Now, in 2024, Waymo operates a rider-only robotaxi service in four

213

161

3K

Charles Qi

@charles_rqi

2 months

Gradually, then suddenly.

Tesla AI

@Tesla_AI

2 months

Hit >1 billion miles driven on FSD

1K

2K

12K

32

65

1K

Charles Qi

@charles_rqi

2 months

First week at Tesla. The vibe felt like a startup: few meetings, lots of hardcore building, flat team structure, direct communication across levels, super fast iterations. I really liked it so far. Referral link below to buy a Tesla car with benefits.

Order Tesla products through Charles's Referral

www.tesla.com

29

64

1K

Charles Qi

@charles_rqi

1 month

I was fortunate to have test driven the car — the acceleration is insane. It also uses premium materials for sports seats and has great sound insulation.

Tesla

@Tesla

1 month

New Model 3 Performance launching today 🏎️ → 0-60 mph in 2.9 510 hp / 741 Nm 163 mph top speed — Performance-tuned chassis Same quiet & comfortable cabin plus bespoke chassis hardware for improved stiffness and higher performance baseline. More power,

1K

3K

18K

15

34

613

Charles Qi

@charles_rqi

1 month

Another perk for Tesla EV is its super low maintenance. The only time I sent my car to maintenance over the past 5 years is for a simple tire rotation.

Elon Musk

@elonmusk

1 month

19K

24K

358K

23

41

470

Charles Qi

@charles_rqi

2 months

Glad to host Prof. Zhu / Lead of NVIDIA robotics lab @yukez for an FSD ride. More to come for physical world AI !

Yuke Zhu

@yukez

2 months

Got a taste of @Tesla 's FSD v12.3.4 last night. By no means flawless, but the human-like driving maneuvers (with no interventions) delivered a magical experience. Excited to witness the recipe of scaling law and data flywheel for full autonomy show signs of life in real products.

32

208

2K

6

31

375

Charles Qi

@charles_rqi

27 days

What useful tasks can a humanoid robot do for a factory? This newly released video from Tesla shows Optimus (Tesla Bot) sorting battery cells, a seemingly simple but nuanced task 🤖 🦾🔋 The video also gives a glimpse of how the bot was trained — through human demonstration

Tesla Optimus

@Tesla_Optimus

27 days

Trying to be useful lately!

2K

6K

31K

11

24

360

Charles Qi

@charles_rqi

1 month

Tesla shareholders, it’s time to vote.

Tesla - Protecting Your Investment and Tesla’s Future

At this year’s Annual General Meeting of Stockholders, we will be asking you to help support the continuation of Tesla’s extraordinary growth by approving two important proposals unanimously recomm...

www.votetesla.com

9

32

315

Charles Qi

@charles_rqi

2 months

FSD subscription: $199 -> $99/month!

Tesla

@Tesla

2 months

You can now subscribe to FSD (Supervised) for $99/month in the US Upgrades > Software Upgrades > Subscribe

2K

14K

7

6

274

Charles Qi

@charles_rqi

2 months

Ideas are cheap — that’s something I learned from my PhD experience too. Often time a researcher sees a paper and says I have thought of the same idea, I could have done this paper! However turning an initial idea to a final product, whether that’s a paper or a business, is the

Elon Musk

@elonmusk

2 months

Precisely

8K

38K

228K

4

21

232

Charles Qi

@charles_rqi

1 month

Beijing is hosting an international auto exhibition this weekend, drawing significant attention to EVs. Although tesla did not participate the exhibit, Elon’s timely presence in the city certainly amplified the buzz.

Elon Musk

@elonmusk

1 month

Honored to meet with Premier Li Qiang. We have known each other now for many years, since early Shanghai days.

11K

20K

203K

3

5

165

Charles Qi

@charles_rqi

2 months

Wonderful fireside chat between Jim Fan @DrJimFan and Percy Laing @percyliang on the Future of Foundation Models at #GTC2024 A few memorable points to myself: 1. Percy is the person who coined the term “foundation models”, or FoMo :) The switch turned on for him in 2020 when

5

6

110

Charles Qi

@charles_rqi

1 month

Training ML models is addictive. YOLO training is even more captivating.

8

10

107

Charles Qi

@charles_rqi

1 month

900+ miles / hour super charging ⚡️⚡️⚡️

3

1

103

Charles Qi

@charles_rqi

2 months

In-depth meditation (3+ day continued practice) is worth a try for everyone. It takes 2-3 days to get over the drowsiness and to calm our thoughtful mind. Then with good practice one can enter a mode of flow, to experience the true peacefulness. With enough practice, such a

Andrej Karpathy

@karpathy

2 months

Returning from an experimental ~2 week detox from the internet. Main takeaway is that I didn't realize how unsettled the mind can get when over-stimulating on problems/information (like a stirred liquid), and ~2 weeks is enough to settle into a lot more zen state. I'm struck by

536

926

13K

4

97

Charles Qi

@charles_rqi

2 months

Thank you all for the kind words and support!

3

1

79

Charles Qi

@charles_rqi

27 days

The evolution of deep learning frameworks never stops. My own journey started with Matlab, Theano, Caffe, torch, and then went on to Tensorflow, PyTorch and JAX: 2013-2014: Matlab. Studying Andrew Ng’s early day DL tutorials and rotating in Andrew’s lab working on NN

Andrej Karpathy

@karpathy

28 days

# CUDA/C++ origins of Deep Learning Fun fact many people might have heard about the ImageNet / AlexNet moment of 2012, and the deep learning revolution it started. What's maybe a bit less known is that the code backing this winning submission to the

147

944

7K

1

3

78

Charles Qi

@charles_rqi

2 months

@ray4tesla I just got the update at 4:30pm PST today on my personal model 3. It is coming to more cars.

4

1

75

Charles Qi

@charles_rqi

22 days

Impulse shopping! I just ordered a CyberHammer — limited edition, only 800 of them in the world.

Javier Verdura

@JVerdura

23 days

It’s Hammer Time!

10

5

135

3

0

55

Charles Qi

@charles_rqi

2 months

@1stPrincipleInv I was considering to take a two-week vacation and join next week. Zero regret that I joined in March to witness these moments.

1

4

51

Charles Qi

@charles_rqi

7 months

I’m starting to use Twitter/X. Let’s start with 🐱🐈 — welcome our new family member NiuNiu! (取名叫牛牛因为她是一只社牛小猫)

2

0

47

Charles Qi

@charles_rqi

21 days

+1 The bitter lesson in AI is that that general methods that leverage computation are ultimately the most effective, and by a large margin. This is due to the quickly falling cost of compute (e.g. Moore’s law, GPU). Exponential data growth is another contributor that is not

Noam Brown

@polynoamial

21 days

I wish every grad student in AI would read The Bitter Lesson

23

53

478

1

5

45

Charles Qi

@charles_rqi

3 months

A typical day in Mountain View, California. Riding in FSD following a Waymo car, smoothly passing through a construction zone. Autonomy is closer than you think. #fsd #waymo

1

41

Charles Qi

@charles_rqi

2 months

@teslaownersSV 👀

2

1

40

Charles Qi

@charles_rqi

3 months

I’m attending Nvidia GTC next week in person, for the first time ever! I’m going with three questions in mind: 1. Is there a “Moore’s law” for AI computing? How soon can we reduce the cost of compute by 1000x? What is the bottleneck, is it memory, bandwidth or energy? 2. How

5

1

34

Charles Qi

@charles_rqi

7 months

Riding a Waymo robotaxi (rider only, no safety driver!) in San Francisco downtown from the finance district to little Italy. Smooth drive and no disengage. Will post another one with FSD in the next tweet. #waymo #robotaxi #SanFrancisco

1

2

30

Charles Qi

@charles_rqi

2 months

This beast 2-meter-tall rack can host 72 GB200 GPUs with 72KW liquid cooling. Estimated to cost $3M (my rough guess) and arriving in Q4 2024. #GTC2024 #nvidia #B200

1

4

29

Charles Qi

@charles_rqi

1 month

@Trescend What YOLO training is: When you have limited resources, you couldn’t iterate over all design choices. So you collect all good ideas and trust your intuition, then You Only Live/Train Once for the best configuration.

3

1

25

Charles Qi

@charles_rqi

2 months

Agent, search, system 2, slow thinker. Large models can/should do better by thinking harder and using tools & memory.

Andrew Ng

@AndrewYNg

2 months

I think AI agentic workflows will drive massive AI progress this year — perhaps even more than the next generation of foundation models. This is an important trend, and I urge everyone who works in AI to pay attention to it. Today, we mostly use LLMs in zero-shot mode, prompting

209

1K

5K

4

1

24

Charles Qi

@charles_rqi

1 month

@lexfridman I like this date. It’s also when Bitcoin reward halved in 2024. A memorable date.

2

1

19

Charles Qi

@charles_rqi

3 months

Congratulations @SpaceX ! Very successful 3rd launch of the Starship. We are a big step closer to landing on Mars!

1

0

21

Charles Qi

@charles_rqi

3 months

All those “AGI will arrive by 20XX” predictions remind me of the hype time of autonomous driving in 2012-2017, when Google/GM/Ford/Tesla etc. all made various predictions of widespread L4 autonomy by early 2020s. In reality it’s much easier to show demos and small scale success

4

1

20

Charles Qi

@charles_rqi

3 months

With the advent of LLMs, tech companies (product & org) are changing rapidly, or even being completely redefined. This results in a "high frequency trading" mode of top AI talents. As an individual, you need to know your position in this paradigm shift. Are you the one being

2

1

19

Charles Qi

@charles_rqi

4 months

In the era of neural network 3.0: novelty is no longer inventing exotic / special architectures. NN1.0: the perceptron era. shallow learning. NN2.0: the deep learning era. all kinds of archs were invented (AlexNet, ResNet, LSTM, PointNet, …, Transformers) NN3.0: now

Saining Xie

@sainingxie

4 months

well diffusion transformer was rejected at CVPR 2023 due to limited novelty.

19

54

726

1

0

17

Charles Qi

@charles_rqi

21 days

When I review an academic paper, I often ask myself: Is the method scalable to leverage more compute and data? for a modeling work, I consider: significance = metric gains / complexity. Often times, an over engineered architecture or complex pipeline does not survive this

1

0

17

Charles Qi

@charles_rqi

7 months

Trying out FSD at Berkeley (part 1/2)

1

0

16

Charles Qi

@charles_rqi

3 months

Waymo has a Research Scientist opening with a focus on computer vision and generative modeling for the application of road understanding/mapping. Preferred background: MS/PhD with top-tier CV/ML publications or/and ML experiences in industry.

0

1

14

Charles Qi

@charles_rqi

3 months

Bonus package: Vision Pro, Waymo robotaxi member, X Premium/Grok. What else?

0

13

Charles Qi

@charles_rqi

22 days

Sold out.

1

0

12

Charles Qi

@charles_rqi

3 months

Recent AI generative models like stable diffusion and Sora showcase neural compression as a pivotal tech. Initially I viewed these neural image/video compression models as novel yet impractical due to its much higher computation cost. I couldn’t realize its potential links to

1

9

Charles Qi

@charles_rqi

2 months

#GTC2024 is on! First impression: it’s like a workshop/tutorial day in an academic conference (CVPR, ICCV) with various tech talks and Jensen’s keynote coming in the afternoon.

0

8

Charles Qi

@charles_rqi

3 months

Search is a nice way to trade time for accuracy. NN is a “fast” thinker. Search is a “slow” thinker. Fast and slow gets you the best results over all. There is still much to explore before we get AGI.

Noam Brown

@polynoamial

3 months

8 years ago today, AlphaGo beat Lee Sedol in a milestone for AI. Unlike typical neural nets, AlphaGo spent ~1 minute per move improving its policy via search. This boosted its Elo by more than a 1000x bigger model. Even today, nobody has trained a raw NN that is superhuman in Go.

35

89

784

0

1

8

Charles Qi

@charles_rqi

3 months

This is huge, if it contains both training and inference code with a permissive license.

Elon Musk

@elonmusk

3 months

This week, @xAI will open source Grok

9K

11K

94K

0

7

Charles Qi

@charles_rqi

2 months

What’s your role in AI? A message to the young audience from Fei-Fei Li as a closing remark of the fireside chat between Fei-Fei and Bill Dally. #GTC2024 Other key points: 1. Compared to worrying about AI singularity, there are more urgent near term topics such as its impact

1

0

7

Charles Qi

@charles_rqi

2 months

Re Q1: For the last 8 years (2016-2024), AI compute has grown 1000x. If this trend continues, we’ll see another 1000x of compute by 2032 — that likely means training the GPT4 within 2 hours (vs the current 90 days with 8000 H100).

1

2

7

Charles Qi

@charles_rqi

4 months

How to build the data flywheel for robots? - I believe it is ultimately by directly observing humans (no extra HW required) + low level control learning with SSL/RL. - The humanoid form factor is needed to align with human. - UMI is a great step towards it (lower cost).

Cheng Chi

@chichengcc

4 months

Can we collect robot data without any robots? Introducing Universal Manipulation Interface (UMI) An open-source $400 system from @Stanford designed to democratize robot data collection 0 teleop -> autonomously wash dishes (precise), toss (dynamic), and fold clothes (bimanual)

42

343

2K

0

1

7

Charles Qi

@charles_rqi

2 months

Key points from the fireside chat between David Luan (Adept CEO) and Bryan (Nvidia VP). #GTC2024 1. We have used up text tokens (all human content created ever in all languages) but the models are not done. Moving forward we will see increased specialization (e.g. finetuning and

0

7

Charles Qi

@charles_rqi

3 months

Exciting news. Waymo can now operate robotaxis in the Peninsula and LA areas!

San Francisco Chronicle

@sfchronicle

3 months

JUST IN: State regulators say Waymo can expand its driverless ride-hailing operations to the Peninsula and Los Angeles, marking a massive commercial expansion that launched last year in San Francisco.

13

22

251

0

6

Charles Qi

@charles_rqi

3 months

Search 2.0 is personal assistant. Every one of us will be the CEO of our self company and has AI executive assistants serving us. Such AI driven search is generative. Results are generated instead of retrieved. They are customized to your question and your context (why you ask,

0

6

Charles Qi

@charles_rqi

3 months

What about generating highly compressed VQ (vector quantized) tokens that are discrete? That makes the video generation problem much more manageable. Some examples are VQVAE tokens used in DALL-E and MAGVITv2 tokens used in VideoPoet.

Yann LeCun

@ylecun

3 months

Modeling the world for action by generating pixel is as wasteful and doomed to failure as the largely-abandoned idea of "analysis by synthesis". Decades ago, there was a big debate in ML about the relative advantages of generative methods vs discriminative methods for

123

234

2K

0

5

Charles Qi

@charles_rqi

4 months

1. Money is a tool invented by human for value exchange across space and time. 2. Bitcoin is a “better” form of money than gold and fiat currencies. 3. Adoption rate of the better form of money will go up over time. (This is not financial advice)

3

0

4

Charles Qi

@charles_rqi

2 months

Stanley has proved itself to get into the Nvidia gear store at GTC 😂 #StanleyCup #GTC2024

1

0

5

Charles Qi

@charles_rqi

1 month

@Trescend Yeah

1

0

5

Charles Qi

@charles_rqi

2 months

@TeslaTrillion Food trucks are not bad.

0

5

Charles Qi

@charles_rqi

7 months

Trying out FSD at Berkeley (part 2/2)

0

5

Charles Qi

@charles_rqi

1 month

@hackbearterry Welcome to X, Terry! I like your YouTube channel.

2

0

5

Charles Qi

@charles_rqi

2 months

@Bobby_Downes @Tesla In the first order now link (under Explore Inventory), one can filter the model of the car to select Model 3.

0

5

Charles Qi

@charles_rqi

3 months

Happy International Women’s Day! 🎉 Last week we hosted a fireside chat with four incredible female leaders in tech. Check out the post for takeaways! Follow the program on LinkedIn or follow me on X for more updates of future events.

The Lighthouse Mentorship Program 领航计划 on LinkedIn: #internationalwomensday2024 #internationalwom...

In honor of #internationalwomensday2024 , we recently invited four female leaders Yanbing Li, Wenjia Fang, Ph.D, Xin (Maggie) Du, Emma Yu to share their…

www.linkedin.com

0

4

Charles Qi

@charles_rqi

2 months

@ByeonChansoo It’s possible as a result of mastery but should probably not as a way of practice.

0

4

Charles Qi

@charles_rqi

2 months

Typo: Laing -> Liang. “AI” is embedded in our Professor’s last name, just not in the right order :)

0

1

4

Charles Qi

@charles_rqi

2 months

@IvanCartago @DrJimFan @percyliang They will be released 72 hours after the event if I remember correctly

1

3

Charles Qi

@charles_rqi

2 months

@PaulyK59060518 Maybe you can donate the ideas at X :)

2

0

4

Charles Qi

@charles_rqi

3 months

@chuhang1122 Manufacturing capacity is a challenge for many companies.

1

0

2

Charles Qi

@charles_rqi

3 months

@chris_j_paxton Agree. An iterative approach with a viable business model is key here. It needs time to mature.

1

0

3

Charles Qi

@charles_rqi

3 months

@StefanoErmon Competition is on. In video gen, AR Model is catching up diffusion. In text gen, now we have diffusion competitive as AR. Wondering what’s your opinion on the end game?

0

3

Charles Qi

@charles_rqi

2 months

Re Q2: Nvidia continues to lead by co-optimizing HW, system and algorithms. This lets it provide AI computing solutions 1-2 magnitude better than alternatives (including its own previous gen solutions!) in terms of throughput and energy efficiency, which gives Nvidia an edge in

0

1

3

Charles Qi

@charles_rqi

3 months

@xf1280 @remyxai Congrats!

0

2

Charles Qi

@charles_rqi

4 months

@wsunid Any new discovery about data, model, scaling is a solid contribution. It’s about the information gain.

0

2

Charles Qi

@charles_rqi

3 months

@JieWang_ZJUI Since the industry is moving so fast. The opportunity cost of doing a PhD is higher than before. Master with research experiences is a good balance. PhD will be for those who value freedom in research or interested in faculty jobs.

1

0

1

Charles Qi

@charles_rqi

3 months

@space_tintin Would love to play it. What’s the name of the game. AI-Arena?

1

0

2

Charles Qi

@charles_rqi

3 months

When network capacity becomes really big and computers are really cheap, this seems very possible. Probably we also need self learning on network pruning to recycle unused connections (like in human baby brain’s “pretraining”). It will take a while (5+ years?) to get there but

AK

@_akhaliq

3 months

Beyond Language Models Byte Models are Digital World Simulators Traditional deep learning often overlooks bytes, the basic units of the digital world, where all forms of information and operations are encoded and manipulated in binary format. Inspired by the success of next

15

132

806

0

1

Charles Qi

@charles_rqi

3 months

Looks like a fun game to play. @karpathy

space tintin

@space_tintin

3 months

Who are you picking?

71

82

1K

0

1

Charles Qi

@charles_rqi

3 months

Interestingly these generative models themselves can be viewed as a compressor of data. E.g. GPT compresses the internet text, as there is a strong tie between predicting next token accurately (less entropy) and efficient compression.

0

1

Charles Qi

@charles_rqi

3 months

@AravSrinivas This quality is insane

0

1

Charles Qi

@charles_rqi

3 months

@bowenc0221 @Tesla @OpenAI Congrats on your new role!

0

1

Charles Qi

@charles_rqi

3 months

@karpathy We need an AI personal assistant to filter, select and rank such media flow for us..

0

1

Charles Qi

@charles_rqi

3 months

Highlights: 9:30am, 5:05pm, 1am

Jason Wei

@_jasonwei

3 months

My typical day as a Member of Technical Staff at OpenAI: [9:00am] Wake up [9:30am] Commute to Mission SF via Waymo. Grab avocado toast from Tartine [9:45 am] Recite OpenAI charter. Pray to optimization Gods. Learn the Bitter Lesson [10:00am] Meetings (Google Meet). Discuss how to

156

311

4K

1

0

1

Charles Qi

@charles_rqi

3 months

@JitendraMalikCV Nice work! Curious how do you think about the path forward to scale up the data for such approach?

1

0

1

Charles Qi

@charles_rqi

3 months

@karpathy Thank you Andrej! I really like your lectures and found them actually quite enjoyable to watch with its clarity and hands on coding.

0

1

Charles Qi

@charles_rqi

3 months

@Ishan345 @Waymo Good to hear! Progresses are definitely significant. Many thought L4 in San Francisco was impossible in this century, but look we are here in 2024. There are both dramatic over and under predictions

0

1

Charles Qi

@charles_rqi

1 year

I signed up for The San Francisco Marathon on Jul 23, 2023!

1

0

1

Charles Qi

@charles_rqi

4 months

The devil is in the data

0

1

Charles Qi

@charles_rqi

4 months

From their tech report/blog: Sora is based on Latent Diffusion with DiT (diffusion transformers). This is in contrast to the recent Google work “VideoPoet” with autoregressive models and discrete tokens. Is it the model or the data that matters most?

OpenAI

@OpenAI

4 months

Introducing Sora, our text-to-video model. Sora can create videos of up to 60 seconds featuring highly detailed scenes, complex camera motion, and multiple characters with vibrant emotions. Prompt: “Beautiful, snowy