Loading video...

Video Failed to Load

There was a problem loading this video. This could be due to a temporary network issue or the video might be unavailable.

ELON: DATA COMPRESSION IS TESLA’S BIGGEST AI CHALLENGE "The single biggest technical challenge of Tesla self-driving AI is context compression of 1.5GB/s of video to ~2kB/s of control outputs using a puny inference computer without making any mistakes. Super hard to avoid overly lossy compression at any given step."

Mario Nawfal

3,609,010 subscribers

486,215 views • 8 months ago •via X (Twitter)

News & Politics Science & Technology

Anya Rossi• Live Now

Private livecam show

0 Comments

No comments available

Comments from the original post will appear here

Related Videos

AI has a "last-mile problem" similar to self-driving cars. With self-driving cars, early demos impressed, but real-world deployment took years. It's easy to hack up a prototype, but making it work reliably at scale is hard. If each step of an AI agent is only 95% accurate, none of the 30-step workflows will work reliably. Going from 95% to 99.9% accuracy is the real challenge.

AI has a "last-mile problem" similar to self-driving cars. With self-driving cars, early demos impressed, but real-world deployment took years. It's easy to hack up a prototype, but making it work reliably at scale is hard. If each step of an AI agent is only 95% accurate, none of the 30-step workflows will work reliably. Going from 95% to 99.9% accuracy is the real challenge.

Richard Socher

14,926 views • 1 year ago

Elon Musk explains why AI could be capable of discovering new physics. “You can think of intelligence like a compression problem. For any given field, it can be physics, or any field, you can distill things down to their most basic axiomatic elements, the things that are extremely likely to be true. And then, you can reason up from there to reach a cogent conclusion. I think if you apply that sort of first-principles compression to a field, and then you decompress it and contest the results against reality, then AI is capable of innovation and discovering new physics.” Town Hall in Philadelphia, October 18, 2024

Elon Musk explains why AI could be capable of discovering new physics. “You can think of intelligence like a compression problem. For any given field, it can be physics, or any field, you can distill things down to their most basic axiomatic elements, the things that are extremely likely to be true. And then, you can reason up from there to reach a cogent conclusion. I think if you apply that sort of first-principles compression to a field, and then you decompress it and contest the results against reality, then AI is capable of innovation and discovering new physics.” Town Hall in Philadelphia, October 18, 2024

ELON CLIPS

1,054,036 views • 1 year ago

Elon Musk says Tesla swallows 1.5 GB of video into 2 KB of control. "We really are photons in, controls out. That is the vast majority of your life: vision, photons in, and then motor controls out." Musk reduced the human nervous system to a single equation. That theory runs in every Tesla on the road. "Your Tesla is taking in one and a half gigabytes a second of video and outputting two kilobytes a second of control outputs with the video at 36 hertz and the control frequency at 18." A 750,000-to-1 compression ratio. The car decides which photons matter. "You don't care about the details of the leaves on the tree on the side of the road, but you care a lot about the road signs and the traffic lights, the pedestrians, and even whether someone in another car is looking at you or not looking at you." Then Musk extended the equation past Tesla. The same architecture runs Optimus. **Photons in. Controls out.** Different number of degrees of freedom. A car has steering and acceleration. A robot arm has dozens of joints. Same compression. More fingers on the other side. Musk, who built one Tesla AI engine across two products: "AI is mostly compression and correlation of two bitstreams." The implication: a human, a Tesla, and an Optimus all run the same loop. Musk, looking back at the human equation: "This is what happens with humans." If you're new here, GeniusThinking is a gallery for the greatest minds in economics, psychology, and history. Follow along for more similar content. P.S. I made a playbook breaking down 100+ mental models used by history's greatest thinkers. 5,000+ downloads. 113 five-star reviews. Comment "MODELS" + Follow GeniusThinking and I'll send it to you. — Elon Musk ( Elon Musk ), CEO of Tesla and SpaceX, on Dwarkesh Patel's ( Dwarkesh Patel ) podcast

Elon Musk says Tesla swallows 1.5 GB of video into 2 KB of control. "We really are photons in, controls out. That is the vast majority of your life: vision, photons in, and then motor controls out." Musk reduced the human nervous system to a single equation. That theory runs in every Tesla on the road. "Your Tesla is taking in one and a half gigabytes a second of video and outputting two kilobytes a second of control outputs with the video at 36 hertz and the control frequency at 18." A 750,000-to-1 compression ratio. The car decides which photons matter. "You don't care about the details of the leaves on the tree on the side of the road, but you care a lot about the road signs and the traffic lights, the pedestrians, and even whether someone in another car is looking at you or not looking at you." Then Musk extended the equation past Tesla. The same architecture runs Optimus. Photons in. Controls out. Different number of degrees of freedom. A car has steering and acceleration. A robot arm has dozens of joints. Same compression. More fingers on the other side. Musk, who built one Tesla AI engine across two products: "AI is mostly compression and correlation of two bitstreams." The implication: a human, a Tesla, and an Optimus all run the same loop. Musk, looking back at the human equation: "This is what happens with humans." If you're new here, GeniusThinking is a gallery for the greatest minds in economics, psychology, and history. Follow along for more similar content. P.S. I made a playbook breaking down 100+ mental models used by history's greatest thinkers. 5,000+ downloads. 113 five-star reviews. Comment "MODELS" + Follow GeniusThinking and I'll send it to you. — Elon Musk ( Elon Musk ), CEO of Tesla and SpaceX, on Dwarkesh Patel's ( Dwarkesh Patel ) podcast

GeniusThinking

160,972 views • 2 months ago

Jeff Dean says models hallucinate because their training data is “squishy.” But what’s in the context window is sharp, the exact text or video frame right in front of them. The goal is to bring more of the world into that context. Compression gives it intuition. Context gives it precision.

Jeff Dean says models hallucinate because their training data is “squishy.” But what’s in the context window is sharp, the exact text or video frame right in front of them. The goal is to bring more of the world into that context. Compression gives it intuition. Context gives it precision.

vitrupo

41,600 views • 3 months ago

Matryoshka Representation Learning (MRL) is a super exciting approach to improving the quality and efficiency of embedding models and strategies ✨ MRL allows models to store more information in the earlier dimensions of their data vectors. This method not only boosts performance in tasks like classification and retrieval, but is also a super cool compression technique! Paper: For compression: It’s been so much fun learning and making this video with @DanielW966, thanks for all your help!

Matryoshka Representation Learning (MRL) is a super exciting approach to improving the quality and efficiency of embedding models and strategies ✨ MRL allows models to store more information in the earlier dimensions of their data vectors. This method not only boosts performance in tasks like classification and retrieval, but is also a super cool compression technique! Paper: For compression: It’s been so much fun learning and making this video with @DanielW966, thanks for all your help!

Victoria Slocum

120,436 views • 1 year ago

DOJO: ELON’S AI GAME-CHANGER Elon: “[Tesla’s Dojo] is the biggest computer overhaul since Cray”—and he’s not kidding. Cray ruled the ‘70s, but Dojo is here to crush AI tasks at scale, training Tesla’s Full Self-Driving system to make cars smarter, safer, and greener. And it’s not just about cars. Dojo’s insane power (shoutout to Tesla’s D1 chip) could spark breakthroughs in healthcare, finance, and chip design. From jobs to science, Dojo isn’t just a computer—it’s a revolution!

DOJO: ELON’S AI GAME-CHANGER Elon: “[Tesla’s Dojo] is the biggest computer overhaul since Cray”—and he’s not kidding. Cray ruled the ‘70s, but Dojo is here to crush AI tasks at scale, training Tesla’s Full Self-Driving system to make cars smarter, safer, and greener. And it’s not just about cars. Dojo’s insane power (shoutout to Tesla’s D1 chip) could spark breakthroughs in healthcare, finance, and chip design. From jobs to science, Dojo isn’t just a computer—it’s a revolution!

Mario Nawfal

115,934 views • 1 year ago

A banger just got released 💥 Here is a snapshot of L2D, the biggest self-driving dataset by far! - 90 TeraBytes of data - 5000 hours of driving - 6 surrounding HD cameras - OPENLY AVAILABLE - Train your car to drive like Tesla at home 🧵 More details in thread

A banger just got released 💥 Here is a snapshot of L2D, the biggest self-driving dataset by far! - 90 TeraBytes of data - 5000 hours of driving - 6 surrounding HD cameras - OPENLY AVAILABLE - Train your car to drive like Tesla at home 🧵 More details in thread

Remi Cadene

171,802 views • 1 year ago

I appreciate your diligence, Sergio, however image compression is a real thing and sometimes indiscernible from the signs of AI. This account will continue to avoid it as much as possible though so keep that skeptical eye, my friend.

I appreciate your diligence, Sergio, however image compression is a real thing and sometimes indiscernible from the signs of AI. This account will continue to avoid it as much as possible though so keep that skeptical eye, my friend.

Awful Taste But Great Execution

277,319 views • 7 months ago

🚨 $TSLA AI VP ASHOK ELLUSWAMY SAYS CURSE OF DIMENSIONALITY MAKES SELF-DRIVING AN EXTREME COMPRESSION CHALLENGE • Tesla cars stream ~2 billion tokens from 8 high-res cameras + history buffer every few seconds, but must compress it all down to just 2 actions (steering + accel/brake) • The real difficulty is learning the correct causal reasons without spurious correlations (e.g., brake because of turn signal, not a moving tree branch) • Tesla fleet generates ~500 years of driving data daily, but most is boring → they proactively identify and collect only rare/interesting events (school buses, skids, hazards) • Large fleet allows fast capture of extremely rare scenarios that are hard or unsafe to stage • With enough rare data, the system achieves proactive safety and early prediction (e.g., braking seconds before a potential crash by reading abnormal vehicle behavior) • Bottom line: self-driving needs massive high-quality edge-case data for real predictive intelligence in real time

🚨 $TSLA AI VP ASHOK ELLUSWAMY SAYS CURSE OF DIMENSIONALITY MAKES SELF-DRIVING AN EXTREME COMPRESSION CHALLENGE • Tesla cars stream ~2 billion tokens from 8 high-res cameras + history buffer every few seconds, but must compress it all down to just 2 actions (steering + accel/brake) • The real difficulty is learning the correct causal reasons without spurious correlations (e.g., brake because of turn signal, not a moving tree branch) • Tesla fleet generates ~500 years of driving data daily, but most is boring → they proactively identify and collect only rare/interesting events (school buses, skids, hazards) • Large fleet allows fast capture of extremely rare scenarios that are hard or unsafe to stage • With enough rare data, the system achieves proactive safety and early prediction (e.g., braking seconds before a potential crash by reading abnormal vehicle behavior) • Bottom line: self-driving needs massive high-quality edge-case data for real predictive intelligence in real time

Tsla Archive

79,757 views • 5 months ago

3Blue1Brown’s new video explains why every LLM is actually a compression machine. everyone describes pre-training as “next token prediction” but that’s just the surface-level objective. in reality it is a means to making the most efficient text compressor. prediction and compression are two sides of the same coin. when you train the model to predict the next token you’re not just teaching it to guess the next word but how to best encode the human knowledge it sees. better compression means better abstraction means better reasoning at some point, compression stops looking like storage or a database (as some like to call it on X) and looks like an approximation of understanding.

3Blue1Brown’s new video explains why every LLM is actually a compression machine. everyone describes pre-training as “next token prediction” but that’s just the surface-level objective. in reality it is a means to making the most efficient text compressor. prediction and compression are two sides of the same coin. when you train the model to predict the next token you’re not just teaching it to guess the next word but how to best encode the human knowledge it sees. better compression means better abstraction means better reasoning at some point, compression stops looking like storage or a database (as some like to call it on X) and looks like an approximation of understanding.

ℏεsam

119,751 views • 1 month ago

🚨 | Lewis Hamilton on Mercedes’ compression ratio: “What’s clear is that they didn’t show the engine power through any of the practice because of the whole talk on the compression issue. They’ve done a really solid job with their engine, which we have as well but I’m trying to understand why it's 2 tenths or more just through power per sector... And so, if it is the compression thing, I wanna understand why the FIA haven't done anything [or] what’s being done to rectify it.”

🚨 | Lewis Hamilton on Mercedes’ compression ratio: “What’s clear is that they didn’t show the engine power through any of the practice because of the whole talk on the compression issue. They’ve done a really solid job with their engine, which we have as well but I’m trying to understand why it's 2 tenths or more just through power per sector... And so, if it is the compression thing, I wanna understand why the FIA haven't done anything [or] what’s being done to rectify it.”

deni

960,023 views • 4 months ago

ELON MUSK: "One of the biggest mistakes people generally make, and I'm guilty of it too is wishful thinking… You ignore the real truth because of what you want to be true. This is a very difficult trap to avoid."

ELON MUSK: "One of the biggest mistakes people generally make, and I'm guilty of it too is wishful thinking… You ignore the real truth because of what you want to be true. This is a very difficult trap to avoid."

Resist the Mainstream

93,418 views • 3 months ago

Time to cook an old Sith recipe, and Darth Vader is up to the task. Welcome to Sith Kitchen🍽️ The biggest challenge in making those types of videos is consistency. I hope the new generation of AI will make it much easier.

Time to cook an old Sith recipe, and Darth Vader is up to the task. Welcome to Sith Kitchen🍽️ The biggest challenge in making those types of videos is consistency. I hope the new generation of AI will make it much easier.

Sith Daddy

68,655 views • 4 months ago

Elon Musk on the biggest danger of AI and robotics: “I think probably the biggest danger of AI or maybe the biggest danger of AI and robotics going wrong is government People who are opposed to corporations or worried about corporations should really worry the most about government. Government is just a corporation in the limit; it is the biggest corporation with a monopoly on violence The government could potentially use AI and robotics to suppress the population. That’s a serious concern” Elon says the real threat is Big Government armed with superintelligence

Elon Musk on the biggest danger of AI and robotics: “I think probably the biggest danger of AI or maybe the biggest danger of AI and robotics going wrong is government People who are opposed to corporations or worried about corporations should really worry the most about government. Government is just a corporation in the limit; it is the biggest corporation with a monopoly on violence The government could potentially use AI and robotics to suppress the population. That’s a serious concern” Elon says the real threat is Big Government armed with superintelligence

X Freeze

719,542 views • 4 months ago

Elon Musk: “The only company with all the ingredients for making intelligent humanoid robots at scale is Tesla” Elon believes Optimus will become the biggest product of all time — 10x bigger than the the biggest product ever made Pilot production of Optimus has already begun and Tesla will be building thousands of bots to put to work at Tesla factories later this year $TSLA

Elon Musk: “The only company with all the ingredients for making intelligent humanoid robots at scale is Tesla” Elon believes Optimus will become the biggest product of all time — 10x bigger than the the biggest product ever made Pilot production of Optimus has already begun and Tesla will be building thousands of bots to put to work at Tesla factories later this year $TSLA

Dalton Brewer

314,213 views • 1 year ago

Sky's economics and data editor Ed Conway takes a look at "the biggest" intervention yet of any UK government minister following the U.S.'s decision to go to war with Iran. Live updates:

Sky's economics and data editor Ed Conway takes a look at "the biggest" intervention yet of any UK government minister following the U.S.'s decision to go to war with Iran. Live updates:

Sky News

17,451 views • 3 months ago

TESLA ISN'T JUST AHEAD IN AI, IT'S AN ORDER OF MAGNITUDE MORE EFFICIENT THAN THE REST OF THE WORLD "I actually think Tesla is ahead of the rest of the world in intelligence density of AI by an order of magnitude or more This is going to sound like a pretty bold statement, but I know what the intelligence efficiency of the big models are like, Grok and a bunch of the other models Tesla’s AI is, in terms of its memory efficiency, I think more than an order of magnitude better. The Tesla AI is very compute-efficient and very memory-efficient One of the metrics one should consider for any given AI model is the intelligence per gigabyte. Especially when you’re constrained on RAM, having an AI that has very high intelligence density per gigabyte for a given number of gigabytes, how much functionality can you get out of it" - Elon Musk

TESLA ISN'T JUST AHEAD IN AI, IT'S AN ORDER OF MAGNITUDE MORE EFFICIENT THAN THE REST OF THE WORLD "I actually think Tesla is ahead of the rest of the world in intelligence density of AI by an order of magnitude or more This is going to sound like a pretty bold statement, but I know what the intelligence efficiency of the big models are like, Grok and a bunch of the other models Tesla’s AI is, in terms of its memory efficiency, I think more than an order of magnitude better. The Tesla AI is very compute-efficient and very memory-efficient One of the metrics one should consider for any given AI model is the intelligence per gigabyte. Especially when you’re constrained on RAM, having an AI that has very high intelligence density per gigabyte for a given number of gigabytes, how much functionality can you get out of it" - Elon Musk

X Freeze

62,736 views • 5 months ago