Video wird geladen...

Video konnte nicht geladen werden

Beim Laden dieses Videos ist ein Problem aufgetreten. Dies könnte an einem vorübergehenden Netzwerkproblem liegen oder das Video ist möglicherweise nicht verfügbar.

Tesla Full Self-Driving capability demo in Austin FSD capability features use vision neural networks to perceive & understand the world, just like humans do. Via our unique fleet learning approach, we are able to collect anonymized data from our vehicles—meaning the neural nets have learned from a few orders... show more

Tesla

24,784,973 subscribers

6,804,722 Aufrufe • vor 2 Jahren •via X (Twitter)

Autos & Fahrzeuge Wissenschaft & Technologie Gesundheit & Wellness

Anya Rossi• Live Now

Private livecam show

10 Kommentare

Profilbild von Sawyer Merritt

Sawyer Merrittvor 2 Jahren

Great video. You should post more of these in the future so more people see how far along this technology is.

Profilbild von Chuck Cook

Chuck Cookvor 2 Jahren

Demo'ing in Elon mode! Thats not fair, we all want this mode .. for demonstration purposes only of course.

Profilbild von Joe Tegtmeyer 🚀 🤠🛸😎

Joe Tegtmeyer 🚀 🤠🛸😎vor 2 Jahren

It’s getting better but still has areas that need improvement. Navigation database errors or omissions, speed limit awareness and acting accordingly, lane selection and awareness of actions like hard braking with traffic directly behind are areas that need more attention.

Profilbild von Quinn Nelson

Quinn Nelsonvor 2 Jahren

Don't disable the hands-on steering wheel requirement just for an ad. Not cool.

Profilbild von Gail Alfar

Gail Alfarvor 2 Jahren

this part is crucial, exiting freeway and moving to immediate right and fsd nailed it - this is the future - ❤️‍🔥 @Tesla_AI

Profilbild von Dirty Tesla

Dirty Teslavor 2 Jahren

Everyone head to to see many hours of FSD footage :)

Profilbild von Tesla Shawn - Comedian/ Anti-Hero

Tesla Shawn - Comedian/ Anti-Herovor 2 Jahren

I hope it stops accidents like this in the future 🤙 Drunk driver hit me head on at 55 mph and Tesla safety standards allowed me to walk away! It’s day 51 thanking @elonmusk and @tesla for their approach to safety first 🙏

Profilbild von Teslaconomics

Teslaconomicsvor 2 Jahren

No line of code needed - video in, controls out

Profilbild von yas

yasvor 2 Jahren

We need @CommunityNotes to restate the disclaimer because most won't read it in the tiny text. HANDS-ON REQUIREMENT WAS DISABLED (CUSTOMER CANNOT DISABLE THIS FEATURE). NOT TRUE DEMO OF FSD.

Profilbild von Tesla Owners Silicon Valley

Tesla Owners Silicon Valleyvor 2 Jahren

This is a great video. More needed.

Ähnliche Videos

You think you've seen it all, FSD actually has It's trained on what amounts to over 100 years of real-world driving scenarios from our fleet of >6M vehicles A lifetime of driving scenarios in 10 minutes

You think you've seen it all, FSD actually has It's trained on what amounts to over 100 years of real-world driving scenarios from our fleet of >6M vehicles A lifetime of driving scenarios in 10 minutes

Tesla

881,694 Aufrufe • vor 8 Monaten

Elon Musk: “In order to solve full self-driving properly, you have to solve real-world AI. The road networks are designed to work with a biological neural net — our brains — and with vision — our eyes. So, in order to make it work with computers, you need to solve real-world AI and vision. We need cameras and silicon neural nets for self-driving to work in a system designed for eyes and biological neural nets. When you put it that way, it’s quite obvious that the only way to solve full self-driving is to solve real-world AI and sophisticated vision.”

Elon Musk: “In order to solve full self-driving properly, you have to solve real-world AI. The road networks are designed to work with a biological neural net — our brains — and with vision — our eyes. So, in order to make it work with computers, you need to solve real-world AI and vision. We need cameras and silicon neural nets for self-driving to work in a system designed for eyes and biological neural nets. When you put it that way, it’s quite obvious that the only way to solve full self-driving is to solve real-world AI and sophisticated vision.”

Elonogy

14,331 Aufrufe • vor 1 Monat

Elon Musk: The only way to solve full self-driving properly is to solve real-world AI. ”What are the road networks designed to work with? They're designed to work with a biological neural net, our brains, and with vision, our eyes. And so, in order to make it work with computers, you basically need to solve real-world AI and vision. We need cameras and silicon neural nets in order to have self-driving work for a system that was designed for eyes and biological neural nets. I guess when you put it that way, it's quite obvious that the only way to solve full self-driving is to solve real-world AI and sophisticated vision.” TED Interview, April 14, 2022

Elon Musk: The only way to solve full self-driving properly is to solve real-world AI. ”What are the road networks designed to work with? They're designed to work with a biological neural net, our brains, and with vision, our eyes. And so, in order to make it work with computers, you basically need to solve real-world AI and vision. We need cameras and silicon neural nets in order to have self-driving work for a system that was designed for eyes and biological neural nets. I guess when you put it that way, it's quite obvious that the only way to solve full self-driving is to solve real-world AI and sophisticated vision.” TED Interview, April 14, 2022

ELON CLIPS

20,411 Aufrufe • vor 1 Jahr

ELON: TO SOLVE FULL SELF-DRIVING, YOU NEED TO SOLVE REAL-WORLD AI "What are the road networks designed to work with? They're designed to work with a biological neural net, our brains, and with vision, our eyes. And so, in order to make it work with computers, you basically need to solve real-world AI and vision. We need cameras and silicon neural nets in order to have self-driving work for a system that was designed for eyes and biological neural nets. I guess when you put it that way, it's quite obvious that the only way to solve full self-driving is to solve real-world AI and sophisticated vision." Source: Elon Musk TED Interview, April 2022

ELON: TO SOLVE FULL SELF-DRIVING, YOU NEED TO SOLVE REAL-WORLD AI "What are the road networks designed to work with? They're designed to work with a biological neural net, our brains, and with vision, our eyes. And so, in order to make it work with computers, you basically need to solve real-world AI and vision. We need cameras and silicon neural nets in order to have self-driving work for a system that was designed for eyes and biological neural nets. I guess when you put it that way, it's quite obvious that the only way to solve full self-driving is to solve real-world AI and sophisticated vision." Source: Elon Musk TED Interview, April 2022

Mario Nawfal

56,936 Aufrufe • vor 1 Jahr

ELON MUSK: "Tesla full-self driving software, which is really just AI and cameras - we don't use radars or lidar or anything like that. It's really trying to drive the car in the same way that human drives the car, which is human drives the car primarily with vision and with a biological neural net, that we take the same approach with our vehicles, which is a digital neural net, and cameras, and we expect this approach to ultimately be at least an order of magnitude safer than humans driving. It is quite magical, because the car feels like it is sentient. It feels actually feels like it's alive. We already have some vehicles operating with no people inside and no safety monitors in three cities in Texas, and probably will be widespread in the US."

ELON MUSK: "Tesla full-self driving software, which is really just AI and cameras - we don't use radars or lidar or anything like that. It's really trying to drive the car in the same way that human drives the car, which is human drives the car primarily with vision and with a biological neural net, that we take the same approach with our vehicles, which is a digital neural net, and cameras, and we expect this approach to ultimately be at least an order of magnitude safer than humans driving. It is quite magical, because the car feels like it is sentient. It feels actually feels like it's alive. We already have some vehicles operating with no people inside and no safety monitors in three cities in Texas, and probably will be widespread in the US."

DogeDesigner

29,208 Aufrufe • vor 2 Monaten

Elon Musk in a new interview this morning on Tesla FSD: "In terms of having self-driving be ubiquitous, I think we’re making steady progress. The Tesla Full Self-Driving software, which is really just AI and cameras, we don’t use radars or LIDAR or anything like that. It’s really trying to drive the car in the same way that a human drives the car, primarily with vision and with a biological neural net. We take the same approach with our vehicles, which is a digital neural net and cameras. I expect this approach to ultimately be at least an order of magnitude safer than humans driving. It is quite magical because the car feels like it is sentient. It actually feels like it’s alive and you can actually, as we improve the software, feel the sentience growing in the car. We already have some vehicles operating with no people inside with no safety monitors in three cities in Texas and probably will be widespread in the US by the end of this year and hopefully in Israel too" Tesla FSD is beginning to feel more and more sentient with each update, it’s incredible. All without the use of Radar or LiDAR. It’s all about the 🧠.

Elon Musk in a new interview this morning on Tesla FSD: "In terms of having self-driving be ubiquitous, I think we’re making steady progress. The Tesla Full Self-Driving software, which is really just AI and cameras, we don’t use radars or LIDAR or anything like that. It’s really trying to drive the car in the same way that a human drives the car, primarily with vision and with a biological neural net. We take the same approach with our vehicles, which is a digital neural net and cameras. I expect this approach to ultimately be at least an order of magnitude safer than humans driving. It is quite magical because the car feels like it is sentient. It actually feels like it’s alive and you can actually, as we improve the software, feel the sentience growing in the car. We already have some vehicles operating with no people inside with no safety monitors in three cities in Texas and probably will be widespread in the US by the end of this year and hopefully in Israel too" Tesla FSD is beginning to feel more and more sentient with each update, it’s incredible. All without the use of Radar or LiDAR. It’s all about the 🧠.

Nic Cruz Patane

104,146 Aufrufe • vor 2 Monaten

To push self-driving into situations wilder than reality, we built a neural network world simulator that can create entirely synthetic worlds for the Tesla to drive in. Video below is fully generated & not a real video

To push self-driving into situations wilder than reality, we built a neural network world simulator that can create entirely synthetic worlds for the Tesla to drive in. Video below is fully generated & not a real video

Tesla

1,698,515 Aufrufe • vor 9 Monaten

NEWS: For the first time, Tesla has launched a whole page on its website dedicated just to FSD, and they revealed a couple new pieces of info on it! • 3.8 billion miles have bene driven on FSD (Supervised) • FSD (Supervised) is 54% safer than a human driver when it's engaged • FSD (Supervised) is trained on what amounts to over 100 years of anonymous real-world driving scenarios from Tesla's fleet of over six million vehicles. Tesla's fleet collectively experiences a lifetime of driving scenarios in 10 minutes. Link to the FSD page in the thread below.

NEWS: For the first time, Tesla has launched a whole page on its website dedicated just to FSD, and they revealed a couple new pieces of info on it! • 3.8 billion miles have bene driven on FSD (Supervised) • FSD (Supervised) is 54% safer than a human driver when it's engaged • FSD (Supervised) is trained on what amounts to over 100 years of anonymous real-world driving scenarios from Tesla's fleet of over six million vehicles. Tesla's fleet collectively experiences a lifetime of driving scenarios in 10 minutes. Link to the FSD page in the thread below.

Sawyer Merritt

367,796 Aufrufe • vor 1 Jahr

Cybertruck Tesla Energy FSD Supervised V12 went to wide release earlier this year We remain convinced that vision-based autonomy with end-to-end neural nets is the right architecture for scalable, profitable autonomy

Cybertruck Tesla Energy FSD Supervised V12 went to wide release earlier this year We remain convinced that vision-based autonomy with end-to-end neural nets is the right architecture for scalable, profitable autonomy

Tesla

146,368 Aufrufe • vor 2 Jahren

🧵 THREAD: Compilation of Tesla Robotaxi clips. As I compile this list, I have to restrain myself from going on Tesla 's website to buy one ☺️ 1. Tesla Robotaxi slows for guy getting arrested by police. Tesla's robotaxi, built on its Full Self-Driving (FSD) technology, primarily relies on a vision-based system using cameras, neural networks, and AI to sense and identify vehicles like police cars and school buses. Per WholeMarsBlog: "The more I use Robotaxi the more impressed I am." Credit: Whole Mars Catalog

🧵 THREAD: Compilation of Tesla Robotaxi clips. As I compile this list, I have to restrain myself from going on Tesla 's website to buy one ☺️ 1. Tesla Robotaxi slows for guy getting arrested by police. Tesla's robotaxi, built on its Full Self-Driving (FSD) technology, primarily relies on a vision-based system using cameras, neural networks, and AI to sense and identify vehicles like police cars and school buses. Per WholeMarsBlog: "The more I use Robotaxi the more impressed I am." Credit: Whole Mars Catalog

Barbie True Blue

110,669 Aufrufe • vor 1 Jahr

The greatest miscalculation the legacy auto industry ever made was doubting Elon's approach to Full Self-Driving For years, Wall Street aggressively shorted Tesla, and the legacy auto industry laughed. They confidently declared that Elon Musk’s approach to autonomy would fail Today, they are realizing they are completely stuck Here is how Elon solved the toughest problem in tech: The First Principles Approach: Elon admitted that FSD turned out to be exponentially harder than anyone expected. Because to solve self-driving, you aren't just writing code.....you basically have to recreate human biology in digital form Think about it: The entire global road system is designed to work for humans. We drive using optical sensors (our eyes) and a biological neural net (our brain) The Vision-Only Breakthrough: While legacy auto relied on expensive crutches like LiDAR, pre-mapped routes, and hard-coded rules, Elon looked at the real world. He knew that if an AI couldn't make decisions entirely on its own, it would eventually encounter edge cases and freeze Tesla took a path everyone deemed impossible: Vision-only, end-to-end neural networks The result was insane The industry that said "it will never work" is now scrambling. They completely misunderstood the problem, and now they are stuck a decade behind. Tesla has achieved what the rest of the automotive world couldn't even comprehend Today, most people using Tesla FSD are blown away by the way it makes a decision just like a human would All achieved with pure vision

The greatest miscalculation the legacy auto industry ever made was doubting Elon's approach to Full Self-Driving For years, Wall Street aggressively shorted Tesla, and the legacy auto industry laughed. They confidently declared that Elon Musk’s approach to autonomy would fail Today, they are realizing they are completely stuck Here is how Elon solved the toughest problem in tech: The First Principles Approach: Elon admitted that FSD turned out to be exponentially harder than anyone expected. Because to solve self-driving, you aren't just writing code.....you basically have to recreate human biology in digital form Think about it: The entire global road system is designed to work for humans. We drive using optical sensors (our eyes) and a biological neural net (our brain) The Vision-Only Breakthrough: While legacy auto relied on expensive crutches like LiDAR, pre-mapped routes, and hard-coded rules, Elon looked at the real world. He knew that if an AI couldn't make decisions entirely on its own, it would eventually encounter edge cases and freeze Tesla took a path everyone deemed impossible: Vision-only, end-to-end neural networks The result was insane The industry that said "it will never work" is now scrambling. They completely misunderstood the problem, and now they are stuck a decade behind. Tesla has achieved what the rest of the automotive world couldn't even comprehend Today, most people using Tesla FSD are blown away by the way it makes a decision just like a human would All achieved with pure vision

X Freeze

603,468 Aufrufe • vor 4 Monaten

It's wild that people post videos of themselves blatantly breaking the law like this. Look, obviously, the idea of being able to do something else, like work, while driving is extremely appealing, but you just can't do it like that. Tesla FSD v14 is extremely impressive, but the best data we have right now points to a critical intervention every ~2,000 miles on average. Human drivers crash about every ~400,000 miles on average. You will hurt yourself, or God forbid, someone else, if you keep doing this, as you won't be able to intervene properly. I know some don't believe because they blindly believe Tesla's claim that "FSD is 10x safer than human drivers", but that's the most misleading claim ever. First off, the data is cherry-picked. I do believe that driving on FSD is safer than driving without, but closer to a factor 3x based on the best, not-Tesla-massaged, data. But more importantly, that's 3x safer than FSD plus human. Not just FSD. It is safer because it is supervised by a human paying attention and correcting the system when needed. If you use it like this guy, you are not safer. Period.

It's wild that people post videos of themselves blatantly breaking the law like this. Look, obviously, the idea of being able to do something else, like work, while driving is extremely appealing, but you just can't do it like that. Tesla FSD v14 is extremely impressive, but the best data we have right now points to a critical intervention every ~2,000 miles on average. Human drivers crash about every ~400,000 miles on average. You will hurt yourself, or God forbid, someone else, if you keep doing this, as you won't be able to intervene properly. I know some don't believe because they blindly believe Tesla's claim that "FSD is 10x safer than human drivers", but that's the most misleading claim ever. First off, the data is cherry-picked. I do believe that driving on FSD is safer than driving without, but closer to a factor 3x based on the best, not-Tesla-massaged, data. But more importantly, that's 3x safer than FSD plus human. Not just FSD. It is safer because it is supervised by a human paying attention and correcting the system when needed. If you use it like this guy, you are not safer. Period.

Fred Lambert

12,092 Aufrufe • vor 1 Monat

Tesla's AI data flywheel 🚘♻️🤖 The Tesla fleet gathers more miles of driving data each week than Waymo has in its entire existence. It was a brilliant move to equip all cars with the full sensor stack (just cameras) long before FSD shipped. From 1:25 of

Tesla's AI data flywheel 🚘♻️🤖 The Tesla fleet gathers more miles of driving data each week than Waymo has in its entire existence. It was a brilliant move to equip all cars with the full sensor stack (just cameras) long before FSD shipped. From 1:25 of

Steve Jurvetson

114,962 Aufrufe • vor 11 Monaten

Bro… it’s 2026 and you’re still driving manually? I just press a button, and I don’t even remember how I got to my destination. If it was legal to do so, I’d be 100% comfortable closing my eyes and taking a nap the entire drive. It’s just the honest truth, I’m not hyping anything. Tesla FSD is really that good now. Tesla FSD has now driven more than 8 billion real world miles, including more than 3 billion city miles. In Tesla’s latest safety data, FSD logged about 1 major collision every 5.3 million miles, which Tesla says is about 7x fewer major collisions than the U.S. average. On top of this, Tesla FSD delivers 7x fewer minor collisions and 5x fewer off-highway collisions than the average. It’s significantly safer than anything out there. This AI system is learning from millions of Teslas out in the real world, not from geofenced areas, like literally anywhere a Tesla can drive. Tesla’s vehicles use 360 degree camera visibility, over-the-air updates, and has a neural net that keeps making the system smarter, safer, and better over time, for all Teslas. The FSD experience is so good now that manually driving already feels like the distant past. And once you’ve lived with what so many Tesla owners get to experience every day, going back to any other car feels almost impossible.

Bro… it’s 2026 and you’re still driving manually? I just press a button, and I don’t even remember how I got to my destination. If it was legal to do so, I’d be 100% comfortable closing my eyes and taking a nap the entire drive. It’s just the honest truth, I’m not hyping anything. Tesla FSD is really that good now. Tesla FSD has now driven more than 8 billion real world miles, including more than 3 billion city miles. In Tesla’s latest safety data, FSD logged about 1 major collision every 5.3 million miles, which Tesla says is about 7x fewer major collisions than the U.S. average. On top of this, Tesla FSD delivers 7x fewer minor collisions and 5x fewer off-highway collisions than the average. It’s significantly safer than anything out there. This AI system is learning from millions of Teslas out in the real world, not from geofenced areas, like literally anywhere a Tesla can drive. Tesla’s vehicles use 360 degree camera visibility, over-the-air updates, and has a neural net that keeps making the system smarter, safer, and better over time, for all Teslas. The FSD experience is so good now that manually driving already feels like the distant past. And once you’ve lived with what so many Tesla owners get to experience every day, going back to any other car feels almost impossible.

Teslaconomics

16,734 Aufrufe • vor 4 Monaten

My Tesla drives me anywhere with the press of a button on the screen. So many people have no idea, but autonomy is here. If you have a Tesla, you don't need to drive if you don't want to. The data already proves that Tesla FSD is 8x safer than humans at driving. Eight cameras and the best real-world Al are watching the roads to keep you safe at all times. It can even predict human driver movements before you do, almost like it sees the future. It's so smart. Driving in the future will essentially just become a niche thing.

My Tesla drives me anywhere with the press of a button on the screen. So many people have no idea, but autonomy is here. If you have a Tesla, you don't need to drive if you don't want to. The data already proves that Tesla FSD is 8x safer than humans at driving. Eight cameras and the best real-world Al are watching the roads to keep you safe at all times. It can even predict human driver movements before you do, almost like it sees the future. It's so smart. Driving in the future will essentially just become a niche thing.

Nic Cruz Patane

11,059 Aufrufe • vor 2 Monaten

Here’s my first drive on Tesla FSD V14.2. For the first time on V14, I had zero hesitation or braking issues all day, and no interventions. From a driving perspective, and after 85 miles on and off-camera on V14.2 today, it now feels as comfortable, if not more so, than V13 did, but with all the capabilities of V14. What’s new in FSD V14.2: • Self-Driving stats section that tells you exactly how many Self-Driving miles you’ve driven and the percentage of FSD miles • You can now select speed profiles and parking preferences while in park • Upgraded the neural network vision encoder, leveraging higher resolution features to further improve scenarios like handling emergency vehicles, obstacles on the road, and human gestures • Big improvement (in my experience) to overall everyday driving comfortability • Slightly less strict attention monitoring at times In my opinion, V14.2 is good enough to be sent to every AI4 owner where FSD is legal, and for Tesla to encourage non-FSD owners/subscribers to try it. FSD did have a couple weird navigation decisions it made in my video, but that’s my only complaint. Overall, great version of FSD. 4K YouTube version of my video in thread below:

Here’s my first drive on Tesla FSD V14.2. For the first time on V14, I had zero hesitation or braking issues all day, and no interventions. From a driving perspective, and after 85 miles on and off-camera on V14.2 today, it now feels as comfortable, if not more so, than V13 did, but with all the capabilities of V14. What’s new in FSD V14.2: • Self-Driving stats section that tells you exactly how many Self-Driving miles you’ve driven and the percentage of FSD miles • You can now select speed profiles and parking preferences while in park • Upgraded the neural network vision encoder, leveraging higher resolution features to further improve scenarios like handling emergency vehicles, obstacles on the road, and human gestures • Big improvement (in my experience) to overall everyday driving comfortability • Slightly less strict attention monitoring at times In my opinion, V14.2 is good enough to be sent to every AI4 owner where FSD is legal, and for Tesla to encourage non-FSD owners/subscribers to try it. FSD did have a couple weird navigation decisions it made in my video, but that’s my only complaint. Overall, great version of FSD. 4K YouTube version of my video in thread below:

Sawyer Merritt

144,763 Aufrufe • vor 8 Monaten

Introducing Continuous Thought Machines New Blog: Modern AI is powerful, but it’s still distinct from human-like flexible intelligence. We believe neural timing is key. Our Continuous Thought Machine is built from the ground up to use neural dynamics as a powerful representation for intelligence. Thought takes time, and reasoning is a process. Biological brains inspire us with their complex neural activity, where neural timing is critical to intelligence. We’re exploring how to bring that power to AI. The Continuous Thought Machine (CTM) incorporates neuron-level temporal processing and neural synchronization, moving beyond current AI limitations. Our approach has two core innovations: (1) neuron-level temporal processing, where each neuron uses unique parameters to process a history of incoming signals for fine-grained temporal dynamics, and (2) neural synchronization, used as a direct latent representation to modulate data and produce outputs, encoding information directly in the timing of neural activity. Learn more about our approach: Interactive Report: Full Paper: GitHub :

Introducing Continuous Thought Machines New Blog: Modern AI is powerful, but it’s still distinct from human-like flexible intelligence. We believe neural timing is key. Our Continuous Thought Machine is built from the ground up to use neural dynamics as a powerful representation for intelligence. Thought takes time, and reasoning is a process. Biological brains inspire us with their complex neural activity, where neural timing is critical to intelligence. We’re exploring how to bring that power to AI. The Continuous Thought Machine (CTM) incorporates neuron-level temporal processing and neural synchronization, moving beyond current AI limitations. Our approach has two core innovations: (1) neuron-level temporal processing, where each neuron uses unique parameters to process a history of incoming signals for fine-grained temporal dynamics, and (2) neural synchronization, used as a direct latent representation to modulate data and produce outputs, encoding information directly in the timing of neural activity. Learn more about our approach: Interactive Report: Full Paper: GitHub :

Sakana AI

289,920 Aufrufe • vor 1 Jahr

Elon Musk: "I thought the self-driving problem would be hard, but it was harder than I thought. It's not like I thought it would be easy, I thought it would be very hard, but it was actually way harder than even that. I mean, what it comes down to at the end of the day is to solve self-driving, you basically need to recreate what humans do to drive, which is: humans drive with optical sensors (eyes) and biological neural nets. That's how the entire road system is designed to work, So for actually for full self-driving to work, we have to recreate that in digital form."

Elon Musk: "I thought the self-driving problem would be hard, but it was harder than I thought. It's not like I thought it would be easy, I thought it would be very hard, but it was actually way harder than even that. I mean, what it comes down to at the end of the day is to solve self-driving, you basically need to recreate what humans do to drive, which is: humans drive with optical sensors (eyes) and biological neural nets. That's how the entire road system is designed to work, So for actually for full self-driving to work, we have to recreate that in digital form."

Mars University

120,160 Aufrufe • vor 4 Monaten

Does GPT understand the world? Here is what Ilya Sutskever, co-founder of OpenAI, says during a discussion with Jensen Huang, CEO of Nvidia: (1) When we train a large neural network to accurately predict the next word in lots of different texts from the internet, the AI is learning a world model. (2) On the surface, it may look like learning correlations in text, but it turns out that to 'just learn' statistical correlations in text, to compress information really well, what the neural network learns is some representation of the process that produced the text. (3) This text is a projection of the world...what the neural network is learning is aspects of the world, of people, of the human conditions, their hopes, dreams, motivations, their interactions...the situations we are in. The neural network learns a compressed, abstract, usable representation." Do you think learning representations = understanding? Are large language models simply stochastic parrots, or are they much more?

Does GPT understand the world? Here is what Ilya Sutskever, co-founder of OpenAI, says during a discussion with Jensen Huang, CEO of Nvidia: (1) When we train a large neural network to accurately predict the next word in lots of different texts from the internet, the AI is learning a world model. (2) On the surface, it may look like learning correlations in text, but it turns out that to 'just learn' statistical correlations in text, to compress information really well, what the neural network learns is some representation of the process that produced the text. (3) This text is a projection of the world...what the neural network is learning is aspects of the world, of people, of the human conditions, their hopes, dreams, motivations, their interactions...the situations we are in. The neural network learns a compressed, abstract, usable representation." Do you think learning representations = understanding? Are large language models simply stochastic parrots, or are they much more?

Alex Ker 🔭

1,367,097 Aufrufe • vor 2 Jahren