正在加载视频...

视频加载失败

加载此视频时出现问题。这可能是由于临时网络问题，或视频可能不可用。

On a chip, even something as basic as selecting one input from a register file has to be built out of AND and OR gates wired together. Simple-seeming operations end up surprisingly expensive. Reiner Pope showed me how this stacks up. In older generations of chips, most of the... show more

Dwarkesh Patel

234,334 subscribers

27,876 次观看 • 1 个月前 •via X (Twitter)

教育新闻政治科学技术

Anya Rossi• Live Now

Private livecam show

0 条评论

暂无评论

原始帖子的评论将显示在这里

相关视频

Billions of times a second, all the circuitry on an AI chip pauses, just for a moment. Why? Reiner Pope explained to me what's going on, and what the clock cycle of a chip is actually telling us:

Billions of times a second, all the circuitry on an AI chip pauses, just for a moment. Why? Reiner Pope explained to me what's going on, and what the clock cycle of a chip is actually telling us:

Dwarkesh Patel

28,938 次观看 • 1 个月前

.Reiner Pope explains the difference between a CPU and a GPU. A huge chunk of a CPU die is actually just predicting what code will run next. GPUs strip this out, which is part of why they can fit so much more compute on the chip.

.Reiner Pope explains the difference between a CPU and a GPU. A huge chunk of a CPU die is actually just predicting what code will run next. GPUs strip this out, which is part of why they can fit so much more compute on the chip.

Dwarkesh Patel

62,507 次观看 • 1 个月前

Watch this video of one of the largest data centres on earth, under construction in Texas…. The video shows the cooling system used for the Ai GPU CHIPS…. It’s made out of pure copper! Copper is used in data centres to transfer heat as well as electricity.

Watch this video of one of the largest data centres on earth, under construction in Texas…. The video shows the cooling system used for the Ai GPU CHIPS…. It’s made out of pure copper! Copper is used in data centres to transfer heat as well as electricity.

Robert Friedland

76,966 次观看 • 5 个月前

Food quality control standards in America seem to be plummeting American bought a bad of Lays Pickle flavored potato chips. In the bag at first appears to be a large chip, it’s not a large chip. It seems to potentially be apart of the ceiling from the factory or something else “I’m gonna throw up”

Food quality control standards in America seem to be plummeting American bought a bad of Lays Pickle flavored potato chips. In the bag at first appears to be a large chip, it’s not a large chip. It seems to potentially be apart of the ceiling from the factory or something else “I’m gonna throw up”

Wall Street Apes

60,561 次观看 • 10 个月前

Chips on shoulders put chips in pockets. How Adam Foroughi, founder of Applovin, built one of the most efficient companies by hiring outcasts with something to prove: “One of the things that I always looked at for hiring was to find people who have a chip on their shoulder. Someone who has a reason to push hard. We ended up with this group of people that were outcasts from other places. One of the best hires I ever made was this kid who was a high school dropout. He was not getting a job in big tech. But he had skills that smoke everyone else when it comes to selling. I knew I could bring that person in, empower that person, and they're going to be hungry and they're going to push harder than the others.”

Chips on shoulders put chips in pockets. How Adam Foroughi, founder of Applovin, built one of the most efficient companies by hiring outcasts with something to prove: “One of the things that I always looked at for hiring was to find people who have a chip on their shoulder. Someone who has a reason to push hard. We ended up with this group of people that were outcasts from other places. One of the best hires I ever made was this kid who was a high school dropout. He was not getting a job in big tech. But he had skills that smoke everyone else when it comes to selling. I knew I could bring that person in, empower that person, and they're going to be hungry and they're going to push harder than the others.”

David Senra

195,981 次观看 • 2 个月前

If we lived in a normal world and normal Country, one where we call benefit from exploiting our resources a pipeline would just be built. Who is going to invest in a project that has this level of hoops to jump through in order for it to happen? This has been set up to fail from day one. When the leader of your Country is so visibly uncomfortable talking about something as simple as a pipeline, it’s not happening.

If we lived in a normal world and normal Country, one where we call benefit from exploiting our resources a pipeline would just be built. Who is going to invest in a project that has this level of hoops to jump through in order for it to happen? This has been set up to fail from day one. When the leader of your Country is so visibly uncomfortable talking about something as simple as a pipeline, it’s not happening.

Ryan Gerritsen🇨🇦🇳🇱

18,879 次观看 • 1 个月前

The realization over the last two extensions has come home to roost. There is not going to be a purchase of TikTok with the Chinese algorithm. We know that we can't buy the algorithm so we're gonna have to rewrite this whole thing, but while we're doing that, we're gonna try and solve for all the litigation that social media gets all stemming from the fact that the platform owns the data and the customer doesn't. I think we're gonna have to fix that. So our proposal is to fix that and not use the Chinese algorithm. When you issue a chip, let's say you take an NVIDIA chip, that chip is a queen bee. It sits in the center of the hive and every programmer that uses that chip is the honeybee. It spends millions of hours training on that chip, learning that chip, writing code for that chip and developing AI for that chip. That's the honey. There's only gonna be two kinds of honey on Earth. American or Chinese. Now if you want the world to be running on Chinese honey, stop shipping American queen bee chips, and that is what is happening right now.

The realization over the last two extensions has come home to roost. There is not going to be a purchase of TikTok with the Chinese algorithm. We know that we can't buy the algorithm so we're gonna have to rewrite this whole thing, but while we're doing that, we're gonna try and solve for all the litigation that social media gets all stemming from the fact that the platform owns the data and the customer doesn't. I think we're gonna have to fix that. So our proposal is to fix that and not use the Chinese algorithm. When you issue a chip, let's say you take an NVIDIA chip, that chip is a queen bee. It sits in the center of the hive and every programmer that uses that chip is the honeybee. It spends millions of hours training on that chip, learning that chip, writing code for that chip and developing AI for that chip. That's the honey. There's only gonna be two kinds of honey on Earth. American or Chinese. Now if you want the world to be running on Chinese honey, stop shipping American queen bee chips, and that is what is happening right now.

Kevin O'Leary aka Mr. Wonderful

43,532 次观看 • 1 年前

Many still don’t understand why Elon is building Terafab Terafab is an extension to all the chip makers in the world It’s not about replacement, not a rivalry and absolutely not competing It’s being built to fulfill the massive chip orders that Tesla, SpaceX and xAI actually need TSMC’s most advanced 2nm capacity is totally booked through 2028 Tesla signed a massive $16.5 billion deal with Samsung back in July 2025 to produce AI6 chips at their Taylor, Texas factory and Samsung is building a Tesla Exclusive chip manufacturing plant to full fill this orders When Elon announced Terafab on March 21, 2026...he made it clear: “That rate is much less than we’d like. We either build the Terafab or we don’t have the chips, and we need the chips, so we build the Terafab” He basically told the chip makers: “Produce as much as you comfortably can. We will take them all. Actually we want even more” Even today Elon said: "SpaceX/Tesla will be always be major customers of TSMC and not competitors in the normal sense of the word" Current production rates are much less than they need....That’s why Terafab exists Terafab is an extension to every chip maker… not competition, not rivalry, absolutely not Even Intel has joined as a partner Even if chip supply improves, massive bottlenecks still exist with memory and advanced packaging You simply can't risk those supply chain breaks at this scale That’s why Terafab is being built to vertically integrate everything - chips, memory, advanced packaging all under one roof, targeting 1 terawatt of AI compute capacity per year This is a ludicrous amount of chips that no chipmaker currently produces at this scale. I don't think even TSMC and Samsung truly understand these numbers yet It’s about building the capacity the future actually demands

Many still don’t understand why Elon is building Terafab Terafab is an extension to all the chip makers in the world It’s not about replacement, not a rivalry and absolutely not competing It’s being built to fulfill the massive chip orders that Tesla, SpaceX and xAI actually need TSMC’s most advanced 2nm capacity is totally booked through 2028 Tesla signed a massive $16.5 billion deal with Samsung back in July 2025 to produce AI6 chips at their Taylor, Texas factory and Samsung is building a Tesla Exclusive chip manufacturing plant to full fill this orders When Elon announced Terafab on March 21, 2026...he made it clear: “That rate is much less than we’d like. We either build the Terafab or we don’t have the chips, and we need the chips, so we build the Terafab” He basically told the chip makers: “Produce as much as you comfortably can. We will take them all. Actually we want even more” Even today Elon said: "SpaceX/Tesla will be always be major customers of TSMC and not competitors in the normal sense of the word" Current production rates are much less than they need....That’s why Terafab exists Terafab is an extension to every chip maker… not competition, not rivalry, absolutely not Even Intel has joined as a partner Even if chip supply improves, massive bottlenecks still exist with memory and advanced packaging You simply can't risk those supply chain breaks at this scale That’s why Terafab is being built to vertically integrate everything - chips, memory, advanced packaging all under one roof, targeting 1 terawatt of AI compute capacity per year This is a ludicrous amount of chips that no chipmaker currently produces at this scale. I don't think even TSMC and Samsung truly understand these numbers yet It’s about building the capacity the future actually demands

X Freeze

53,291 次观看 • 2 个月前

.Visibl Semiconductors is building the first coordination layer for chip design. As teams scale, workstreams drift out of sync, issues surface late and drive $XM+ delays and months of rework before tapeout. Visibl is building end-to-end AI to shrink chip design from years to weeks, starting where intent breaks first: coordination. Book a demo at Congrats on the launch Bryce N, and Jordon Kashanchi!

.Visibl Semiconductors is building the first coordination layer for chip design. As teams scale, workstreams drift out of sync, issues surface late and drive $XM+ delays and months of rework before tapeout. Visibl is building end-to-end AI to shrink chip design from years to weeks, starting where intent breaks first: coordination. Book a demo at Congrats on the launch Bryce N, and Jordon Kashanchi!

Y Combinator

22,189 次观看 • 4 个月前

On safari, we woke up from a nap and noticed something moving outside our window as the sun was setting Check out the videos. This was one of the most magical moments of my life

On safari, we woke up from a nap and noticed something moving outside our window as the sun was setting Check out the videos. This was one of the most magical moments of my life

Ramit Sethi

32,187 次观看 • 11 个月前

Lol at these people. One of the oldest image manipulation effects, morphing (Beier-Neely algorithm), which had been around for 33 years at this point, this guy tried to replicate using AI (probably consuming as much as the electricity of my hometown in a day) and it's not even as smooth. This is what I mean when I say people that use AI have no idea even how to assess whether what they made is even good. This is supposed to be such a simple effect to get, it's like the ABC of visual effects. Anyway the one on the left was manipulated by Pacific Data Imagery, responsible for special effects for Terminator 2

Lol at these people. One of the oldest image manipulation effects, morphing (Beier-Neely algorithm), which had been around for 33 years at this point, this guy tried to replicate using AI (probably consuming as much as the electricity of my hometown in a day) and it's not even as smooth. This is what I mean when I say people that use AI have no idea even how to assess whether what they made is even good. This is supposed to be such a simple effect to get, it's like the ABC of visual effects. Anyway the one on the left was manipulated by Pacific Data Imagery, responsible for special effects for Terminator 2

VIKARE

13,235 次观看 • 1 年前

The lack of "feature design" is why so many products over time feel hollow or messy. This isn't visual design. This isn't architectural design. I thought that a short video lecture of what feature design is and a real case study of applying it in Ghostty would be helpful. Feature design is the planning step behind how you're going to solve one or more user problems with a product feature: what that feature looks like, how it feels, and not just how its going to tactically solve these specific problems, but how that solution is going to interface with the edges of other features that currently exist or are planned to exist in the future.

The lack of "feature design" is why so many products over time feel hollow or messy. This isn't visual design. This isn't architectural design. I thought that a short video lecture of what feature design is and a real case study of applying it in Ghostty would be helpful. Feature design is the planning step behind how you're going to solve one or more user problems with a product feature: what that feature looks like, how it feels, and not just how its going to tactically solve these specific problems, but how that solution is going to interface with the edges of other features that currently exist or are planned to exist in the future.

Mitchell Hashimoto

155,378 次观看 • 6 个月前

When it mattered most, we showed up. As Chair of the Nairobi Branch, I rallied fellow advocates to stand together for the profession and for one of us. It was about taking action

When it mattered most, we showed up. As Chair of the Nairobi Branch, I rallied fellow advocates to stand together for the profession and for one of us. It was about taking action

Charles Kanjama SC

77,305 次观看 • 5 个月前

Pope Leo challenges students participating in the Jubilee of the World of Education to work for a better society through education, which is “one of the most beautiful and powerful tools for changing the world.” “How wonderful it would be if one day your generation were remembered as the ‘generation plus,’ remembered for the extra drive you brought to the Church and the world,” Pope Leo said. But he also warned against keeping this desire a dream. The Pope explained the answer to achieving this is through education, “one of the most beautiful and powerful tools for changing the world.”

Pope Leo challenges students participating in the Jubilee of the World of Education to work for a better society through education, which is “one of the most beautiful and powerful tools for changing the world.” “How wonderful it would be if one day your generation were remembered as the ‘generation plus,’ remembered for the extra drive you brought to the Church and the world,” Pope Leo said. But he also warned against keeping this desire a dream. The Pope explained the answer to achieving this is through education, “one of the most beautiful and powerful tools for changing the world.”

Vatican News

20,021 次观看 • 8 个月前

Our goal is to make the best chips physically possible for the large model needs of frontier labs. The MatX One chip delivers higher throughput than any announced product while also achieving the lowest latency. Reiner Pope joined Bloomberg TV this morning to discuss:

Our goal is to make the best chips physically possible for the large model needs of frontier labs. The MatX One chip delivers higher throughput than any announced product while also achieving the lowest latency. Reiner Pope joined Bloomberg TV this morning to discuss:

MatX

40,049 次观看 • 4 个月前

One of the best high level explanations of what the actual goal of Trump’s tariffs are, which is to be used as a bargaining chip against other countries who charge tariffs on our imports. It’s the way to get back to being a country that manufactures and exports items again.

One of the best high level explanations of what the actual goal of Trump’s tariffs are, which is to be used as a bargaining chip against other countries who charge tariffs on our imports. It’s the way to get back to being a country that manufactures and exports items again.

Insurrection Barbie

91,647 次观看 • 1 年前

My enemy AI design philosophy is to make them follow as few and as simple rules as possible. Just a handful of raycasts to look for player and to get the most basic info about the collision around them.

My enemy AI design philosophy is to make them follow as few and as simple rules as possible. Just a handful of raycasts to look for player and to get the most basic info about the collision around them.

Arthur Vyater

38,818 次观看 • 1 年前

ELON MUSK: We believe the AI5 chip will be roughly comparable performance to an NVIDIA Blackwell, and at much less than 10% of the cost Transcription: I'm super hardcore on chips right now as you may be able to tell. I have chips on the brain. I dream about chips, Literally! Because in order to have a functional robot, you have to have a great AI chip. And it needs to be an inexpensive chip and it needs to be very power efficient So we think we believe the AI5 chip will be probably about a third of the power of say something like a Blackwell, an NVIDIA Blackwell, which is a great chip, for roughly comparable performance. And much less than 10% of the cost. This is a chip that is very much optimized for the Tesla AI software stack. So it's not meant to be a general purpose chip, it's meant to be an amazing chip for the Tesla AI software And I mean a couple of things that I think make... like how is Tesla able to achieve such an improvement? I think it is because we are specialized. We're not trying to... you know, NVIDIA has to serve the superset of all past and future customers. So all of their requirements, all of the software that they've written has to work, which is a very difficult problem. Whereas we just need to make it work for our software. And so we're able to simplify the chip dramatically And then we also, I think we're unique in this, but like we have an integer-based system. And integer operations are fundamentally more efficient than floating point operations. So we can do floating point, but the vast majority of our inference is done in integer. Which is, if you're familiar with sort of logic gates, the simplicity of integer... it's integer is much more power efficient, much more silicon efficient, but you have to, you actually have to train for integer inference, which everyone else is training for floating point. That's kind of like a niche technical detail, but it's actually very important. So, yeah, this is going to be a great chip So this chip will be made in basically in four places: TSMC Taiwan, Samsung Korea, TSMC Arizona, and TSMC Texas. And we already know what improvements to make for AI6. So I'm hopeful that we can within less than a year of AI5 starting production, we can actually transition in the same fab to AI6 and double all of the performance metrics

ELON MUSK: We believe the AI5 chip will be roughly comparable performance to an NVIDIA Blackwell, and at much less than 10% of the cost Transcription: I'm super hardcore on chips right now as you may be able to tell. I have chips on the brain. I dream about chips, Literally! Because in order to have a functional robot, you have to have a great AI chip. And it needs to be an inexpensive chip and it needs to be very power efficient So we think we believe the AI5 chip will be probably about a third of the power of say something like a Blackwell, an NVIDIA Blackwell, which is a great chip, for roughly comparable performance. And much less than 10% of the cost. This is a chip that is very much optimized for the Tesla AI software stack. So it's not meant to be a general purpose chip, it's meant to be an amazing chip for the Tesla AI software And I mean a couple of things that I think make... like how is Tesla able to achieve such an improvement? I think it is because we are specialized. We're not trying to... you know, NVIDIA has to serve the superset of all past and future customers. So all of their requirements, all of the software that they've written has to work, which is a very difficult problem. Whereas we just need to make it work for our software. And so we're able to simplify the chip dramatically And then we also, I think we're unique in this, but like we have an integer-based system. And integer operations are fundamentally more efficient than floating point operations. So we can do floating point, but the vast majority of our inference is done in integer. Which is, if you're familiar with sort of logic gates, the simplicity of integer... it's integer is much more power efficient, much more silicon efficient, but you have to, you actually have to train for integer inference, which everyone else is training for floating point. That's kind of like a niche technical detail, but it's actually very important. So, yeah, this is going to be a great chip So this chip will be made in basically in four places: TSMC Taiwan, Samsung Korea, TSMC Arizona, and TSMC Texas. And we already know what improvements to make for AI6. So I'm hopeful that we can within less than a year of AI5 starting production, we can actually transition in the same fab to AI6 and double all of the performance metrics

X Freeze

305,091 次观看 • 8 个月前

One of the most extreme Ferraris ever built. This F40 LM is one of just 19 factory racing variants developed by Michelotto, and it's heading to Monterey 2025. Raw, analog, and built for the track, this is the F40 turned all the way up.⁠ Register to bid at

One of the most extreme Ferraris ever built. This F40 LM is one of just 19 factory racing variants developed by Michelotto, and it's heading to Monterey 2025. Raw, analog, and built for the track, this is the F40 turned all the way up.⁠ Register to bid at

RM Sotheby's

12,322 次观看 • 11 个月前

Ever wondered how we come up with combos with new characters and new moves? It’s all in the frame data! Frame data opens up a new perspective to the game. This is one of the most basic levels of frame data 101. Check it out! #streetfighter #streetfighter6 #kimberly #sf6

Ever wondered how we come up with combos with new characters and new moves? It’s all in the frame data! Frame data opens up a new perspective to the game. This is one of the most basic levels of frame data 101. Check it out! #streetfighter #streetfighter6 #kimberly #sf6

Sherry Nhan

109,126 次观看 • 3 年前