Video yükleniyor...

Video Yüklenemedi

Bu video yüklenirken bir sorun oluştu. Bu geçici bir ağ sorunundan kaynaklanıyor olabilir veya video kullanılamıyor olabilir.

Ana Sayfaya Dön

Narrative violation from Dylan on Dwarkesh: H100s are worth more today than they were 3 years ago. There's a sentiment that data center buildouts are priced into the risk of rapidly depreciating GPUs. But the models want to learn. Token prices are falling so fast that you can get... show more

Sarah Wang

16,813 subscribers

159,395 görüntüleme • 4 ay önce •via X (Twitter)

Bilim & Teknoloji Finans

Anya Rossi• Live Now

Private livecam show

0 Yorum

Yorum bulunmuyor

Orijinal gönderinin yorumları burada görünecek

Benzer Videolar

The value produced by models is getting so much better so fast that old hardware is actually getting *more* expensive to rent. 3 years ago, the best model you could run on a H100 chip was GPT-4. Now, you can run GPT-5.4 on it, which is smaller and cheaper to run while producing much more valuable tokens. w. Dylan Patel

The value produced by models is getting so much better so fast that old hardware is actually getting more expensive to rent. 3 years ago, the best model you could run on a H100 chip was GPT-4. Now, you can run GPT-5.4 on it, which is smaller and cheaper to run while producing much more valuable tokens. w. Dylan Patel

Dwarkesh Patel

77,908 görüntüleme • 3 ay önce

SemiAnalysis's Dylan Patel says AI models will improve faster in the next 6 month to a year than we saw in the past year because there's a new axis of scale that has been unlocked in the form of synthetic data generation, that we are still very early in scaling up

SemiAnalysis's Dylan Patel says AI models will improve faster in the next 6 month to a year than we saw in the past year because there's a new axis of scale that has been unlocked in the form of synthetic data generation, that we are still very early in scaling up

Tsarathustra

174,827 görüntüleme • 1 yıl önce

Tapper: Do you agree that the US is at risk of facing a recession? Bessent: I believe we are in transition period... I think that we're in good shape. Oh really? More Americans are struggling to get by today than they were a year ago.

Tapper: Do you agree that the US is at risk of facing a recession? Bessent: I believe we are in transition period... I think that we're in good shape. Oh really? More Americans are struggling to get by today than they were a year ago.

Ron Smith

38,103 görüntüleme • 8 ay önce

Sam Altman says the leap from GPT-4 to GPT-5 will be as big as that of GPT-3 to 4 and the plan is to integrate the GPT and o series of models into one model that can do everything

Sam Altman says the leap from GPT-4 to GPT-5 will be as big as that of GPT-3 to 4 and the plan is to integrate the GPT and o series of models into one model that can do everything

Tsarathustra

224,102 görüntüleme • 1 yıl önce

.Secretary Chris Wright: “This administration and the President's actions have done a lot to change the trajectory on the energy prices going forward. Most energy prices, actually, that consumers see actually are lower today than they were a year ago, and they were lower today than they were at inauguration.”

.Secretary Chris Wright: “This administration and the President's actions have done a lot to change the trajectory on the energy prices going forward. Most energy prices, actually, that consumers see actually are lower today than they were a year ago, and they were lower today than they were at inauguration.”

U.S. Department of Energy

14,567 görüntüleme • 1 yıl önce

Today, we expose Old World Maps and you don't want to miss the Old World UNDERGROUND Cities that we are going to expose today. A place that will change the way that we all think about the ground that we stand on. We find out that these Tunnel Systems are much more than we were ever told. They are not just hallways, and we expose this to the world today. We are told that these are found by a chicken… Let’s show the world that this place that we live, is so much more than we are taught.

Today, we expose Old World Maps and you don't want to miss the Old World UNDERGROUND Cities that we are going to expose today. A place that will change the way that we all think about the ground that we stand on. We find out that these Tunnel Systems are much more than we were ever told. They are not just hallways, and we expose this to the world today. We are told that these are found by a chicken… Let’s show the world that this place that we live, is so much more than we are taught.

MY LUNCH BREAK

27,329 görüntüleme • 11 ay önce

Sam Altman says the leap from GPT-4 to GPT-5 will be as big as that of GPT-3 to 4 and the plan is to integrate the GPT and o series of models into one model that can do everything - that’s the AGI

Sam Altman says the leap from GPT-4 to GPT-5 will be as big as that of GPT-3 to 4 and the plan is to integrate the GPT and o series of models into one model that can do everything - that’s the AGI

Chubby♨️

231,782 görüntüleme • 1 yıl önce

This is a reminder that Indigenous homelands are more than just a weekend trip this #LaborDay, they're the homes of our ancestors & the Native peoples that are still here today. We encourage you to learn more & find out whose lands you’re on:

This is a reminder that Indigenous homelands are more than just a weekend trip this #LaborDay, they're the homes of our ancestors & the Native peoples that are still here today. We encourage you to learn more & find out whose lands you’re on:

IllumiNative

11,215 görüntüleme • 2 yıl önce

At Every, all employees use agents + have unlimited token budgets. Since GPT-3, they've increased headcount from 4 to 30. "If you're actually on the frontier like we are," Dan Shipper 📧 said, "it seems like there's more human work to do than ever."

At Every, all employees use agents + have unlimited token budgets. Since GPT-3, they've increased headcount from 4 to 30. "If you're actually on the frontier like we are," Dan Shipper 📧 said, "it seems like there's more human work to do than ever."

TBPN

55,494 görüntüleme • 1 ay önce

Base models are insanely powerful, and far more creative than instruction-tuned models like GPT-4 and Claude. But few people have actually had a chance to use them. That changes today. I've built a playground that allows you to interact with Llama 3.1 405B Base:

Base models are insanely powerful, and far more creative than instruction-tuned models like GPT-4 and Claude. But few people have actually had a chance to use them. That changes today. I've built a playground that allows you to interact with Llama 3.1 405B Base:

Matt Shumer

117,782 görüntüleme • 1 yıl önce

Every year someone names a new bottleneck for AI compute scaling. Dylan Patel on why power isn't gonna be the big one over the next few years: fundamentally there's many different ways to generate power (rather than just one company that can produce the EUV tools needed for the chips themselves) and the supply chains are simpler and easier to ramp. You can do jet engines bolted to the ground. Ship engines. Diesel recips from auto manufacturers with declining volumes. Fuel cells. Each category alone delivers tens of gigawatts by end of decade. Combined, hundreds. Even if energy costs double, a GPU goes from $1.40/hr to $1.50/hr. Nobody notices a dime when the models are improving so fast the value dwarfs the cost. Even if you don't add more power, but simply add more batteries, you can unlock 20% more of the US's terawatt scale power grid. This is because grid utilities want to make sure they're sized for peak summer load that hits a few hours a year. With enough batteries, you can make this guarantee, even without turning on more power plants! Fundamentally, there's a lot of different ways to bring power online over the next few years. Building more logic and memory is far more difficult and centralized, so that's where Dylan thinks the bottleneck will be.

Every year someone names a new bottleneck for AI compute scaling. Dylan Patel on why power isn't gonna be the big one over the next few years: fundamentally there's many different ways to generate power (rather than just one company that can produce the EUV tools needed for the chips themselves) and the supply chains are simpler and easier to ramp. You can do jet engines bolted to the ground. Ship engines. Diesel recips from auto manufacturers with declining volumes. Fuel cells. Each category alone delivers tens of gigawatts by end of decade. Combined, hundreds. Even if energy costs double, a GPU goes from $1.40/hr to $1.50/hr. Nobody notices a dime when the models are improving so fast the value dwarfs the cost. Even if you don't add more power, but simply add more batteries, you can unlock 20% more of the US's terawatt scale power grid. This is because grid utilities want to make sure they're sized for peak summer load that hits a few hours a year. With enough batteries, you can make this guarantee, even without turning on more power plants! Fundamentally, there's a lot of different ways to bring power online over the next few years. Building more logic and memory is far more difficult and centralized, so that's where Dylan thinks the bottleneck will be.

Dwarkesh Patel

214,172 görüntüleme • 3 ay önce

"Compute is the currency of the future." "Compute is the new OIL." Great, so how do I invest? You can play with the NVIDIA stock, or spin up an AI-Ready Data Center. Challenges of Spinning up AI-Ready Compute: Cost: Even a small cluster of H100 machines is upward to $10M GPU Scarcity: Unallocated or “underutilized” H100, A100, and RTX4090 GPUs are almost non-existent. Very Specific Physical & Technical Requirements: A data center with the right infrastructure is also very rare. There are only 10 in North America that deploy thousands of GPUs, in the open market. Domain Expertise: Teams of experts with the years of experience required to operate, deploy, and manage AI compute at scale are as rare as the GPUs they deploy. Exabits checks all of these boxes and we are the premier Compute Tokenization Platform. We are backed by REAL COMPUTE used by REAL CUSTOMERS.

"Compute is the currency of the future." "Compute is the new OIL." Great, so how do I invest? You can play with the NVIDIA stock, or spin up an AI-Ready Data Center. Challenges of Spinning up AI-Ready Compute: Cost: Even a small cluster of H100 machines is upward to $10M GPU Scarcity: Unallocated or “underutilized” H100, A100, and RTX4090 GPUs are almost non-existent. Very Specific Physical & Technical Requirements: A data center with the right infrastructure is also very rare. There are only 10 in North America that deploy thousands of GPUs, in the open market. Domain Expertise: Teams of experts with the years of experience required to operate, deploy, and manage AI compute at scale are as rare as the GPUs they deploy. Exabits checks all of these boxes and we are the premier Compute Tokenization Platform. We are backed by REAL COMPUTE used by REAL CUSTOMERS.

Exabits

13,895 görüntüleme • 1 yıl önce

🚨 NEW DATA: California fast food prices are now more than 20% higher than the national average. Twice as big of a gap as two years ago. Thanks to Democrat policies, the 'California Premium' for your Double Double - just doubled. Thanks Gavin.

🚨 NEW DATA: California fast food prices are now more than 20% higher than the national average. Twice as big of a gap as two years ago. Thanks to Democrat policies, the 'California Premium' for your Double Double - just doubled. Thanks Gavin.

steve hilton

46,295 görüntüleme • 8 ay önce

The Trump campaign keeps asking, “Are you better off now than you were 4 years ago?” So, let’s answer that question. Are you better off now than you were on March 13, 2020?

The Trump campaign keeps asking, “Are you better off now than you were 4 years ago?” So, let’s answer that question. Are you better off now than you were on March 13, 2020?

All In with Chris Hayes

310,355 görüntüleme • 2 yıl önce

Interesting that in three of my recent conversations with Krishna at Anthropic, Dylan Patel, and Gavin Baker, each said that frontier tokens are capturing the majority of the economic value. Gavin: "An overwhelming amount of the economic returns to AI at the model layer have been at the frontier. That's surprising to me, and I think it's been surprising to a lot of people. This is one of the most important questions to be answered, and you need to have a hypothesis on it as an investor." Krishna: "We think the returns to frontier intelligence are extremely high. Customers invest really heavily in more tokens with the newer models. The ones at the frontier clearly are capturing this economic value, driving meaningful ROI for customers. The returns to frontier intelligence are not slowing down." Dylan: "No one gives a crap about GPT-4 class models. They want the frontier because the frontier lets them create the economically valuable things."

Interesting that in three of my recent conversations with Krishna at Anthropic, Dylan Patel, and Gavin Baker, each said that frontier tokens are capturing the majority of the economic value. Gavin: "An overwhelming amount of the economic returns to AI at the model layer have been at the frontier. That's surprising to me, and I think it's been surprising to a lot of people. This is one of the most important questions to be answered, and you need to have a hypothesis on it as an investor." Krishna: "We think the returns to frontier intelligence are extremely high. Customers invest really heavily in more tokens with the newer models. The ones at the frontier clearly are capturing this economic value, driving meaningful ROI for customers. The returns to frontier intelligence are not slowing down." Dylan: "No one gives a crap about GPT-4 class models. They want the frontier because the frontier lets them create the economically valuable things."

Patrick OShaughnessy

121,110 görüntüleme • 1 ay önce

20 years ago Jane Street's entire compute cluster was six Dell boxes stacked on the floor at the end of an office row. They called it the Hive. Last week they let Dwarkesh Patel walk through their new Texas data center. 4,032 GPUs. Each rack pulls 140 kilowatts. A normal data center rack pulls 10. 8,000 kilometers of fiber inside one building, and the fastest links are still copper because light moves more slowly through glass than electrons through metal. Bookmark it tonight. Then read the article below.

20 years ago Jane Street's entire compute cluster was six Dell boxes stacked on the floor at the end of an office row. They called it the Hive. Last week they let Dwarkesh Patel walk through their new Texas data center. 4,032 GPUs. Each rack pulls 140 kilowatts. A normal data center rack pulls 10. 8,000 kilometers of fiber inside one building, and the fastest links are still copper because light moves more slowly through glass than electrons through metal. Bookmark it tonight. Then read the article below.

zostaff

236,142 görüntüleme • 1 ay önce

62 years ago today, Bob Dylan recorded “The Times They Are a-Changin’”

62 years ago today, Bob Dylan recorded “The Times They Are a-Changin’”

Melodies & Masterpieces

35,517 görüntüleme • 8 ay önce

Californians across the state are being priced out of homeownership, rents are skyrocketing, and more than 170,000 Californians are living on our streets. It’s unacceptable. Two years ago today, I launched CA DOJ’s Housing Justice Team to tackle this crisis from all angles. 🧵

Californians across the state are being priced out of homeownership, rents are skyrocketing, and more than 170,000 Californians are living on our streets. It’s unacceptable. Two years ago today, I launched CA DOJ’s Housing Justice Team to tackle this crisis from all angles. 🧵

Rob Bonta

40,583 görüntüleme • 2 yıl önce