Loading video...

Video Failed to Load

Go Home

While the Internet was freaking out about the Bryan Johnson post on his GF’s vagina and the J.P. Morgan Dom… Anthropic introduced “introspection adapters” which is a way to monitor what a model is doing under the hood in real time, and how safety in AI could get built. 👇

bigsuey's profile picture

Sue Ennis

17,322 subscribers

18,002 views • 3 months ago •via X (Twitter)

Science & Technology
LIVE

Anya Rossi• Live Now

Private livecam show

0 Comments

No comments available

Comments from the original post will appear here

Related Videos

This was a MAGA before The BBB passed and then this is after he actually read the BBB and found out the real truth about what was in it and how Trump lied to everyone! 👇👇👇
1:27

This was a MAGA before The BBB passed and then this is after he actually read the BBB and found out the real truth about what was in it and how Trump lied to everyone! 👇👇👇

Suzie rizzio

178,862 views • 10 months ago

What is left when AI runs it all? In this Unchained, Michael J. Casey and David Mattin join me to discuss: 💡 How Moltbook points to where the AI meta is headed 😬 How AI could impact jobs ❕️ Which country is best positioned to win the AI race ⁉️ What a post-human economy looks like 👀 Which jobs survive in a post-human economy Timestamps: 🚀 0:29 Introduction 🧏‍♂️ 2:10 How the Moltbook saga offers a window into where the AI meta could be headed 🤔 9:29 Why Michael wants a sovereign AI model 🌎 17:42 How AI could impact jobs ⚠️ 25:31 How AI could have a worse effect on the mental health young people than social media ❕️ 30:27 Which country is best positioned to win the AI race? 📍 35:02 What money looks like in a post-human economy 🤔 52:00 Which jobs flourish in a post-human economy? 💡 1:00:33 Michael and David share tokens and projects they find intriguing
59:23

What is left when AI runs it all? In this Unchained, Michael J. Casey and David Mattin join me to discuss: 💡 How Moltbook points to where the AI meta is headed 😬 How AI could impact jobs ❕️ Which country is best positioned to win the AI race ⁉️ What a post-human economy looks like 👀 Which jobs survive in a post-human economy Timestamps: 🚀 0:29 Introduction 🧏‍♂️ 2:10 How the Moltbook saga offers a window into where the AI meta could be headed 🤔 9:29 Why Michael wants a sovereign AI model 🌎 17:42 How AI could impact jobs ⚠️ 25:31 How AI could have a worse effect on the mental health young people than social media ❕️ 30:27 Which country is best positioned to win the AI race? 📍 35:02 What money looks like in a post-human economy 🤔 52:00 Which jobs flourish in a post-human economy? 💡 1:00:33 Michael and David share tokens and projects they find intriguing

Laura Shin

16,726 views • 6 months ago

OpenAI and Anthropic haven't won the enterprise AI market yet. “When you get a CIO one-on-one, they’re not happy with what Anthropic is trying to do. They're not happy with what OpenAI is bringing to the enterprise.” "I think ultimately, OpenAI is a consumer AI company. Anthropic is a coding model company, and the enterprise is actually really, really wide open and incredibly diverse in terms of the winners" — May Habib, CEO of WRITER
0:59

OpenAI and Anthropic haven't won the enterprise AI market yet. “When you get a CIO one-on-one, they’re not happy with what Anthropic is trying to do. They're not happy with what OpenAI is bringing to the enterprise.” "I think ultimately, OpenAI is a consumer AI company. Anthropic is a coding model company, and the enterprise is actually really, really wide open and incredibly diverse in terms of the winners" — May Habib, CEO of WRITER

The Information

39,146 views • 7 days ago

Penelope Ford was injured doing a moonsault to the floor which led to an impromptu finish in the tag title match on Dynamite. "The fact is doing a moonsault off the post to the floor... I mean at some point you're gonna get hurt and that goes for Hangman Page and Charlotte and Tiffany and Sol Ruca and everybody that does this." - Bryan Alvarez
1:02

Penelope Ford was injured doing a moonsault to the floor which led to an impromptu finish in the tag title match on Dynamite. "The fact is doing a moonsault off the post to the floor... I mean at some point you're gonna get hurt and that goes for Hangman Page and Charlotte and Tiffany and Sol Ruca and everybody that does this." - Bryan Alvarez

Wrestling Observer

21,129 views • 5 months ago

Today is Day 11 of 12 Days of Agent Two. Agent Two isn't just Claude under the hood - we get asked this a lot. The truth is - while we use all the state of the art models out there, the Agent's real intelligence comes from its harness: a domain-specific layer built from continuous creative research that shapes how every model is actually used - more about this in the video. But this harness is what makes Agent Two frontier intelligence for creative work - at 4x the speed, and half the price.
5:00

Today is Day 11 of 12 Days of Agent Two. Agent Two isn't just Claude under the hood - we get asked this a lot. The truth is - while we use all the state of the art models out there, the Agent's real intelligence comes from its harness: a domain-specific layer built from continuous creative research that shapes how every model is actually used - more about this in the video. But this harness is what makes Agent Two frontier intelligence for creative work - at 4x the speed, and half the price.

Invideo

2,541,249 views • 10 days ago

last month was my final month at Sophon, my time there has deepened something i already believed in: brand building is a perception game. media is shifting in real time with ai. it's already rewriting how attention works, how stories get told, and how brands get built. im truly excited about the future. so what's next? the intersection of crypto, ai, payments, and privacy excite me the most right now. each one of theses verticals play a key role in shaping what our society could look like a decade from today and i want to be a part of that process.
1:06

last month was my final month at Sophon, my time there has deepened something i already believed in: brand building is a perception game. media is shifting in real time with ai. it's already rewriting how attention works, how stories get told, and how brands get built. im truly excited about the future. so what's next? the intersection of crypto, ai, payments, and privacy excite me the most right now. each one of theses verticals play a key role in shaping what our society could look like a decade from today and i want to be a part of that process.

Abhz 💭

17,084 views • 5 months ago

As expected, Tesla’s Model Y ride-hailing service in the Bay Area currently has a safety driver in the driver’s seat, which is why Tesla is calling it a ride-hailing service rather than a Robotaxi service. It is still a paid service though and is likely using the same software as the Austin Robotaxis. Once Tesla receives the required permit, they will remove the safety driver and instead have a safety monitor in the passenger seat like they do in Austin. That could happen within the next 1–2 months, according to Elon.
0:49

As expected, Tesla’s Model Y ride-hailing service in the Bay Area currently has a safety driver in the driver’s seat, which is why Tesla is calling it a ride-hailing service rather than a Robotaxi service. It is still a paid service though and is likely using the same software as the Austin Robotaxis. Once Tesla receives the required permit, they will remove the safety driver and instead have a safety monitor in the passenger seat like they do in Austin. That could happen within the next 1–2 months, according to Elon.

Sawyer Merritt

304,428 views • 1 year ago

Dylan Patel, founder of SemiAnalysis: "The upper bound on how much compute can be produced by 2030 is around 200 gigawatts a year." The entire world has about 20 gigawatts of AI deployed right now. The ceiling is 10x what exists today, and it still isn't enough to feed what Sam, Elon, Dario and Demis are racing to build. Everyone argues about which model wins. The real limit is a number measured in gigawatts, and it's already maxed out years in advance.
7:13

Dylan Patel, founder of SemiAnalysis: "The upper bound on how much compute can be produced by 2030 is around 200 gigawatts a year." The entire world has about 20 gigawatts of AI deployed right now. The ceiling is 10x what exists today, and it still isn't enough to feed what Sam, Elon, Dario and Demis are racing to build. Everyone argues about which model wins. The real limit is a number measured in gigawatts, and it's already maxed out years in advance.

zostaff

163,567 views • 2 months ago

The crovwd is seeing a man get molested in real time and doing nothing about it😭😭😂
0:31

The crovwd is seeing a man get molested in real time and doing nothing about it😭😭😂

ARYA™

42,104 views • 5 months ago

IMO — Decentralized AI is more than: - an AI model in the sky, with good external auditing - an AI model in the sky, which people vote on how to use - an AI model in the sky, which is free for anyone to use - open source AI - federated training None of these are truly an interface to the world's collective intelligence. Each is actually... *mostly* centralized AI... but with the right ambitions!!! In this podcast, I lay out what I think a true decentralized AI ecosystem looks like, and my guesses on how to get there. The key use-case is broad listening (video below describes broad listening) (link to full podcast in reply)
1:03

IMO — Decentralized AI is more than: - an AI model in the sky, with good external auditing - an AI model in the sky, which people vote on how to use - an AI model in the sky, which is free for anyone to use - open source AI - federated training None of these are truly an interface to the world's collective intelligence. Each is actually... *mostly* centralized AI... but with the right ambitions!!! In this podcast, I lay out what I think a true decentralized AI ecosystem looks like, and my guesses on how to get there. The key use-case is broad listening (video below describes broad listening) (link to full podcast in reply)

⿻ Andrew Trask

65,233 views • 11 months ago

Asked on MSNBC how will respond to President Donald J. Trump asserting the National Guard will be deployed to other states, Mayor Brandon Johnson said: "The community coming together to push back in a very dramatic way." Johnson knows exactly what he is doing. The mayor is calling for his progressive allies to stage mass protests in the streets. Johnson is inciting mass protests and violence.
0:32

Asked on MSNBC how will respond to President Donald J. Trump asserting the National Guard will be deployed to other states, Mayor Brandon Johnson said: "The community coming together to push back in a very dramatic way." Johnson knows exactly what he is doing. The mayor is calling for his progressive allies to stage mass protests in the streets. Johnson is inciting mass protests and violence.

Chicago Contrarian

27,919 views • 1 year ago

Apple built a large foundation model and fine-tuned it on multiple tasks. But they are doing something very clever: They load a single model in memory and use different adapters to specialize the model on the fly. I recorded a video to show you how to write the code to do the same thing Apple is doing. I explain everything step by step. Here is what I'll show you in the video: 1. We'll load two datasets 2. Then load a large model 3. Then, we'll fine-tune the model on both datasets I'll use LoRA to fine-tune the model. This process creates two small adapters, each specializing in solving one of the datasets. The base model's original parameters will remain unchanged. From here: 4. We'll generate a list of tasks 5. We'll load the correct adapter to solve each task The large model I'm using needs 346 MB of memory, but I only need to load it once. Each adapter is only 2.7 MB. I only need to load the base model once and pair it with any of the fine-tuned adapters. Minimum memory footprint and I can solve multiple tasks. Hope this helps!
38:02

Apple built a large foundation model and fine-tuned it on multiple tasks. But they are doing something very clever: They load a single model in memory and use different adapters to specialize the model on the fly. I recorded a video to show you how to write the code to do the same thing Apple is doing. I explain everything step by step. Here is what I'll show you in the video: 1. We'll load two datasets 2. Then load a large model 3. Then, we'll fine-tune the model on both datasets I'll use LoRA to fine-tune the model. This process creates two small adapters, each specializing in solving one of the datasets. The base model's original parameters will remain unchanged. From here: 4. We'll generate a list of tasks 5. We'll load the correct adapter to solve each task The large model I'm using needs 346 MB of memory, but I only need to load it once. Each adapter is only 2.7 MB. I only need to load the base model once and pair it with any of the fine-tuned adapters. Minimum memory footprint and I can solve multiple tasks. Hope this helps!

Santiago

84,747 views • 2 years ago

Anthropic has created an AI model that “can get into almost any software in the world,” former AI company founder and CEO Matt Shumer says — but the company is not releasing it to the public: “I have never seen something like this — a company that is about to IPO and having such a moral compass.” Shumer shares what makes the AI model so powerful and potentially dangerous.
4:55

Anthropic has created an AI model that “can get into almost any software in the world,” former AI company founder and CEO Matt Shumer says — but the company is not releasing it to the public: “I have never seen something like this — a company that is about to IPO and having such a moral compass.” Shumer shares what makes the AI model so powerful and potentially dangerous.

CBS Mornings

98,892 views • 4 months ago

Elon Musk says every job done at a computer is going like lightning. The last time this happened, the people it happened to came out ahead. Musk: “Anything that is digital, which is like just someone at a computer doing something, AI is going to take over those jobs like lightning.” Most people in his position soften that. He said it on a podcast without hedging once. Musk: “Just like digital computers took over the job of people doing manual calculations.” Computer was a job title before it was a machine. Rooms full of people working out by hand the arithmetic that kept aircraft in the air and put rockets past the atmosphere. The machines arrived and the title disappeared. Everyone who held it kept working. Dorothy Vaughan taught herself to program the IBM brought in to replace her section, then ran it. Katherine Johnson stopped doing arithmetic and started calculating the path to the moon. Musk: “But much faster.” Speed is the part everyone hears as the threat. It is closer to mercy. They spent years in limbo while an institution worked out what to do with them. Faster means less time waiting for permission to matter. The machine takes the part of your work that was already mechanical, the part you could write down as a procedure. What survives is the judgment about what is worth doing, which never fit in a procedure to begin with. The obvious objection is scale. Manual calculation was one occupation, and this reaches every screen in every industry at once. That same breadth is why the tools are everywhere. They needed an institution to put them on the new machine. Nobody is waiting on that now. The thing doing the replacing runs on a laptop and answers to whoever opens it. Musk gets called a doomer for saying this out loud. Telling people while there is still time to move is the opposite of that. The arithmetic left the building. The people who had been doing it went to the moon.
0:24

Elon Musk says every job done at a computer is going like lightning. The last time this happened, the people it happened to came out ahead. Musk: “Anything that is digital, which is like just someone at a computer doing something, AI is going to take over those jobs like lightning.” Most people in his position soften that. He said it on a podcast without hedging once. Musk: “Just like digital computers took over the job of people doing manual calculations.” Computer was a job title before it was a machine. Rooms full of people working out by hand the arithmetic that kept aircraft in the air and put rockets past the atmosphere. The machines arrived and the title disappeared. Everyone who held it kept working. Dorothy Vaughan taught herself to program the IBM brought in to replace her section, then ran it. Katherine Johnson stopped doing arithmetic and started calculating the path to the moon. Musk: “But much faster.” Speed is the part everyone hears as the threat. It is closer to mercy. They spent years in limbo while an institution worked out what to do with them. Faster means less time waiting for permission to matter. The machine takes the part of your work that was already mechanical, the part you could write down as a procedure. What survives is the judgment about what is worth doing, which never fit in a procedure to begin with. The obvious objection is scale. Manual calculation was one occupation, and this reaches every screen in every industry at once. That same breadth is why the tools are everywhere. They needed an institution to put them on the new machine. Nobody is waiting on that now. The thing doing the replacing runs on a laptop and answers to whoever opens it. Musk gets called a doomer for saying this out loud. Telling people while there is still time to move is the opposite of that. The arithmetic left the building. The people who had been doing it went to the moon.

Dustin

33,072 views • 8 days ago

Introducing Lemon Slice Live: the first real-time audio-video AI model Anyone can video chat with any character by just uploading a single photo. Every pixel on the screen is rendered at 25fps in real-time by a transformer model. Here’s how we built it + a live demo! 👇
0:23

Introducing Lemon Slice Live: the first real-time audio-video AI model Anyone can video chat with any character by just uploading a single photo. Every pixel on the screen is rendered at 25fps in real-time by a transformer model. Here’s how we built it + a live demo! 👇

LemonSlice

171,007 views • 1 year ago

MITRE is under no obligation to adhere to any of these requests, which is why this entire line of inquiry is a waste of time. The system was setup in a way so that the technology could not be revealed by public request. The only way we're getting the truth is the MH370 operation. Release the files and these videos will be proven real.
1:07

MITRE is under no obligation to adhere to any of these requests, which is why this entire line of inquiry is a waste of time. The system was setup in a way so that the technology could not be revealed by public request. The only way we're getting the truth is the MH370 operation. Release the files and these videos will be proven real.

Ashton Forbes

26,685 views • 4 days ago

🚨 There is a country in Europe in which press freedom is under fire. The State is attempting to ban a TV broadcaster. The broadcaster never broke the law and was never convicted in court. I am shocked by this. This is a violation of the #RuleOfLaw and of the freedom of speech. What is going on? And which country is this about? I explain it in this video... ⤵️
2:42

🚨 There is a country in Europe in which press freedom is under fire. The State is attempting to ban a TV broadcaster. The broadcaster never broke the law and was never convicted in court. I am shocked by this. This is a violation of the #RuleOfLaw and of the freedom of speech. What is going on? And which country is this about? I explain it in this video... ⤵️

Rob Roos 🇳🇱

72,217 views • 2 years ago

ICP MCP making games on Internet Computer a reality...?! DFINITY Foundation's Model Context Protocol (MCP), a standard introduced to let AI models connect securely to external data and tools, is already being put to good use. A new post shares X2 using Anthropic's Claude to play a generated video game with instructions routed through $ICP's MCP. Are we seeing the dawn of AI on Internet Computer happening in real time?
0:20

ICP MCP making games on Internet Computer a reality...?! DFINITY Foundation's Model Context Protocol (MCP), a standard introduced to let AI models connect securely to external data and tools, is already being put to good use. A new post shares X2 using Anthropic's Claude to play a generated video game with instructions routed through $ICP's MCP. Are we seeing the dawn of AI on Internet Computer happening in real time?

BSCN

21,334 views • 24 days ago

Noted economist Ruchir Sharma says 👇 • The entire world today has a monomaniacal focus on AI. • It’s all about which countries will be the winners and losers in AI. • Unfortunately for India, in contrast to what happened during the tech boom of the 1990s and 2000s, foreign investors have concluded that India is a loser in the AI space. • The world is obsessed with building infrastructure for AI, something India sorely lacks. • Countries like South Korea and Taiwan are doing far better, while India is nowhere close. This is the stark reality. Hype and photo-ops won’t suffice. India, which became the flagbearer of the IT and telecom revolution thanks to visionary leadership at the time, cannot afford to let the AI revolution pass us by. But sadly, that is exactly what is happening - and the world is not going to wait.
2:58

Noted economist Ruchir Sharma says 👇 • The entire world today has a monomaniacal focus on AI. • It’s all about which countries will be the winners and losers in AI. • Unfortunately for India, in contrast to what happened during the tech boom of the 1990s and 2000s, foreign investors have concluded that India is a loser in the AI space. • The world is obsessed with building infrastructure for AI, something India sorely lacks. • Countries like South Korea and Taiwan are doing far better, while India is nowhere close. This is the stark reality. Hype and photo-ops won’t suffice. India, which became the flagbearer of the IT and telecom revolution thanks to visionary leadership at the time, cannot afford to let the AI revolution pass us by. But sadly, that is exactly what is happening - and the world is not going to wait.

Supriya Shrinate

47,780 views • 3 months ago