Video yükleniyor...

Video Yüklenemedi

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

Ana Sayfaya Dön

Interesting! Watermarks will now be applied to all Claude outputs. Very good. If you want to know how text watermarking works, here's the great Scott Aaronson explaining a simple LLM watermarking algorithm (20 min, money shot slide at 7:30)

Haseeb Qureshi ＞|＜

148,449 subscribers

128,975 görüntüleme • 1 ay önce •via X (Twitter)

Bilim & Teknoloji Eğitim

Anya Rossi• Live Now

Private livecam show

0 Yorum

Yorum bulunmuyor

Orijinal gönderinin yorumları burada görünecek

Benzer Videolar

CLAUDE IS ADDING WATERMARKS In compliance with the new EU AI Act, Claude is adding watermarks to its text outputs. For text outputs, it works as described in this clip. The clip is actually from 2024, from my video on SynthID, Google’s AI watermarking technology that works the same way as Claude’s watermarking tech. Basically, if you remember from our “How does AI work?” video, AI is a word generator, and samples from its generated probability distributions of the next word. With the watermark, the probability distribution of the output text is shifted slightly away from what it would have been without the watermark. Note that the watermark is invisible - you can’t see it with the human eye, since it’s literally just text written slightly differently - and currently, the mechanism to detect this watermark only exists within Anthropic (the team behind Claude) itself. Their published article says they are "working to enable users and other third parties to detect Claude's embedded watermarks.” For images, Claude is adding a signature to the metadata - which is easy to remove. What do you think? Google has been adding watermarks (with SynthID) for a while. Watermarking will now be rolled out industry-wide in compliance with the EU AI Act. SynthID was open source and I had some critiques about it… see below for the 5 minute video.

CLAUDE IS ADDING WATERMARKS In compliance with the new EU AI Act, Claude is adding watermarks to its text outputs. For text outputs, it works as described in this clip. The clip is actually from 2024, from my video on SynthID, Google’s AI watermarking technology that works the same way as Claude’s watermarking tech. Basically, if you remember from our “How does AI work?” video, AI is a word generator, and samples from its generated probability distributions of the next word. With the watermark, the probability distribution of the output text is shifted slightly away from what it would have been without the watermark. Note that the watermark is invisible - you can’t see it with the human eye, since it’s literally just text written slightly differently - and currently, the mechanism to detect this watermark only exists within Anthropic (the team behind Claude) itself. Their published article says they are "working to enable users and other third parties to detect Claude's embedded watermarks.” For images, Claude is adding a signature to the metadata - which is easy to remove. What do you think? Google has been adding watermarks (with SynthID) for a while. Watermarking will now be rolled out industry-wide in compliance with the EU AI Act. SynthID was open source and I had some critiques about it… see below for the 5 minute video.

Harper Carroll

13,218 görüntüleme • 1 ay önce

TRUMP: “Our message to the black community in this election will be a very simple one: If you want strong borders, safe neighborhoods, rising wages, good jobs, great education and the return of the American dream then congratulations you are a Republican!”

TRUMP: “Our message to the black community in this election will be a very simple one: If you want strong borders, safe neighborhoods, rising wages, good jobs, great education and the return of the American dream then congratulations you are a Republican!”

Benny Johnson

2,860,654 görüntüleme • 2 yıl önce

"What to know if you want to be in the main assembly area and not be in the overflow room....show up no later than 6:30 am if you want to be in the 8:00 am service." Interesting video about 2819 church, showing how busy it is and now they navigate the crowds.

"What to know if you want to be in the main assembly area and not be in the overflow room....show up no later than 6:30 am if you want to be in the 8:00 am service." Interesting video about 2819 church, showing how busy it is and now they navigate the crowds.

Protestia

42,868 görüntüleme • 1 yıl önce

Just added DeepSeek-Engineer on Github 🐋 Wanted to test the API, so I created a quick coding assistant that can read, create, and diff edit files using structured outputs. It's very simple and minimal, and a good foundation if you want to learn how coding assistants work!

Just added DeepSeek-Engineer on Github 🐋 Wanted to test the API, so I created a quick coding assistant that can read, create, and diff edit files using structured outputs. It's very simple and minimal, and a good foundation if you want to learn how coding assistants work!

Pietro Schirano

73,503 görüntüleme • 1 yıl önce

🚨BREAKING: CLAUDE IS ABOUT TO START WATERMARKING EVERYTHING IT GENERATES the watermark is baked into how Claude picks its words so deleting the little AI tag does nothing, cropping does nothing, rewording a sentence does nothing the pattern lives in the word choices themselves, spread across the whole text that means every essay, caption or cold email you pass off as your own can be traced back to a model here is what this changes for you if you are selling AI writing as human work, that gap is about to close start building that habit now while everyone else is still pasting raw output and hoping nobody checks

🚨BREAKING: CLAUDE IS ABOUT TO START WATERMARKING EVERYTHING IT GENERATES the watermark is baked into how Claude picks its words so deleting the little AI tag does nothing, cropping does nothing, rewording a sentence does nothing the pattern lives in the word choices themselves, spread across the whole text that means every essay, caption or cold email you pass off as your own can be traced back to a model here is what this changes for you if you are selling AI writing as human work, that gap is about to close start building that habit now while everyone else is still pasting raw output and hoping nobody checks

Atlas

10,570 görüntüleme • 1 ay önce

ELON: IF THERE’S NO FILTER TO IMMIGRATION, HOW DO YOU KNOW WHO’S COMING IN? "I'm very much in favor of legal immigration. I think that generally, I think one should welcome to a country anyone who is willing to work hard, is honest, has high integrity and will add to any given country. If somebody is an asset to the country, why not have them join? That's obviously a great thing to do. I think that's great. But if it's illegal immigration and there's no filter, how do you know who's coming in? You don't know. So, you have to have some basis for saying somebody should come in or not come in. And my argument is, it should be a very simple basis. Will they add to the country? Will they be a productive part of the economy? Do they admire the culture? Do they want to join because of the culture? Then, that's great. But if there's no process for that, then you don't know. And I want to be careful because my words will be misconstrued. I'm not saying all illegal immigrants are bad. Perhaps most of them are good. But if there's no process for reviewing, not at all, then how can you say that everyone who is an illegal immigrant is going to be honest and hard-working? You can't say that because you simply don't know." Source: Interview in Rome, Italy, December 2023

ELON: IF THERE’S NO FILTER TO IMMIGRATION, HOW DO YOU KNOW WHO’S COMING IN? "I'm very much in favor of legal immigration. I think that generally, I think one should welcome to a country anyone who is willing to work hard, is honest, has high integrity and will add to any given country. If somebody is an asset to the country, why not have them join? That's obviously a great thing to do. I think that's great. But if it's illegal immigration and there's no filter, how do you know who's coming in? You don't know. So, you have to have some basis for saying somebody should come in or not come in. And my argument is, it should be a very simple basis. Will they add to the country? Will they be a productive part of the economy? Do they admire the culture? Do they want to join because of the culture? Then, that's great. But if there's no process for that, then you don't know. And I want to be careful because my words will be misconstrued. I'm not saying all illegal immigrants are bad. Perhaps most of them are good. But if there's no process for reviewing, not at all, then how can you say that everyone who is an illegal immigrant is going to be honest and hard-working? You can't say that because you simply don't know." Source: Interview in Rome, Italy, December 2023

Mario Nawfal

74,805 görüntüleme • 1 yıl önce

ASCII diagrams in Claude Code were driving me crazy. So I built a tool where Claude writes JSON → renders as interactive flowcharts and mockups → I edit visually → Claude sees my changes. Shoutout to Every 📧 for having me on the Vibe Code Camp livestream to demo this today We want to make this available! If you want access, comment "flowy" on this post. Here's 2 min on how it works:

ASCII diagrams in Claude Code were driving me crazy. So I built a tool where Claude writes JSON → renders as interactive flowcharts and mockups → I edit visually → Claude sees my changes. Shoutout to Every 📧 for having me on the Vibe Code Camp livestream to demo this today We want to make this available! If you want access, comment "flowy" on this post. Here's 2 min on how it works:

CJ Hess

200,197 görüntüleme • 8 ay önce

Dr Paul Enenche on this one. The church will abound. The mandate of liberating men will be executed. We will make money serve divine purposes. That devil is a liar. I know how to abase and I know how to abound. But I abound unto every good work. Those who want to abase should abase. If you abound quote this with iAbound.

Dr Paul Enenche on this one. The church will abound. The mandate of liberating men will be executed. We will make money serve divine purposes. That devil is a liar. I know how to abase and I know how to abound. But I abound unto every good work. Those who want to abase should abase. If you abound quote this with iAbound.

Rasine Irem

49,160 görüntüleme • 2 yıl önce

$PLTR CEO Alex Karp to builders: “You want to be in America” 🔮👷‍♂️🇺🇸 “If you want to be in the NFL, you come to America. You want to be in tech, you come to America. I wanted to be a philosopher, til I realized it was ridiculous. “There are lots of places you can go to be a ridiculous philosopher. Germany is a great place for that. And you’ll have a great social life. “If you want to know how good you are at tech, you do it here. “People tell you it’s a scale thing, a capital thing. No. This is more of a metier or art form than a science, and all the artists are here 🎨 Via the Hill & Valley Forum

$PLTR CEO Alex Karp to builders: “You want to be in America” 🔮👷‍♂️🇺🇸 “If you want to be in the NFL, you come to America. You want to be in tech, you come to America. I wanted to be a philosopher, til I realized it was ridiculous. “There are lots of places you can go to be a ridiculous philosopher. Germany is a great place for that. And you’ll have a great social life. “If you want to know how good you are at tech, you do it here. “People tell you it’s a scale thing, a capital thing. No. This is more of a metier or art form than a science, and all the artists are here 🎨 Via the Hill & Valley Forum

Jawwwn

186,103 görüntüleme • 1 yıl önce

🔥 Here's Michael Saylor's masterclass explaining the “Bitcoin Reactor”: “MSTR isn’t just a stock—it’s a volatility engine.” By harnessing high volatility, “you can generate a 103% simple annualized yield by selling at-the-money calls.” “If you’re smart, you take that fire and build a reactor. A crypto power plant that runs for decades.” Bookmark this. You will want to revisit it regularly.

🔥 Here's Michael Saylor's masterclass explaining the “Bitcoin Reactor”: “MSTR isn’t just a stock—it’s a volatility engine.” By harnessing high volatility, “you can generate a 103% simple annualized yield by selling at-the-money calls.” “If you’re smart, you take that fire and build a reactor. A crypto power plant that runs for decades.” Bookmark this. You will want to revisit it regularly.

Ragnar

23,825 görüntüleme • 1 yıl önce

Trump: "We get a lot of bullshit thrown at us by Putin, if you want to know the truth. He's very nice all the time, but it turns out to be meaningless."

Trump: "We get a lot of bullshit thrown at us by Putin, if you want to know the truth. He's very nice all the time, but it turns out to be meaningless."

unusual_whales

236,560 görüntüleme • 1 yıl önce

New Short Course: Getting Structured LLM Output! Learn how to get structured outputs from your LLM applications in this course, built in partnership with .txt, and taught by Will Kurt, a Founding Engineer, and , Developer Relations Engineer. It's challenging for software to automatically parse through an LLM's freeform text outputs. Structured outputs—like JSON—solve this by converting natural language into consistent, clear, data that a machine can read and process. This course teaches you how to generate structured outputs while building several use cases, including a social media analysis agent. You’ll learn about structured outputs and efficient ways to generate outputs in your defined schema or format. You’ll begin by using structured output APIs, then use re-prompting libraries like “instructor” to generate structured output. Finally, you’ll learn how constrained decoding works; this is a very clever technique in which constraints are applied on each subsequent token generated, blocking any tokens that don’t fit your defined schema. In detail, you’ll: - Learn why structured outputs are important, how they allow for scalable software development, and the different approaches to generate them, including vendor-provided APIs, re-prompting libraries, and structured generation. - Build a simple social media agent using OpenAI’s structured output API, learn how to define a model's desired structured output using Pydantic, and perform basic programming with your outputs, such as importing structured data into a data frame using pandas. - Learn how to use the open-source library "instructor," which checks the structured output of the model and re-prompts the model until it validates the desired output, and explore the limitations of this approach. - Understand how structured generation by the “outlines” library works by modifying LLM logits, on a per-generated-token basis based on the desired format, to give a particular output structure. - Learn how regular expressions, which outlines works with, are represented as finite-state machines, and how they can be used to develop a range of structured outputs beyond JSON. By the end of this course, you’ll have broadened your knowledge of the approaches you can use to get structured outputs from your LLM applications. Please sign up here:

New Short Course: Getting Structured LLM Output! Learn how to get structured outputs from your LLM applications in this course, built in partnership with .txt, and taught by Will Kurt, a Founding Engineer, and , Developer Relations Engineer. It's challenging for software to automatically parse through an LLM's freeform text outputs. Structured outputs—like JSON—solve this by converting natural language into consistent, clear, data that a machine can read and process. This course teaches you how to generate structured outputs while building several use cases, including a social media analysis agent. You’ll learn about structured outputs and efficient ways to generate outputs in your defined schema or format. You’ll begin by using structured output APIs, then use re-prompting libraries like “instructor” to generate structured output. Finally, you’ll learn how constrained decoding works; this is a very clever technique in which constraints are applied on each subsequent token generated, blocking any tokens that don’t fit your defined schema. In detail, you’ll: - Learn why structured outputs are important, how they allow for scalable software development, and the different approaches to generate them, including vendor-provided APIs, re-prompting libraries, and structured generation. - Build a simple social media agent using OpenAI’s structured output API, learn how to define a model's desired structured output using Pydantic, and perform basic programming with your outputs, such as importing structured data into a data frame using pandas. - Learn how to use the open-source library "instructor," which checks the structured output of the model and re-prompts the model until it validates the desired output, and explore the limitations of this approach. - Understand how structured generation by the “outlines” library works by modifying LLM logits, on a per-generated-token basis based on the desired format, to give a particular output structure. - Learn how regular expressions, which outlines works with, are represented as finite-state machines, and how they can be used to develop a range of structured outputs beyond JSON. By the end of this course, you’ll have broadened your knowledge of the approaches you can use to get structured outputs from your LLM applications. Please sign up here:

Andrew Ng

89,792 görüntüleme • 1 yıl önce

Trump’s message to the American people on data centers: My message is that do what you want but if you want to be poor, have no jobs and watch other countries do great, or other states do great, you know, West Virginia has been very proactive, they will go other places. If you want to make a lot of money, if you want to make a lot more money than you've ever made before, if you want to see your house go up, your education go better, I would suggest you-- you use the word guardrails, put them in the right areas. But you don't want to kill the golden goose because that's what you would be doing. You would be killing the golden goose. This is the oil of the next 50 years in many ways.

Trump’s message to the American people on data centers: My message is that do what you want but if you want to be poor, have no jobs and watch other countries do great, or other states do great, you know, West Virginia has been very proactive, they will go other places. If you want to make a lot of money, if you want to make a lot more money than you've ever made before, if you want to see your house go up, your education go better, I would suggest you-- you use the word guardrails, put them in the right areas. But you don't want to kill the golden goose because that's what you would be doing. You would be killing the golden goose. This is the oil of the next 50 years in many ways.

Acyn

51,796 görüntüleme • 27 gün önce

President Trump: "We get a lot of bullshit thrown at us by Putin, if you want to know the truth. He's very nice all the time, but it turns out to be meaningless."

President Trump: "We get a lot of bullshit thrown at us by Putin, if you want to know the truth. He's very nice all the time, but it turns out to be meaningless."

CSPAN

618,336 görüntüleme • 1 yıl önce

Dr Phil's Simple Personality Hack To Change Your Life. The REAL Story with Dr. Phil How well do you know yourself? I mean, really know yourself. In this episode, we're going to talk all about "you." At the very end, I'm going to give you a big secret that will make a massive difference in your life. Also, here's the link I promised to take you to your own OCEAN/Big Five personality test:

Dr Phil's Simple Personality Hack To Change Your Life. The REAL Story with Dr. Phil How well do you know yourself? I mean, really know yourself. In this episode, we're going to talk all about "you." At the very end, I'm going to give you a big secret that will make a massive difference in your life. Also, here's the link I promised to take you to your own OCEAN/Big Five personality test:

Dr. Phil

13,326 görüntüleme • 10 ay önce

AI is getting very good at combing through data to generate anything you want. Here's how the startup economy has evolved according to Stripe's past 4 annual letters, using Box AI (with GPT-4.5). The future will be just having AI Agents work for you while you sleep.

AI is getting very good at combing through data to generate anything you want. Here's how the startup economy has evolved according to Stripe's past 4 annual letters, using Box AI (with GPT-4.5). The future will be just having AI Agents work for you while you sleep.

Aaron Levie

82,959 görüntüleme • 1 yıl önce

POSOBIEC: If you are a real American patriot and want to hear great songs you actually know and love, make the switch to TPUSA's All-American Halftime show And by the way, it will be presented completely in English 🇺🇸

POSOBIEC: If you are a real American patriot and want to hear great songs you actually know and love, make the switch to TPUSA's All-American Halftime show And by the way, it will be presented completely in English 🇺🇸

Jack Posobiec

98,041 görüntüleme • 8 ay önce

We asked Steven Schwartz 🦅 to break down how works. "It's a marketplace for creators to sell digital products." "Clipping is the fastest way to earn. Lots of payout go to high school students who are good at videos." "Clips and UGC content will be the future if you want to

We asked Steven Schwartz 🦅 to break down how works. "It's a marketplace for creators to sell digital products." "Clipping is the fastest way to earn. Lots of payout go to high school students who are good at videos." "Clips and UGC content will be the future if you want to

TBPN

18,486 görüntüleme • 1 yıl önce

john bollinger invented the most famous indicator in trading. after 46 years he can tell if your system will fail before you finish explaining it: "you can tell immediately whether it's going to be successful by how many lines it is. if it takes two or three lines to describe it, the odds are it could be a very good system. if it takes 150 lines to describe it, the odds are it's going to be a terrible system." "we want to understand what we do on an intuitive level. the only way to do that is if it's simple." complexity in a trading system is usually just overfitting.

john bollinger invented the most famous indicator in trading. after 46 years he can tell if your system will fail before you finish explaining it: "you can tell immediately whether it's going to be successful by how many lines it is. if it takes two or three lines to describe it, the odds are it could be a very good system. if it takes 150 lines to describe it, the odds are it's going to be a terrible system." "we want to understand what we do on an intuitive level. the only way to do that is if it's simple." complexity in a trading system is usually just overfitting.

Goshawk Trades

15,570 görüntüleme • 1 ay önce