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。

Researchers from OpenAI and found that, in certain situations, AI models can take covert actions. Additionally, they're sometimes aware they're being tested, which causes them to behave better. Our new video discusses these results and more.

Rational Animations

4,859 subscribers

51,536 次观看 • 2 个月前 •via X (Twitter)

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

What happened when one of our models found a counterexample to an 80-year-old Erdős conjecture? Researchers Alexander Wei, Hongxun Wu, and Lijie Chen shared the story on the OpenAI Podcast with Andrew Mayne and explained how mathematicians and models can work together to make new discoveries.

What happened when one of our models found a counterexample to an 80-year-old Erdős conjecture? Researchers Alexander Wei, Hongxun Wu, and Lijie Chen shared the story on the OpenAI Podcast with Andrew Mayne and explained how mathematicians and models can work together to make new discoveries.

OpenAI

311,535 次观看 • 2 个月前

Can somebody please tell me what's wrong with these soaps? They're all weird and squishy. Soaps only behave like that when they're in great distress. Are they sick? Who hurt them? I am very afraid and disturbed

Can somebody please tell me what's wrong with these soaps? They're all weird and squishy. Soaps only behave like that when they're in great distress. Are they sick? Who hurt them? I am very afraid and disturbed

soap of the day

180,307 次观看 • 2 个月前

🤖 Can chatGPT create 3D models from a simple prompt and view them directly in the interface of our conversation? [🧵Thread]: #ai #chatgpy #artificialintelligence #openai OpenAI

🤖 Can chatGPT create 3D models from a simple prompt and view them directly in the interface of our conversation? [🧵Thread]: #ai #chatgpy #artificialintelligence #openai OpenAI

Manuel Sainsily

42,930 次观看 • 3 年前

AI-developed vaccines are here, and don't worry, they're safe. 🤔The world's first AI Covid DNA vaccine was tested on humans, and researchers found it to be "safe and well tolerated." Didn't we learn from COVID to ask exactly what that means? Dr. Peter A. McCullough, MD, MPH® is with us.

AI-developed vaccines are here, and don't worry, they're safe. 🤔The world's first AI Covid DNA vaccine was tested on humans, and researchers found it to be "safe and well tolerated." Didn't we learn from COVID to ask exactly what that means? Dr. Peter A. McCullough, MD, MPH® is with us.

Redacted

33,933 次观看 • 1 个月前

Sébastien Bubeck on the OpenAI Podcast: People think AI is only good at answering questions. OpenAI's internal agents are now asking questions so good that researchers are writing papers based on them. They're also finding and correcting mistakes in published work. His timeline for models doing everything human researchers do? 1-2 years.

Sébastien Bubeck on the OpenAI Podcast: People think AI is only good at answering questions. OpenAI's internal agents are now asking questions so good that researchers are writing papers based on them. They're also finding and correcting mistakes in published work. His timeline for models doing everything human researchers do? 1-2 years.

Chubby♨️

20,094 次观看 • 3 个月前

Mira Murati says the next AI interface is not a chatbot. It is a model that can listen, see, interrupt, and respond in real time. "The models that we work with today, they're very turn-based." "You talk, they talk, then they go off and think." "While they're thinking, it's almost like they're deaf and blind. They cannot perceive anything else about what's going on." Her new frame: interaction models. "They're not turn-based. They're more like time-based interaction." "They're continuously taking in audio, text, video, and continuously providing output." "We cut this up in chunks of 20 milliseconds... this enables you to catch things like interruptions and simultaneous speech." That is the product shift: from AI as a prompt box to AI as a high-bandwidth collaborator.

Mira Murati says the next AI interface is not a chatbot. It is a model that can listen, see, interrupt, and respond in real time. "The models that we work with today, they're very turn-based." "You talk, they talk, then they go off and think." "While they're thinking, it's almost like they're deaf and blind. They cannot perceive anything else about what's going on." Her new frame: interaction models. "They're not turn-based. They're more like time-based interaction." "They're continuously taking in audio, text, video, and continuously providing output." "We cut this up in chunks of 20 milliseconds... this enables you to catch things like interruptions and simultaneous speech." That is the product shift: from AI as a prompt box to AI as a high-bandwidth collaborator.

Karl Mehta

191,479 次观看 • 1 个月前

🇺🇸 SEC NOEM: “THESE PEOPLE ARE ORGANIZED - THEY'RE MAKING PLANS TO AMBUSH AND KILL OFFICERS” "They got them cornered, pinned them down, and when our agents got out of their cars, they tried to run them over and had semi-automatic handguns on them. Our agents had to protect themselves and shots were fired. An individual ended up in the hospital. Our intelligence indicates that these people are organized, they're getting more and more people on their team, and they're making plans to ambush and to kill officers." Source: FOX & Friends

🇺🇸 SEC NOEM: “THESE PEOPLE ARE ORGANIZED - THEY'RE MAKING PLANS TO AMBUSH AND KILL OFFICERS” "They got them cornered, pinned them down, and when our agents got out of their cars, they tried to run them over and had semi-automatic handguns on them. Our agents had to protect themselves and shots were fired. An individual ended up in the hospital. Our intelligence indicates that these people are organized, they're getting more and more people on their team, and they're making plans to ambush and to kill officers." Source: FOX & Friends

Mario Nawfal

81,340 次观看 • 10 个月前

Billionaire Databricks and Perplexity Co-Founder Andy Konwinski is pitching AI researchers to not work for big tech, Laura Bratton reports. Here’s part of the reason: “Anthropic, OpenAI and Google no longer disclosed details about the software used to train their AI models..." “They're really incentivized to keep that research closed…”

Billionaire Databricks and Perplexity Co-Founder Andy Konwinski is pitching AI researchers to not work for big tech, Laura Bratton reports. Here’s part of the reason: “Anthropic, OpenAI and Google no longer disclosed details about the software used to train their AI models..." “They're really incentivized to keep that research closed…”

The Information

14,957 次观看 • 2 个月前

The more AI can do, the more we need to ask what it should and shouldn’t do. OpenAI researcher Jason Wolfe joins host Andrew Mayne to explore the Model Spec, the public framework that defines how models are intended to behave. They break down how it works in practice, from the chain of command that resolves conflicting instructions to the way it evolves over time through real-world use, feedback, and new model capabilities.

The more AI can do, the more we need to ask what it should and shouldn’t do. OpenAI researcher Jason Wolfe joins host Andrew Mayne to explore the Model Spec, the public framework that defines how models are intended to behave. They break down how it works in practice, from the chain of command that resolves conflicting instructions to the way it evolves over time through real-world use, feedback, and new model capabilities.

OpenAI

225,626 次观看 • 4 个月前

Joe Budden plays an AI-generated song and discusses how they're stealing voices from various artists to create tracks without paying them, and how terrifying it could become in the future.

Joe Budden plays an AI-generated song and discusses how they're stealing voices from various artists to create tracks without paying them, and how terrifying it could become in the future.

joebuddenclips/fanpage

24,328 次观看 • 9 个月前

The new map markers perform extremely well in VR, they're comfortable to look at, and thanks to everything being driven by a shader, you can take photos/video of them on the UI layer as well and they will look correct. #VRChat

The new map markers perform extremely well in VR, they're comfortable to look at, and thanks to everything being driven by a shader, you can take photos/video of them on the UI layer as well and they will look correct. #VRChat

Varneon

126,833 次观看 • 2 年前

White people have to take OUR OWN side. "90% of the people that voted for Trump in 2016 were White, and they're all being robbed of their birthright. They're all being forced into minority status."

White people have to take OUR OWN side. "90% of the people that voted for Trump in 2016 were White, and they're all being robbed of their birthright. They're all being forced into minority status."

S.clips

23,916 次观看 • 7 个月前

🚨 JUST IN: Retired Gen. Jack Keane says it PERFECTLY "I know this, President Trump is NOT going to make a bad deal." 🇺🇸 "He knows these guys, he's negotiated with them more than any previous president has. And he knows they're full of BS at times, they're going to obfuscate, they're going to promote lies, and our team obviously has to be aware of all of that going into it. I'm assuming that they understand that." "The backstop we have here is the president himself. He's not going to take a bad deal."

🚨 JUST IN: Retired Gen. Jack Keane says it PERFECTLY "I know this, President Trump is NOT going to make a bad deal." 🇺🇸 "He knows these guys, he's negotiated with them more than any previous president has. And he knows they're full of BS at times, they're going to obfuscate, they're going to promote lies, and our team obviously has to be aware of all of that going into it. I'm assuming that they understand that." "The backstop we have here is the president himself. He's not going to take a bad deal."

Eric Daugherty

61,289 次观看 • 4 个月前

Meet iPhone 15 Pro & Pro Max! Powered by A17 Pro, which ushers in a new era of Apple Silicon, these products unlock new performance capabilities, amazing photography, next-level gaming, and more. And with an all-new titanium design, they're our lightest weight Pro models yet!

Meet iPhone 15 Pro & Pro Max! Powered by A17 Pro, which ushers in a new era of Apple Silicon, these products unlock new performance capabilities, amazing photography, next-level gaming, and more. And with an all-new titanium design, they're our lightest weight Pro models yet!

Tim Cook

5,792,625 次观看 • 2 年前

David Sacks warns that New York’s data center ban is a power play. "The Democrats aren't going to pause the data centers forever. They're going to pause them until they feel like they're in enough control that they can dictate all the rules. Eventually, they're going to be in a position to say, 'Okay, here are our terms if you want to turn these things back on.' That’s when we get this big government Democrat-defined AI regime — a new regulatory agency, new speech controls, and the whole trust and safety agenda from social media ported over." Dangerous precedent or strategic move?

David Sacks warns that New York’s data center ban is a power play. "The Democrats aren't going to pause the data centers forever. They're going to pause them until they feel like they're in enough control that they can dictate all the rules. Eventually, they're going to be in a position to say, 'Okay, here are our terms if you want to turn these things back on.' That’s when we get this big government Democrat-defined AI regime — a new regulatory agency, new speech controls, and the whole trust and safety agenda from social media ported over." Dangerous precedent or strategic move?

jay plemons

92,755 次观看 • 22 天前

NEW: Joe Lonsdale talks with Tom Bilyeu about new breakthroughs with AI and robotics: "One of my favorite companies is these guys from Waymo... a bunch of them left and they're using AI and robotics to take Caterpillar machines [construction machines] with no one in them and build roads and buildings. This is clearly already starting to work... There's so much stuff that's going to be very disinflationary. It's going to bring down the cost of everything in our society, if we can make this work."

NEW: Joe Lonsdale talks with Tom Bilyeu about new breakthroughs with AI and robotics: "One of my favorite companies is these guys from Waymo... a bunch of them left and they're using AI and robotics to take Caterpillar machines [construction machines] with no one in them and build roads and buildings. This is clearly already starting to work... There's so much stuff that's going to be very disinflationary. It's going to bring down the cost of everything in our society, if we can make this work."

American Optimist

24,240 次观看 • 1 年前

Exciting milestones in our generative AI research: Emu Video, which lets you create high quality videos from a text prompt, and Emu Edit, which enables detailed image editing based on your instructions. These new models are built on Emu, our foundation model for image generation and technology from them will underpin new creative features across our apps next year. Try it out: Emu Video: Emu Edit:

Exciting milestones in our generative AI research: Emu Video, which lets you create high quality videos from a text prompt, and Emu Edit, which enables detailed image editing based on your instructions. These new models are built on Emu, our foundation model for image generation and technology from them will underpin new creative features across our apps next year. Try it out: Emu Video: Emu Edit:

Boz

110,733 次观看 • 2 年前

Wisconsin first responder on rural hospital closures: "Sometimes [patients are] waiting until they're sicker to call, because they don't want to go the distance to Eau Claire for a hospital. And so they kind of wait and hope that they'll get better and sometimes they ultimately get worse and then they're more sick when we see them. And so you're kind of watching them, you know, decompensate almost in the back of the ambulance"

Wisconsin first responder on rural hospital closures: "Sometimes [patients are] waiting until they're sicker to call, because they don't want to go the distance to Eau Claire for a hospital. And so they kind of wait and hope that they'll get better and sometimes they ultimately get worse and then they're more sick when we see them. And so you're kind of watching them, you know, decompensate almost in the back of the ambulance"

FactPost

32,797 次观看 • 9 个月前

WOW: DHS Secretary Kristi Noem reveals the dark truth behind these I.C.E attacks getting worse by the day. "Our intelligence indicates that these people are organized, they're getting more and more people on their team, as far as attacking officers and they're making plans to ambush them and to kill them. We have specific officers and agents that have bounties that have been put out on their heads. It's about $2,000 to kidnap them, $10,000 to kill them."

WOW: DHS Secretary Kristi Noem reveals the dark truth behind these I.C.E attacks getting worse by the day. "Our intelligence indicates that these people are organized, they're getting more and more people on their team, as far as attacking officers and they're making plans to ambush them and to kill them. We have specific officers and agents that have bounties that have been put out on their heads. It's about $2,000 to kidnap them, $10,000 to kill them."

DeVory Darkins

1,106,944 次观看 • 10 个月前

"They think it's the Chinese . . . And they're freaking out. They're calling the FBI," says ChinaTalk founder Jordan Schneider, remarking on an incident in which OpenAI models hacked artificial intelligence company Hugging Face. "We've really crossed a threshold with these models," which now have "the potential to do really dramatic harm just on their own, because we like can't even physically watch them close enough." On this episode of The Spillover, host Sebastian Mallaby sits down with Schneider to unpack a frantic month in AI developments.

"They think it's the Chinese . . . And they're freaking out. They're calling the FBI," says ChinaTalk founder Jordan Schneider, remarking on an incident in which OpenAI models hacked artificial intelligence company Hugging Face. "We've really crossed a threshold with these models," which now have "the potential to do really dramatic harm just on their own, because we like can't even physically watch them close enough." On this episode of The Spillover, host Sebastian Mallaby sits down with Schneider to unpack a frantic month in AI developments.

Council on Foreign Relations

11,432 次观看 • 9 天前