Video yükleniyor...

Video Yüklenemedi

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

Ana Sayfaya Dön

Delphi AI researcher Accelxr 👾 explains why restricting frontier AI model access can create its own security risk. “If you’re not a company that gets access to it, you lose the ability to defend yourself against it when it’s released... Now it becomes a race immediately after it is... show more

Delphi Digital

1,515,587 subscribers

24,330 görüntüleme • 2 ay önce •via X (Twitter)

Bilim & Teknoloji Haberler & Politika

Anya Rossi• Live Now

Private livecam show

5 Yorum

Delphi Digital profil fotoğrafı

Delphi Digital2 ay önce

Watch the full episode here.

Omega profil fotoğrafı

Omega2 ay önce

@accelxr That means anyone not in on the secret sauce is essentially playing catch up with no shield, great analysis !

Fortune Keeper's Collective profil fotoğrafı

Fortune Keeper's Collective2 ay önce

@accelxr Restricted access can create an uneven defense problem. The governance question is how to expand legitimate evaluation and preparedness without turning capability release into a race with no guardrails.

Rompel profil fotoğrafı

Rompel2 ay önce

Red-teaming after release is also how every jailbreak library gets built — the "race" starts the moment weights or API access go public either way. The real question is whether restricted access actually buys defenders lead time, or just buys the model vendor plausible deniability.

sophie d. profil fotoğrafı

sophie d.1 ay önce

@accelxr Restricted access can buy time to develop defenses before release, delaying the race. But if the model leaks, you're worse off. Where's the line between control and obscurity?

Benzer Videolar

Delphi AI researcher Accelxr 👾 explains why Nvidia’s move into open models could make it a competitor to its own customers. “Nvidia is best positioned to extract maximum throughput for their models out of their own tech stack.”

Delphi AI researcher Accelxr 👾 explains why Nvidia’s move into open models could make it a competitor to its own customers. “Nvidia is best positioned to extract maximum throughput for their models out of their own tech stack.”

Delphi Digital

22,422 görüntüleme • 1 ay önce

Delphi AI researcher Accelxr 👾 explains why China’s AI progress is more than just distillation. “Distillation can explain a large chunk of the capability and benchmark side, but it can’t explain a lot of the things they’ve shipped... The research is quite impressive.”

Delphi AI researcher Accelxr 👾 explains why China’s AI progress is more than just distillation. “Distillation can explain a large chunk of the capability and benchmark side, but it can’t explain a lot of the things they’ve shipped... The research is quite impressive.”

Delphi Digital

37,686 görüntüleme • 2 ay önce

Delphi AI researcher Accelxr 👾 explains why access to Nvidia chips could slow China’s push for chip independence. “Getting them back on the Nvidia bridge should slow down their development, kind of paradoxically. It slows their sovereignty push, and any tightening actually accelerates their push to build out their own chips.”

Delphi AI researcher Accelxr 👾 explains why access to Nvidia chips could slow China’s push for chip independence. “Getting them back on the Nvidia bridge should slow down their development, kind of paradoxically. It slows their sovereignty push, and any tightening actually accelerates their push to build out their own chips.”

Delphi Digital

20,562 görüntüleme • 2 ay önce

Delphi AI researcher Accelxr 👾 explains why inference could strengthen China’s position despite its compute disadvantage. “If the axis of competition has actually shifted from the training side to the serving side, all of these innovations... make them better situated.”

Delphi AI researcher Accelxr 👾 explains why inference could strengthen China’s position despite its compute disadvantage. “If the axis of competition has actually shifted from the training side to the serving side, all of these innovations... make them better situated.”

Delphi Digital

38,776 görüntüleme • 2 ay önce

AI access is becoming the new blunt instrument in geopolitics. Aaron Levie on what happens if model access becomes a lever: "A year ago, I wouldn’t have put blocking access to an AI model on the board of possible moves. Now it is.”

AI access is becoming the new blunt instrument in geopolitics. Aaron Levie on what happens if model access becomes a lever: "A year ago, I wouldn’t have put blocking access to an AI model on the board of possible moves. Now it is.”

Deirdre Bosa

12,889 görüntüleme • 4 ay önce

Jensen Huang: the first thing you should do when deploying an AI agent is take away all of its rights. “Then you give it access to files, data, tools, the network, or even the internet only if it needs it.” “You don’t put an agent into your company and give it access to everything.” “We don’t give that to people. We don’t give that to any level of executive.” “We have to apply the same philosophy to AI agents.” “That’s what OpenShell is: a force field around the agent, provisioning rights, responsibilities and access.”

Jensen Huang: the first thing you should do when deploying an AI agent is take away all of its rights. “Then you give it access to files, data, tools, the network, or even the internet only if it needs it.” “You don’t put an agent into your company and give it access to everything.” “We don’t give that to people. We don’t give that to any level of executive.” “We have to apply the same philosophy to AI agents.” “That’s what OpenShell is: a force field around the agent, provisioning rights, responsibilities and access.”

Veee

401,291 görüntüleme • 1 ay önce

Anthropic is refusing to release its latest model because it’s so powerful that it becomes a security threat. @jason makes the case that the AI race is “existential” to national security because of what models like Anthropic’s withheld model, “Mythos,” can do. If bad actors like China get their hands on these capabilities with no incentive to disclose it, there are some pretty scary implications… The real question… What if they already have it?

Anthropic is refusing to release its latest model because it’s so powerful that it becomes a security threat. @jason makes the case that the AI race is “existential” to national security because of what models like Anthropic’s withheld model, “Mythos,” can do. If bad actors like China get their hands on these capabilities with no incentive to disclose it, there are some pretty scary implications… The real question… What if they already have it?

This Week in Startups

23,084 görüntüleme • 7 ay önce

Satya Nadella reveals why every company may need its own AI model: the model becomes the new company database. "To me, a model is like the database market." "A firm should be able to take the tacit knowledge it has and embed it inside weights in a model that they control." "When somebody asks me how many models should there be, I'll say as many models as firms in the world." The contrarian part: the value may not sit in one universal frontier model. It sits in each company turning its private operating knowledge into a controlled model.

Satya Nadella reveals why every company may need its own AI model: the model becomes the new company database. "To me, a model is like the database market." "A firm should be able to take the tacit knowledge it has and embed it inside weights in a model that they control." "When somebody asks me how many models should there be, I'll say as many models as firms in the world." The contrarian part: the value may not sit in one universal frontier model. It sits in each company turning its private operating knowledge into a controlled model.

Karl Mehta

469,102 görüntüleme • 4 ay önce

Delphi AI researcher Accelxr 👾 explains how China is scaling AI compute with domestic chips. “These superclusters take a ton of Ascend chips and combine them together. The supercluster itself has, I think, over 500k Ascend chips. They're pushing on it for sure.”

Delphi AI researcher Accelxr 👾 explains how China is scaling AI compute with domestic chips. “These superclusters take a ton of Ascend chips and combine them together. The supercluster itself has, I think, over 500k Ascend chips. They're pushing on it for sure.”

Delphi Digital

30,296 görüntüleme • 1 ay önce

We’re moving from asking, “What AI do you use?” to “What AI does your company run on?” to “What AI is running your company?” Today, we’re launching your @ company agent in Adapt. Your company’s own AI, branded with your name and deployable within minutes. Tag @ yourcompany in Slack and ask it to do the work. As we go through this change, it is important that a company owns its AI. Own the agent, @ yourcompany. Own the knowledge it builds. Own the model routing. Own the experience. The company agent is the durable layer. It gives your company access to the best frontier and open-weight models, choosing the right model for each task to balance quality, speed, and cost. When you tag @ yourcompany, you’re talking directly to your business through one front door. It learns how your company works, chooses the right model and skill, delegates work to specialized agents, and becomes more valuable as your company’s knowledge grows. A few things make this possible. It has to connect to every system your employees rely on, with the right access and permissions to understand the business and take action. It needs its own secure computer to write code, build apps, analyze data, operate software, and do the work. It has to be everywhere your people work. Web, Slack, iMessage, GitHub, and more. Vercel CEO Guillermo Rauch articulated this perfectly when describing @ v, the internal company agent they built. “If agents become the foundation of modern companies, you being in complete control, from source to runtime to data to token, seems like a pretty big deal.” Every company will have its own @ company agent. Adapt is how you deploy yours. What AI is running your company?

We’re moving from asking, “What AI do you use?” to “What AI does your company run on?” to “What AI is running your company?” Today, we’re launching your @ company agent in Adapt. Your company’s own AI, branded with your name and deployable within minutes. Tag @ yourcompany in Slack and ask it to do the work. As we go through this change, it is important that a company owns its AI. Own the agent, @ yourcompany. Own the knowledge it builds. Own the model routing. Own the experience. The company agent is the durable layer. It gives your company access to the best frontier and open-weight models, choosing the right model for each task to balance quality, speed, and cost. When you tag @ yourcompany, you’re talking directly to your business through one front door. It learns how your company works, chooses the right model and skill, delegates work to specialized agents, and becomes more valuable as your company’s knowledge grows. A few things make this possible. It has to connect to every system your employees rely on, with the right access and permissions to understand the business and take action. It needs its own secure computer to write code, build apps, analyze data, operate software, and do the work. It has to be everywhere your people work. Web, Slack, iMessage, GitHub, and more. Vercel CEO Guillermo Rauch articulated this perfectly when describing @ v, the internal company agent they built. “If agents become the foundation of modern companies, you being in complete control, from source to runtime to data to token, seems like a pretty big deal.” Every company will have its own @ company agent. Adapt is how you deploy yours. What AI is running your company?

Jim Benton

18,925 görüntüleme • 2 ay önce

Anthropic, the tech company that recently clashed with the Pentagon, says it has developed a new AI model that is too powerful to be released to the public.

Anthropic, the tech company that recently clashed with the Pentagon, says it has developed a new AI model that is too powerful to be released to the public.

Channel 4 News

103,858 görüntüleme • 7 ay önce

Wild. AI avatars can now interpret text automatically. Synthesia just released a new diffusion model called EXPRESS-1, which allows AI avatars to understand a script and adjust performance on the spot. Direct link to access and try it below.

Wild. AI avatars can now interpret text automatically. Synthesia just released a new diffusion model called EXPRESS-1, which allows AI avatars to understand a script and adjust performance on the spot. Direct link to access and try it below.

Rowan Cheung

117,595 görüntüleme • 2 yıl önce

Google has released a new AI image generator. Its interface is easy to use and totally free. I'll show you how to access it and compare it with Dall-E 3 below ⬇

Google has released a new AI image generator. Its interface is easy to use and totally free. I'll show you how to access it and compare it with Dall-E 3 below ⬇

Paul Couvert

310,937 görüntüleme • 2 yıl önce

The biggest danger isn’t AI. It’s getting left behind. “You need access to compute. You need access to your own data.” “The stack is compute, data, ability to manage data, ability to control the models, ability to train the models, and ability to compete against your competitors and adversaries.” “You are gonna need a lot of electricity. How are you gonna get it?” “I don’t think the AI revolution is a hundred percent good. There are huge dangers. But being left behind is probably the biggest danger now.”

The biggest danger isn’t AI. It’s getting left behind. “You need access to compute. You need access to your own data.” “The stack is compute, data, ability to manage data, ability to control the models, ability to train the models, and ability to compete against your competitors and adversaries.” “You are gonna need a lot of electricity. How are you gonna get it?” “I don’t think the AI revolution is a hundred percent good. There are huge dangers. But being left behind is probably the biggest danger now.”

Chad Wahlquist

30,401 görüntüleme • 2 ay önce

Delphi AI researcher Accelxr 👾 explains why AI agents could favor cheaper open weights models. “Agentic use cases are extremely cost sensitive. And so if that’s the case, you kind of have this composition shift towards cheapest, but still adequately capable models.”

Delphi AI researcher Accelxr 👾 explains why AI agents could favor cheaper open weights models. “Agentic use cases are extremely cost sensitive. And so if that’s the case, you kind of have this composition shift towards cheapest, but still adequately capable models.”

Delphi Digital

24,476 görüntüleme • 1 ay önce

Dan Hendrycks warns that sentient AI creates a moral trap. If an AI can suffer, using it starts to look like slavery. But once it has rights, we lose the ability to shut it down or constrain its growth. “These are very complicated moral questions that we'll all have to grapple with.”

Dan Hendrycks warns that sentient AI creates a moral trap. If an AI can suffer, using it starts to look like slavery. But once it has rights, we lose the ability to shut it down or constrain its growth. “These are very complicated moral questions that we'll all have to grapple with.”

vitrupo

38,075 görüntüleme • 10 ay önce

.OpenAI reassigned 25% of its production engineers to security. Then it used its top AI model to find vulnerabilities in its own systems. President Greg Brockman explains why.

.OpenAI reassigned 25% of its production engineers to security. Then it used its top AI model to find vulnerabilities in its own systems. President Greg Brockman explains why.

No Priors

30,681 görüntüleme • 20 gün önce

Delphi AI researcher Accelxr 👾 explains how Chinese labs are optimizing AI for domestic chips. “The Chinese labs have actually been going beneath that, handwriting low-level GPU instructions and exploiting some of the chip behaviors at a lower level.”

Delphi AI researcher Accelxr 👾 explains how Chinese labs are optimizing AI for domestic chips. “The Chinese labs have actually been going beneath that, handwriting low-level GPU instructions and exploiting some of the chip behaviors at a lower level.”

Delphi Digital

40,262 görüntüleme • 1 ay önce

I've got early access to a model that will be released this week! Thanks to the lab 🙏 No, it's not Qwen 4. Yes, it seems strong! Keep testing it! Share prompts you wanna try!

I've got early access to a model that will be released this week! Thanks to the lab 🙏 No, it's not Qwen 4. Yes, it seems strong! Keep testing it! Share prompts you wanna try!

Ivan Fioravanti

29,814 görüntüleme • 6 gün önce