NVIDIA Lends Support to Washington's Efforts to Ensure AI Safety

[T0@st]

In an event at the White House today, NVIDIA announced support for voluntary commitments that the Biden Administration developed to ensure advanced AI systems are safe, secure and trustworthy. The news came the same day NVIDIA's chief scientist, Bill Dally, testified before a U.S. Senate subcommittee seeking input on potential legislation covering generative AI. Separately, NVIDIA founder and CEO Jensen Huang will join other industry leaders in a closed-door meeting on AI Wednesday with the full Senate.

Seven companies including Adobe, IBM, Palantir and Salesforce joined NVIDIA in supporting the eight agreements the Biden-Harris administration released in July with support from Amazon, Anthropic, Google, Inflection, Meta, Microsoft and OpenAI.



The commitments are designed to advance common standards and best practices to ensure the safety of generative AI systems until regulations are in place, the White House said. They include:

• Testing the safety and capabilities of AI products before they're deployed
• Safeguarding AI models against cyber and insider threats, and...
• Using AI to help meet society's greatest challenges, from cancer to climate change.

Dally Shares NVIDIA's Experience
In his testimony, Dally told the Senate subcommittee that government and industry should balance encouraging innovation in AI with ensuring models are deployed responsibly. The subcommittee's hearing, "Oversight of AI: Rules for Artificial Intelligence," is among actions from policymakers around the world trying to identify and address potential risks of generative AI.

Earlier this year, the subcommittee heard testimonies from leaders of Anthropic, IBM and OpenAI, as well as academics such as Yoshua Bengio, a University of Montreal professor considered one of the godfathers of AI. Dally, who leads a global team of more than 300 at NVIDIA Research, shared the witness table on Tuesday with Brad Smith, Microsoft's president and vice chair. Dally's testimony briefly encapsulated NVIDIA's unique role in the evolution of AI over the last two decades.

How Accelerated Computing Sparked AI
He described how NVIDIA invented the GPU in 1999 as a graphics processing unit, then fit it for a broader role in parallel processing in 2006 with the CUDA programming software. Over time, developers across diverse scientific and technical computing fields found this new form of accelerated computing could significantly advance their work.

Along the way, researchers discovered GPUs also were a natural fit for AI's neural networks, because they require massive parallel processing. In 2012, the AlexNet model, trained on two NVIDIA GPUs, demonstrated human-like capabilities in image recognition. That result helped spark a decade of rapid advances using GPUs, leading to ChatGPT and other generative AI models used by hundreds of millions worldwide.

Today, accelerated computing and generative AI are showing the potential to transform industries, address global challenges and profoundly benefit society, said Dally, who chaired Stanford University's computer science department before joining NVIDIA.

AI's Potential and Limits
In written testimony, Dally provided examples of how AI is empowering professionals to do their jobs better than they might have imagined in fields as diverse as business, healthcare and climate science. Like any technology, AI products and services have risks and are subject to existing laws and regulations that aim to mitigate those risks.

Industry also has a role to play in deploying AI responsibly. Developers set limits for AI models when they train them and define their outputs. Dally noted that NVIDIA released in April NeMo Guardrails, open-source software developers can use to guide generative AI applications in producing accurate, appropriate and secure text responses. He said that NVIDIA also maintains internal risk-management guidelines for AI models.

Eyes on the Horizon
Making sure that new and exceptionally large AI models are accurate and safe is a natural role for regulators, Dally suggested. He said that these "frontier" models are being developed at a gigantic scale. They exceed the capabilities of ChatGPT and other existing models that have already been well-explored by developers and users. Dally urged the subcommittee to balance thoughtful regulation with the need to encourage innovation in an AI developer community that includes thousands of startups, researchers and enterprises worldwide. AI tools should be widely available to ensure a level playing field, he said.

During questioning, Senator Amy Klobuchar (D-MN) asked Dally why NVIDIA announced in March it's working with Getty Images. "At NVIDIA, we believe in respecting people's intellectual property rights," Dally replied. "We partnered with Getty to train large language models with a service called Picasso, so people who provided the original content got remunerated."

In closing, Dally reaffirmed NVIDIA's dedication to innovating generative AI and accelerated computing in ways that serve the best interests of all.

View at TechPowerUp Main Site | Source

 

[TheinsanegamerN]

This sounds like a GREAT way to use "safety" as an excuse to lock down the AI market so only multi billion dollar corpos can participate, while ensuring the typical citizen has no power to use these tools. OpenAI tried that once already.

Government and corporations are like bleach and ammonia. Both are bad, mixing them is FAR worse.

 

[Avro Arrow]

This is perhaps the most disturbing part of the entire article:

"Separately, NVIDIA founder and CEO Jensen Huang will join other industry leaders in a closed-door meeting on AI Wednesday with the full Senate."

A bunch of billionaires in a closed-door meeting with the full Senate. Why would this meeting need to be closed-door? Billionaires and politicians having secret meetings that the general populace isn't privy to? If that's not special treatment for the rich by the taxpayer-funded government, I don't know what is.

This is how things got as bad as they are. I hope Bernie throws something at their heads.

 

[Count von Schwalbe]

If that's not special treatment for the rich by the taxpayer-funded government, I don't know what is.

I wonder if it isn't more of the other way around...

These people control the flow of information. What government wouldn't want to court them?

 

[notaburner]

I wonder if it isn't more of the other way around...

These people control the flow of information. What government wouldn't want to court them?

Just seems like the typical approach to regulation. If you can't outright lobby the government into not taking action, the next best thing is to be in the room when the rules are written. Allows you to "soften" the impact on your business and use the rules to gain a competitive advantage against those who aren't party to the discussions. Obviously this isn't how things should be but it's not at all surprising or unique to the tech sector.

 

[R-T-B]

This sounds like a GREAT way to use "safety" as an excuse to lock down the AI market so only multi billion dollar corpos can participate, while ensuring the typical citizen has no power to use these tools. OpenAI tried that once already.

Government and corporations are like bleach and ammonia. Both are bad, mixing them is FAR worse.

I mean you can literally clone most of the AI programs source on github (including OpenAI) so that genie isn't going back in the bottle. Not sure this concern is remotely valid, but the concern about what the genie may do is.

 

[Dr. Dro]

This sounds like a GREAT way to use "safety" as an excuse to lock down the AI market so only multi billion dollar corpos can participate, while ensuring the typical citizen has no power to use these tools. OpenAI tried that once already.

Government and corporations are like bleach and ammonia. Both are bad, mixing them is FAR worse.

Because that is precisely the endgame. There has always been big money in generative AI (which is essentially the automation of intellectual work - computer kiosks have put McDonald's cashiers out of a job, in the future, advanced generative AI will put researchers and consultants out of a job, and eventually even lawyers, attorneys and prosecutors, saving megacorporations trillions if not quadrillions of dollars), but now that this AI cat got out of the bag they want to ensure that it will be fed data that causes the algorithm to comply and work to their own purpose.

I envision a future where the result of the marriage between advanced robotics and automation may very well lead to the near-complete replacement of both the low-skill and the specialized workforce, which will be left to fend for itself. There is still relative safety in trade jobs, but even those are in jeopardy. This will happen, it's not if, it's just a matter of when.

You know, I'm somewhat sad about having been born in the early 1990's. I haven't lived through the period where the foundation of modern society and technology was developed, and I won't live long enough to see the fruit of the early days of the space age. Instead, we have this transitional period ahead of us in which I am sure that difficulty and violence are sure to follow.

 

[AusWolf]

I stopped reading here:

He (Dally) described how NVIDIA invented the GPU in 1999 as a graphics processing unit

Nvidia did not invent the GPU. Nvidia only invented the term "GPU". GPUs have existed long before Nvidia, they just weren't called as such. What a douchebag this Dally guy must be!

 

[Pumper]

The wolf lending support in the security of the sheep.

 

[Tsukiyomi91]

The wolf lending support in the security of the sheep.

assuming the wolf doesn't devour the sheep in the end, which is kinda inevitable...

 

[Avro Arrow]

I wonder if it isn't more of the other way around...

These people control the flow of information. What government wouldn't want to court them?

Because the Declaration of Independence begins with "We, The People" not "We, The Multinational Corporations". The state's separation from big business is just as important as separation from religion (which is just another form of big business). Having the rich involved in ruling the country is called feudalism and has been most recently been seen in only two developed countries. The first was Hitler's Germany and the second is Putin's Russia (with his oligarch friends).

Having a government that's in bed with big business like that is inherently dangerous.

Just seems like the typical approach to regulation. If you can't outright lobby the government into not taking action, the next best thing is to be in the room when the rules are written. Allows you to "soften" the impact on your business and use the rules to gain a competitive advantage against those who aren't party to the discussions. Obviously this isn't how things should be but it's not at all surprising or unique to the tech sector.

I know, but what's disturbing in this case is how brazen they're being about it. Previously, it wasn't disclosed when it occurred because they knew that they were doing wrong. Now they're trying to make it seem...

 

[Count von Schwalbe]

Because the Declaration of Independence begins with "We, The People" not "We, The Multinational Corporations". The state's separation from big business is just as important as separation from religion (which is just another form of big business). Having the rich involved in ruling the country is called feudalism and has been most recently been seen in only two developed countries. The first was Hitler's Germany and the second is Putin's Russia (with his oligarch friends).

Having a government that's in bed with big business like that is inherently dangerous.

Much as I agree with what you say, I tend to be cynical about how our dear government views this...

But we stray dangerously close to incurring the wrath of the mods.

 

[Avro Arrow]

Much as I agree with what you say, I tend to be cynical about how our dear government views this...

But we stray dangerously close to incurring the wrath of the mods.

Why is that? We haven't said anything inappropriate or off-topic.

 

[kapone32]

Since when did Nvidia invent the GPU in 1999? We seem to forget Creative, Voodoo and a few others including ATI.

This is perhaps the most disturbing part of the entire article:

"Separately, NVIDIA founder and CEO Jensen Huang will join other industry leaders in a closed-door meeting on AI Wednesday with the full Senate."

A bunch of billionaires in a closed-door meeting with the full Senate. Why would this meeting need to be closed-door? Billionaires and politicians having secret meetings that the general populace isn't privy to? If that's not special treatment for the rich by the taxpayer-funded government, I don't know what is.

This is how things got as bad as they are. I hope Bernie throws something at their heads.

Did he not have a closed door meeting with India's Prime Minister last week?

 

[Avro Arrow]

Since when did Nvidia invent the GPU in 1999? We seem to forget Creative, Voodoo and a few others including ATI.


Did he not have a closed door meeting with India's Prime Minister last week?

I honestly don't know. I don't pay attention to news in India.

 

[Dr. Dro]

Since when did Nvidia invent the GPU in 1999?

GeForce 256. It was the first processor that could be fully called a GPU, a single chip which could fully process graphics on its own: it would do transform and lighting, triangle clipping, cube and bump mapping, video processing (it was the first card with motion compensation) as well as all "2D" tasks without offloading the processing of any of these elements to the CPU. The RIVA 128 was actually one of the first cards to combine the 2D and 3D acceleration in a single chip, the Voodoo required a separate 2D card and an external bridge cord hooking the two cards in the back of the PC.

ATI's contemporary at the time, the Rage Fury, did not have many of these capabilities, most notable the hardware transform and lighting, which was the killer feature back then: just as ray tracing is today and tessellation was a decade ago. These things wouldn't appear on the red side until the original Radeon in 2000. Rage 128/Fury's performance and feature set was closer to that of the RIVA TNT2, not exactly a match for the GeForce.

So while not entirely accurate to the letter, as in, "what GeForce 256 did was never seen before in history", its NV10 processor was the first one which did it all without external help, which merits Nvidia's claim to the creation of the GPU as "graphics processing unit" and not as a "3D accelerator card".

 

[AusWolf]

GeForce 256. It was the first processor that could be fully called a GPU, a single chip which could fully process graphics on its own: it would do transform and lighting, triangle clipping, cube and bump mapping, video processing (it was the first card with motion compensation) as well as all "2D" tasks without offloading the processing of any of these elements to the CPU. The RIVA 128 was actually one of the first cards to combine the 2D and 3D acceleration in a single chip, the Voodoo required a separate 2D card and an external bridge cord hooking the two cards in the back of the PC.

ATI's contemporary at the time, the Rage Fury, did not have many of these capabilities, most notable the hardware transform and lighting, which was the killer feature back then: just as ray tracing is today and tessellation was a decade ago. These things wouldn't appear on the red side until the original Radeon in 2000. Rage 128/Fury's performance and feature set was closer to that of the RIVA TNT2, not exactly a match for the GeForce.

So while not entirely accurate to the letter, as in, "what GeForce 256 did was never seen before in history", its NV10 processor was the first one which did it all without external help, which merits Nvidia's claim to the creation of the GPU as "graphics processing unit" and not as a "3D accelerator card".

Also:
"In 1997, Rendition collaborated with Hercules and Fujitsu on a "Thriller Conspiracy" project which combined a Fujitsu FXG-1 Pinolite geometry processor with a Vérité V2200 core to create a graphics card with a full T&L engine years before Nvidia's GeForce 256; This card, designed to reduce the load placed upon the system's CPU, never made it to market."

Graphics processing unit - Wikipedia

en.wikipedia.org

 

[kapone32]

GeForce 256. It was the first processor that could be fully called a GPU, a single chip which could fully process graphics on its own: it would do transform and lighting, triangle clipping, cube and bump mapping, video processing (it was the first card with motion compensation) as well as all "2D" tasks without offloading the processing of any of these elements to the CPU. The RIVA 128 was actually one of the first cards to combine the 2D and 3D acceleration in a single chip, the Voodoo required a separate 2D card and an external bridge cord hooking the two cards in the back of the PC.

ATI's contemporary at the time, the Rage Fury, did not have many of these capabilities, most notable the hardware transform and lighting, which was the killer feature back then: just as ray tracing is today and tessellation was a decade ago. These things wouldn't appear on the red side until the original Radeon in 2000. Rage 128/Fury's performance and feature set was closer to that of the RIVA TNT2, not exactly a match for the GeForce.

So while not entirely accurate to the letter, as in, "what GeForce 256 did was never seen before in history", its NV10 processor was the first one which did it all without external help, which merits Nvidia's claim to the creation of the GPU as "graphics processing unit" and not as a "3D accelerator card".

This i what I miss about TPU...context. I agree mostly with what you are saying but we know that until they got Voodoo the 3D engine was weak.

 

[R-T-B]

Why is that? We haven't said anything inappropriate or off-topic.

Politics are not permitted.

 

[Dr. Dro]

Also:
"In 1997, Rendition collaborated with Hercules and Fujitsu on a "Thriller Conspiracy" project which combined a Fujitsu FXG-1 Pinolite geometry processor with a Vérité V2200 core to create a graphics card with a full T&L engine years before Nvidia's GeForce 256; This card, designed to reduce the load placed upon the system's CPU, never made it to market."

Graphics processing unit - Wikipedia

en.wikipedia.org

Yeah, they had a full engine but it was the combination of two different chips and that also never made it to market. I mentioned "while not a never before seen, it was a world-first in it all being fully integrated into a single package"

Which was quite the achievement at the time, I think the claim has merit, until the TNT2 and the Fury, both companies only really offered combined 3D accelerators, really. Also, you ever read about the Fury Maxx? It was some sort of dual-core Rage Fury and the first to market with alternate frame rendering. They used some unorthodox dual-AGP port configuration that wasn't compatible with Windows 2000 or XP so it only really worked on Windows 98. Anandtech still has their review of it up, apparently

ATI Rage Fury MAXX

www.anandtech.com

 

[Why_Me]

Because the Declaration of Independence begins with "We, The People" not "We, The Multinational Corporations". The state's separation from big business is just as important as separation from religion (which is just another form of big business). Having the rich involved in ruling the country is called feudalism and has been most recently been seen in only two developed countries. The first was Hitler's Germany and the second is Putin's Russia (with his oligarch friends).

Having a government that's in bed with big business like that is inherently dangerous.


I know, but what's disturbing in this case is how brazen they're being about it. Previously, it wasn't disclosed when it occurred because they knew that they were doing wrong. Now they're trying to make it seem...
View attachment 313807

Pfizer, Facebook, pre Musk Twitter, MSM, Boeing, Raytheon, etc ... it's been happening for a while now.

 

[claes]

It has already happened everywhere, even post-Musk Twitter

Under Elon Musk, Twitter has approved 83% of censorship requests by authoritarian governments

The social network has restricted and withdrawn content critical of the ruling parties in Turkey and India, among other countries, including during electoral campaigns

english.elpais.com

 

[Why_Me]

It has already happened everywhere, even post-Musk Twitter

Under Elon Musk, Twitter has approved 83% of censorship requests by authoritarian governments

The social network has restricted and withdrawn content critical of the ruling parties in Turkey and India, among other countries, including during electoral campaigns

english.elpais.com

That article contradicts itself. It starts off pushing free speech and then later on says more moderation on social media sites such as Facebook and Twitter need more moderation. Who ever wrote that piece should think about another career other than journalism.

 

[Dr. Dro]

That article contradicts itself. It starts off pushing free speech and then later on says more moderation on social media sites such as Facebook and Twitter need more moderation. Who ever wrote that piece should think about another career other than journalism.

It's the El País. If you know a thing about that newspaper, you'd very well know they are not favorable to free speech. It's basically a Spanish-language Washington Post of sorts.

 

[Avro Arrow]

Politics are not permitted.

This article is about a closed-doors meeting between tech billionaires and the US Senate so the story itself is about politics. Going by your logic, this article couldn't have a comment section at all because no matter what, every comment made could be construed as political. I believe that politics is only a taboo subject if it's off-topic (which isn't the case here).