Why I left Nvidia at the start of the AI ​​boom

AI News

Photo illustration: Intelligencer, Photo: Jacopo Pantaleoni

You may not know the name Jacopo Pantaleoni, but you certainly know his work. As a computer scientist, Pantaleoni worked on James Cameron's Avatar Mental Images, Inc. matrixPantaleoni, who specialized in turning data into images, went on to spend the next 15 years developing Nvidia's flagship graphics processing unit. His work helped shape the world of hyper-realistic digital images that make up everything from video games to movies to the world of bioinformatics, where they're essential for studying DNA sequences. They're also essential to artificial intelligence, serving as the backbone of OpenAI and its competitors.

But Pantaleoni, who lives in Germany, left the company last July just as the artificial intelligence boom was propelling it to the top of the tech world. Since then, he has become an outspoken critic of not only the technology but also Nvidia's role as a driving force behind that boom. For Pantaleoni, the most worrying outcome of AI technology is not apocalyptic, but matrix-Rather than the kind of scenario sometimes touted by Silicon Valley moguls, he's worried about a more familiar one: a small number of AI companies monopolizing an unhealthy amount of power, influence, and information. The Fastest Revolution: An Insider's Guide to Rapid Technological Change and Its Biggest Threats And now he's advising government regulators on how to rein in the technology. I spoke with Pantaleoni in late May about why he left Big Tech.

What made you leave Nvidia?
It was a realization that the extent to which I contribute to society, whether directly or indirectly, is much greater than I thought, and that I needed time to reflect and think critically about it. Most of my indirect contributions to society have not necessarily been good. Currently, computing technology poses some risks to society that are still underestimated. I'm not talking about existential risks from superintelligent AI or anything like that.

you're not?
No. At this point, it's a known risk, and people like Sam Altman are just proposing it to distract from the most pressing issues. [people like Altman] They often have very strong vested interests.The main risk now is the concentration of power and the resulting loss of jobs.Another big issue that I explore in the book is the increasing cognitive debilitation that computing technology is having across society.

We focus most of our attention on algorithms that are specifically designed to capture and sustain our attention to the fullest extent, primarily in our visual cortex. This is the main area of ​​my career contributions. I have worked primarily on visual computing, and it is the power of visual computing that has enabled companies like Nvidia to succeed.

Tell us about the work you've done and how it has contributed to an issue that you're very concerned about right now.
My background is in rendering technology, which is creating images that look as realistic as possible from 3D dataset models of visual objects that don't exist. I started doing this for the visual effects industry in the 90s. AvatarPrior to that, he worked as a visual scientist at Mental Images. matrix Film production has been around for a while before I joined the company, and my career has been primarily focused on high performance, high quality, high fidelity computer graphics rendering.

The same technology was applied to computer games. That's how companies like Nvidia were born in the late '90s and 2000s. They started producing hardware to bring this kind of rendering for computer games to the masses. I worked at Nvidia Research to develop visual computing technology five years ahead. In 2010, we – researchers like Jensen Huang and myself – realized that the computing power required for computer graphics is exactly the same as the power required for tasks like machine learning and artificial intelligence. That was Jensen Huang's genius moment. People like me contributed by making the hardware more and more programmable and more and more accessible to the masses.

At what point did you start to have doubts?
I do not understand reservation I think that's the right word, but around 2014 or 2015, I started to have this weird feeling about the impact of my work. I was witnessing a dramatic change. I was witnessing companies like Google and Amazon developing massively parallel computing power for very different uses, while also becoming much larger customers of the same kinds of machines. They were starting to adopt machine learning at scale, essentially to drive the advertising and attention economy.

How did it actually work? What exactly were they doing so differently?
In fact, I think that's the secret to Google's success. They were essentially a machine learning company, and they were the first company to realize huge benefits from the scalability of computer equipment. And they were one of the first companies to overcome what was called diminishing returns that plagued all manufacturing companies in the previous era. They actually had accelerating returns.

How did they do it?
Well, if you run a normal manufacturing company and you want to expand, you have to pay more employees more. You have to get more goods, more raw materials. In this case, the raw material for a company like Google is completely free. Just information. And they can easily process this raw material by throwing more computer power at it.

Were you surprised to see the technology you were developing being used in this way?
It was surprising. I've dedicated my life to building technology that turns mathematics into art, and that turns computing into a means of scientific research, like DNA sequencing, so to see this technology being used essentially for advertising and the attention economy was a little disheartening.

Was there anything Nvidia could have done?
I don't think so. Nvidia doesn't contribute to the attention economy, but it drives it. Companies like Nvidia can't say, “We refuse to let this market run.” When I left, I thought it was fundamentally impossible to change the outcomes these companies produce from the inside. The only way to fundamentally change their impact on society is to regulate them.

I said earlier that the real ultimate risk is the concentration of power. From chatbots telling people what to do. Putting glue on pizza For issues you are concerned about?
When you build tools that every person on the planet uses, you're already concentrating the power in your hands, and these tools automate more and more tasks.

Many media outlets, new york The magazine's parent company has partnered with OpenAI. Do you think it will have a similar impact on how the news industry operates?
When you have one or a few companies that own the tools that basically generate news for the whole world, that's political power. And so they have to be regulated. There's no other way. We're in a dangerous situation, and our democracy is in danger.

Do you still own Nvidia shares?
i will do it.

Has Nvidia's stock price impacted your life?
Yes, even before this stock price increase, Nvidia's stock price surge was impacting my financial life and the lives of many of my remaining colleagues.

That was another aspect that set off alarm bells in my head: I realized that I was essentially moving from being a highly paid graphics professional whose job was still niche to a new superstar league of scientists from companies like Google, Nvidia, Amazon, etc., who are now part of the 1 percent in tech in terms of wealth.

Do you feel conflicted about it?
I certainly agree, and I felt like there was a bit of an imbalance between the proliferation of economic wealth and economic value for the computer scientists who work at these companies and what they're actually contributing to the world, especially because I felt like that net impact wasn't inherently positive.

So, does that mean the financial rewards from your work at Nvidia outweigh your contributions to society?
That's a good way to put it. The impact of my work on society has certainly been great, but it hasn't always been as good as I would have hoped.

Do you plan to continue holding the Nvidia shares you have held thus far?
Not necessarily. For now, I just see them as part of savings for my family and not necessarily something I intend to keep.

Are you going to give money?
I don't have enough money to actually donate, Nvidia didn't make me rich, and I can enjoy a year without working.

What do you think you’re doing now?
I pivoted to become a public expert. I've provided consulting to regulators. I've been invited to speak at conferences. I'm not sure if I want to work as a technology expert again. Probably, but I'd have to find something that really aligns with my desire to have a positive impact on society.

What are some tech projects that have a positive impact on society?For example, anything related to healthcare, using technology to solve healthcare problems in a positive and not exploitative way. Or developing technology that is more gentle — that is, not trying to grab your attention, but is more peripheral, so to speak. Or any kind of technology that does the opposite of what's done today with AI assistants, like stimulating kids to be more social, or more focused on learning and learning to use their intelligence to solve problems.

Do you think people are generally aware of how they are being affected by technology?
At the moment, there is a big gap. But it is not based on research. This is just my personal perception of social media and people who use it. For example, many influencers see social media as an opportunity in life. And they believe that having hundreds of thousands of followers actually means something. In reality, they are all strangers who have no influence on their lives outside of advertising.

Do you think AI technology will accelerate that?
Yes, we think so. On the one hand, it's being used to maximize attention on an even larger scale and exploit all the information we have about our users and their preferences. On the other hand, it's enabling our users to generate more content and flood the web with even more distractions, making it more difficult to separate fact from fiction.

This interview has been edited for length and clarity.

Source link

Leave a Reply

Your email address will not be published. Required fields are marked *