OPEN AI Chief Scientist Ilya Sutzkaber: The role of academia in artificial intelligence research is changing

At a meeting held this week at Tel Aviv University with the participation of OPEN AI CEO Sam Altman, Sutzkober said that the academy cannot compete with the industry in computing power or engineering ability, but instead investigate how to better measure the performance of artificial intelligence, and use artificial intelligence to solve large problems

In a conversation that took place yesterday (Monday) at Tel Aviv University with the participation of Sutzkober, OPEN AI CEO Sam Altman and Prof. Nadav Cohen, from the host university, addressed the questions that lie in the seam between the academy and the technology industry when the academy cannot establish a large computing capacity, and although it does not have an engineering heritage it is Can definitely answer the difficult questions that accompany artificial intelligence

OPEN AI's chief scientist, Ilya Sutzkaber says that the role of academia in artificial intelligence research is changing. OPEN AI CEO Sam Altman and Chief Scientist Ilya Sutzkover participated in the event moderated by Dr. Nadav Cohen from Tel Aviv University.  

The two, and especially Sotskovich, referred to issues that are at the intersection between academia and industry, on the one hand, the contribution of universities to the study of artificial intelligence, as well as the role of artificial intelligence in solving major problems of humanity, primarily the climate crisis.

Sutzkaber explains: "Academia used to be the place for the most innovative research in the field, but now it is not so much for two reasons: the computing capacity and the engineering. The academy has less computing capacity than the industry can afford and there is generally no engineering culture. Still the academy can make very dramatic and significant contributions Artificial Intelligence: There are so many mysteries about the operation of neural networks.The production of these complex objects of deep learning perceived, which is like an alchemical process where we take the raw materials of the data plus the energy source of the computing and get the artificial intelligence.

Measuring the quality of artificial intelligence is impossible today

"But what it is, how it works, what its characteristics are, how we control it, how to make it so that we can contain it and also how we measure all these parameters, these disappear. Even the simple task of measuring quality is impossible. In the past the problem was not critical Because AI wasn't that important. But now that AI is important, we realize that we can't measure it. You don't need a huge computing cluster to answer these questions. You don't need a huge engineering team To ask these questions will be a dramatic and significant contribution that everyone will immediately notice." He told the audience that was made up of Israeli high-tech leaders, lecturers and students from Tel Aviv University.

CEO Altman also addressed the issue: there is confusion about what we are doing and how the academy should address it. Many articles have been written, but the important thing is to think about the most important problems, simply focus on them, change thinking about focus; about the most important problems: what is still not It is possible to do, what we still don't know, and how to measure the operation of artificial intelligence."

"First of all, you have to understand the problem, because understanding the problem is a prerequisite to moving towards a solution. This is where we can help. We have access programs to the academy where you can get computing capacity and access to our most advanced models. They research them, write articles. They even did it With GPT3. Even before we had the first product, researchers at many universities wrote articles, studied their models, their features, their biases. We would love to hear any ideas.

Quick solution of difficult problems

In another context related to academia, Sutzkever explained: "We are going to understand the mysteries of the universe and more than that I truly believe that scientific and technological progress is the only sustainable way in which life improves, the world improves."

"These capabilities will allow us to unlock a huge amount of new science, new technological advances. We're already seeing the beginnings of people using these tools to be more efficient, but if you imagine a world where you can ask AI to find cures for diseases or address the climate crisis ."

"For example, climate change is very serious. We believe that dealing with climate change will not be difficult for future generations of artificial intelligence. We will explain to him that we want to build efficient carbon capture systems.

If we can speed up scientific progress, which is something that strong artificial intelligence can do, we can achieve very advanced carbon capture, much faster, generate cheap energy much faster and we can achieve cheap manufacturing. The combination of all three will create a system that would not be possible without the artificial intelligence that accelerates the process."

"We will ask artificial intelligence to tell us how to produce a lot of clean energy cheaply, how to capture carbon efficiently and how to build a plant that will do it on a planetary scale."

Sutzkober later added that today's artificial intelligence, despite the miracle that has come, is only a fraction of what artificial intelligence will be able to do in the future.

More of the topic in Hayadan:

• Sam Altman CEO of OPEN AI: We need to build a control mechanism for artificial intelligence like the Atomic Arms Control Council (video)
• Sam Altman, CEO of OpenAI - the developer of ChatGPT has signed a new open letter warning against the dangers of artificial intelligence
• Five reasons why robots won't take over the world