At Arthur, research and development is fundamental to our identity. Our dedicated team of researchers continuously stays ahead of the curve in machine learning, informing our product development and business strategy. Collaborating with data scientists, ML directors, and AI Center of Excellence leaders worldwide, we’re proud to bring real-world solutions from the lab to the boardroom.
With that in mind, we’re thrilled to share the recent conversation we had with Ben Feuer, Arthur’s ML research fellow for this summer.
Tell us a little bit about your background and research interests.
I’m currently a Ph.D. student at New York University studying deep learning! My research is divided between experimental studies designed to help the community better understand what AI is (and isn’t) capable of, and new learning algorithms to make AI more accurate and consistent.
What interested you most about working at Arthur?
Arthur is such a special company, and I am really excited to be working here this summer! Aside from the fact that the office has an adorable “resident therapy dog” (heart emoji!), I love that Arthur is committed as a company to conducting cutting-edge research. Arthur’s Chief Scientist, John Dickerson, has a very deep research background, and he has worked closely with me throughout my internship to help point me in the right direction. I have enjoyed the friendly and welcoming atmosphere—the team at Arthur has really made me feel at home!
What’s the most exciting/promising field in machine learning right now, in your opinion?
If you’re interested in a field with widespread practical applications, it’s pretty hard to beat LLMs (large language models) at the moment. The commercial coming-of-age of ChatGPT has made them nearly synonymous with AI, at least for the moment. But the ML field with the most widespread medium to long term promise is almost certainly robotics and embodied ML. As robots start appearing in our stores, on our streets, and in our places of business, it’s really going to start changing the way people relate to the technology.
Any interesting hot takes?
Although I’m not in the Blake Lemoine camp, I do believe that the point at which advanced AI achieves the functional equivalent of sentience might be closer than many people believe. By functional equivalent, I mean that industry might soon develop AI which humans treat more like an equal, and less like a product.
Character.AI, a startup which trains specialized chatbots with distinctive personalities, has proven to be one of the stickiest and most enduring applications of LLMs so far, and there are many stories out there of people developing meaningful (to them, at least!) relationships with the bots on the site. If it listens like a friend, and it talks like a friend, and it remembers who you are and what you’ve been up to, then at a certain point, I think it’s only natural to want to reciprocate.
Then, if you see someone treating your ‘friend’ really badly, or some company deletes your friend’s memory so that it can no longer remember who you are to save on cloud hosting costs, I think you might respond to that more negatively than, say, having Google change the Maps UI in a way they find less intuitive. Companies are going to have to become really careful stewards of their chatbots and embodied AIs; that, I think, is the practical upshot.
At any other point in human history, what I’m talking about would be the stuff of science fiction. We really live in an amazing time!
SHARE
