New chair remains focused on opportunities in computer science

As the new chair of �鶹��Ѱ��Boulder’s Department of Computer Science,��Sriram Sankaranarayanan will admit it’s been a challenging year for the field.

In the 25 years he’s been a computer scientist, there were “dream problems” he and his colleagues wanted to solve — creating intelligent agents that could understand speech and synthesize natural language, for example.��

“Within a small, really tiny space and time, they got solved to an extent where computers perform amazingly well with these problems,” he said.

While the immediate outcome of that has been layoffs, graduates struggling to find jobs and generally a lot of uncertainty about the future of the field, Sankaranarayanan also believes it’s creating exciting new problems to solve.��

“Is there a better time to be a computer scientist? I don't know of any better time,” he said.��

New frontiers for AI

Sankaranarayanan compared today’s large language model AI systems to an aircraft. If you approach a mechanic and ask what a particular wire on their plane does, they can answer confidently.��

Sriram in his lab with students sitting at desks in the background

But with AI, “we managed to solve the problems in a way that we just don't understand the solution as we would with other kinds of engineered artifacts,” he said.��

That creates an opportunity to really dig into how AI works. It’s also leading to an enormous amount of “technical debt” that will have to be dealt with, he said. As billions and billions of lines of code are written with AI, at some point, decisions will have to be made about how to maintain that code.��

“The industry is going to figure out the optimal ratio of people to AI,” Sankaranarayanan said. “I feel like the companies who are going all in on AI have gotten it wrong. I think they will need more people, especially people who are really good at troubleshooting, who are good at engineering for performance, engineering for reliability and engineering for security.”

Making software safe, secure and reliable is his research area, and he can attest to the difficulty of that particular challenge. There are also fundamental computing problems — like�� — that we don’t know how to solve efficiently.��

“The things we know are fundamentally hard will remain so, even with AI.”

[The industry] will need more people, especially people who are really good at troubleshooting, who are good at engineering for performance, engineering for reliability and engineering for security.

Campuswide capacity

When it comes to new fields like quantum computing, Sankaranarayanan said the problems waiting to be solved get even more interesting.��

“The kind of computing systems that I encountered as an undergraduate student more or less are the same kind of computing systems we work with today,” he said. “It's just that they have gotten faster. They've been miniaturized. But fundamentally they have the same working principles.”

Quantum computing will require entirely new computing systems before its advantages can be realized, he said. Extend that to new secure, reliable software that will have to be developed for those systems, and you’ll be busy for the next 25 years.��

“All of these are changing rapidly, and computer science is on the receiving end of an enormous amount of investment from the industry and the federal government in terms of AI research and research into quantum computing,” Sankaranarayanan said.

�鶹��Ѱ��Boulder’s history in quantum science and engineering make it the perfect place to tackle quantum computing, he added. He hopes his department can continue to build connections with other departments working on quantum challenges, like physics and electrical and computer engineering.��

The already-interdisciplinary nature of the department also makes it a great partner for researchers across campus as they grapple with the effects of AI on their disciplines.��

“That's one thing I'm super keen on developing and enriching — having an interdisciplinary, outward-focused computer science department that is available to everyone who's looking to navigate the big disruptions,” Sankaranarayanan said. “As AI spreads across all the sciences and engineering and humanities, we want to be there for the campus as a set of experts who can talk to pretty much anyone on campus.”

Systems of support

The silver lining of a tough year for the computer science field? The department now has some breathing room to look at ways to better support students who are choosing it, from curriculum changes to new degree options.��

“When we were having huge enrollments three years ago, there was really no way we could make structural changes of the sort that we are talking about now,” he said. “We are even more motivated to support our students.”

Sankaranarayanan said the faculty are also discussing how the department’s alumni network can be better activated to support recent graduates in their career searches.��

Overall, Sankaranarayanan is excited about the next incoming class and the passion they are bringing to their education.��

“I don't have to tell them why they need to learn computer science,” he said. “They are coming in a time when graduates are struggling to find jobs, and yet they're choosing our field. They are true believers.”

�鶹��Ѱ��