Remembering Cleve Moler (1939 – 2026)

For the past week, I’ve been hearing a distinctive, unforgettable, booming voice in my head. One time, the voice talked about “God’s pi.”

Another time, it posed a puzzle: “I’m thinking of two numbers whose average is 3. What are the numbers?”

Later, the voice offered this dryly skeptical comment: “Well, if you don’t need the answer to be correct, I can give it to you as fast as you want.”

“Booming voice” is one of the most common physical descriptions of Cleve Moler, creator of MATLAB, and co-founder of MathWorks. I worked down the hall from Cleve in my early MathWorks years, so, because of that voice, I heard half of every conversation in his office.

Cleve passed away on May 20, at age 86. Since then, people all over the world have been remembering him for his influence and impact on their lives. I have been deeply moved by the outpouring of feeling expressed in social media such as LinkedIn.

People have shared their experiences of how Cleve has influenced their professional lives and careers. I’ve been struck by the range of ages among the people remembering him. Some knew and worked with him in the 1970s, and some were just starting their careers in the 2010s. Many rather young-looking engineers have posted their selfies that were taken with Cleve. The smiles in all those pictures demonstrate the truth of what Ned Gulley said this week – “He cared, and you felt it.”

My first interaction with Cleve was in the Usenet newsgroup comp.soft-sys.matlab. This forum was brand new in the winter of 1993, when I made my first cssm post, about a bug I had found in MATLAB. Cleve graciously responded within a few hours with an explanation of the cause and the anticipated fix for the bug. Cleve loved the newsgroup and the way it allowed him to interact frequently with MATLAB users all over the world. I saw this happening from afar, and I thought, wow, how cool it must be to work at MathWorks with people like Cleve.

I will remember Cleve as a world champion story teller. He had an inexhaustible supply of stories, and he told them in every way imaginable. In the dining area where most MATLAB developers lunched together in 1994, he regularly captivated the room with tales of earlier computer eras, or of his experiences working on some of the earliest parallel computing machines.

Cleve told many of his stories in Cleve’s Corner, a regular MathWorks newsletter column that began in 1990, when the newsletter was still being printed and physically mailed to MATLAB users around the world. (Thanks, MathWorks, for maintaining an archive of these columns.) The beginning of the second Cleve’s Corner, “The World’s Simplest Impossible Problem,” is a good example of how Cleve imparted personality and wit to mathematical concepts:

The other day at lunch with a couple of other MathWorks people, I posed the following problem:

“I’m thinking of two numbers. Their average is 3. What are the numbers?”

Now stop a minute and, before you read any further, provide your own answer to my question.

Of course, my problem doesn’t have a “right” answer. The solution is not unique. The problem is “underdetermined” or “ill-posed.”

Most of the people I was having lunch with wouldn’t give me an answer. When I pressed them, one person finally said, “Both numbers could be 3.” Another person said, “Yes, but one could be 6 and the other could be 0.” That’s what I hoped they would say. In some sense, these are both “nice” answers. Nobody said “23 and -17,” or “2.71828 and 3.28172.” Those would have also been correct, but not as “nice.”

What does this have to do with MATLAB? Well, MATLAB doesn’t balk at impossible problems. It will solve this one. And it does it without complaining like my buddies at lunch.

Naturally, the problem is a matrix problem.

Cleve’s Corner made the jump over to MATLAB Central blogs in 2012, with this post about Fibonacci matrices. Between June 2012 and April 2026 he published 342 Cleve’s Corner blog posts, averaging a post every other week.

He blogged about interesting matrices (magic squares; a million dollar matrix), prime numbers, and Pascal’s triangle.

He told so, so many stories about the generation of mathematicians and computer scientists (one of his final posts was about Velvel Kahan) who invented whole new ways to solve matrix and linear algebra problems.

He told stories about novel applications of matrices, such as transforming the way everyone used the Internet (Google page rank as “The World’s Largest Matrix Computation”), or identifying animal species in trail camera photos (“An Experiment in Deep Learning with Wild Animal Trail Camera Data”).

He told us about new ways to analyze the human body (“The Eigenwalker Model of the Human Gait”).

He crafted many, many MATLAB animations to tell us about mathematics. One of his favorite uses of animation was for explaining eigenvalues and eigenvectors.

He was even interviewed on a national news broadcast once, to tell Americans all about a computer chip that sometimes divided numbers wrong (“The Pentium Papers”).

And Cleve told us all the story of MATLAB. I never tired of hearing it. He explained the origins and evolution of MATLAB to generations of MathWorkers, and to engineers, scientists, and mathematicians in professional conferences all around the world. When MATLAB was recognized in the Fourth ACM History of Programming Languages Conference, Cleve, along with co-author Jack Little, shared “A History of MATLAB”, a 67-page paper. If you want to see one telling of Cleve’s MATLAB story, check out this “Evolution of MATLAB” video.

To the engineers and scientists who are remembering Cleve this week, I want you to know that Cleve was always eager to hear (and then retell, often) your own stories. Cleve was endlessly fascinated by the work that you are doing. He cared, and he loved it all.

I’ll leave you with Cleve’s final challenge from “The World’s Simplest Impossible Problem,”:

Okay, now I’m thinking of three numbers whose average is $\pi$. What are the numbers?

Hey Cleve – we miss you, and we miss your stories. Thank you for everything.