The End Is Not Clear

In his January 2023 Communications Viewpoint, “The End of Programming,” Matt Welsh wrote “nobody actually understands how large AI models work.” However, already no one person understands existing large computer systems. Indeed, no team of people understands them. Staff turnover and other practicalities of real life mean not even the team that wrote them originally (should it still exist) nor the team currently responsible for maintaining them, fully understands large software systems, which can now exceed a billion lines of code. And yet such systems are in worldwide daily use and deliver economic benefits.

W.B. Langdon, London, U.K.

Back to Top

Author’s response:

It is true that modern software systems are incredibly complex, but fortunately, with the exception of programs written to be intentionally oblique—such as in an esoteric programming language like Malbolge or Whitespace—it is, in fact, possible to directly inspect and understand the source code of such software directly, either by hand or with the aid of automated code analysis tools. A large AI model, on the other hand, is essentially just a bunch of floating point numbers. Just as it is impossible to untangle the operation of the human brain by inspecting the neuronal connections under an electron microscope, AI models are not amenable to direct scrutiny.”

Matt Welsh, Seattle, WA, USA

Back to Top

Theory of Relativity

In his response to Robert Glass’s comments in the February 2023 Communications Letters to the Editor section, Editor-in-Chief James Larus asks the question: Why are more practitioners not members of ACM? I developed software for 44 years and have been a member of the ACM less than half that time. I have mentioned ACM to my workmates and distributed articles I thought would be of interest. My colleagues did not see the relation between the articles and their work. Admittedly, many developers were working on IBM mainframes. Developers had in-depth knowledge of z/OS and its subsystems. I cannot recall any ACM articles that focused on that platform. What does ACM offer those developers that write business applications on z/OS?

I was fortunate to have worked on OS/2, Linux for z, Windows, and z/OS. I have learned, and programmed in, new languages. I have written code that implemented classic algorithms and then found uses for them in my work. I got involved in the Microsoft TEALS program and found the articles on CS in education enlightening. Many practitioners work in the same area for their entire career. Many move into management roles. How would membership in the ACM benefit them?

Personally, I have found membership in the ACM to be of great value. I do not think I am representative of the world of practitioners.

Walter Falby, Detroit, MI, USA

Back to Top

Editor-in-Chief’s response:

Stay tuned! The forthcoming Communications website will allow more space for articles, and Communications will use this to publish more material of interest to practitioners.

James Larus, Editor-in-Chief, Communications of the ACM Lausanne, Switzerland

Back to Top

Binary Balancing

Moshe Y. Vardi’s May 2023 Communications column presents two fundamental—and inseparable—problems. He wants ACM to explicitly commit to the public good.

Problem 1 is there is no agreed-upon definition of the public good. Is pro-abortion part of that good, or is pro-life? Is fare-beating in the subway part of that public good (as the May 2023 Communications Law and Technology column by Kendra Albert and James Grimmelmann Viewpoint implies)?

Problem 2 is Vardi attempts to hijack the sole primary purpose of ACM, which is to advance computing.

There are far too many political activists today insisting all groups must support their agenda, or else. This just leads to polarization and divisiveness, which are clearly growing in society. ACM must not let this happen.

Alex Simonelis, Montreal, Quebec, Canada

Back to Top

Author’s response

Alexander Simonelis does not like my claim that the higher purpose of ACM is to serve the public good. He is not the only one. I have been practically accused of being a Marxist-Leninist because of my advocacy of the public good.

In response, I would point out that positioning the public good as the higher purpose of computing as a profession is explicitly stated in the ACM Code of Professional Ethics. Advancing computing as a science and a profession is therefore a means, not an end. Simonelis also bemoans the growing divisiveness of society. But there is strong evidence this divisiveness has, to a significant degree, been brought about by information technology. So Simonelis is essentially saying that to avoid divisiveness we should ignore the role of technology in fomenting societal divisiveness.

I agree with Simonelis when he says we do not have an agreement on how to best serve the public good. As an example, the Preamble of the U.S. Constitution essentially says the higher purpose of the United States is to advance the public good, and the Constitution is the mechanism by which the citizens of the U.S. reach agreement on how to serve the public good. ACM is an association of members, so the members, collectively, will have to decide on how ACM should serve the public good.

Moshe Y. Vardi, Senior Editor Communications of the ACM Houston, TX, USA