A Turing‑Award Legend on AI, Parallel Computing, and Learning's Future

Parallel Computing's Deep Impact on Computer Science

In the field of computer science, Jeffrey Ullman is an undeniable name. The 83‑year‑old Stanford professor co‑authored the famed "Dragon Book", helped found database theory, and won the 2020 Turing Award. His career spans compiler development, database theory, and algorithm research, shaping generations of programmers through classic textbooks.

In a recent interview, Ullman admits that technology feels like a young‑people's game and that he struggles to keep up with new tools such as large language models, phone navigation, and car‑system integration, illustrating how even a pioneer must adapt to the AI wave.

He reflects on the unpredictability of major tech shifts, recalling how in 1992 few mentioned the World Wide Web while discussing the "information highway". He notes that breakthroughs often emerge before experts anticipate them.

Looking ahead, he says the biggest change in software engineering was the move from machine language to high‑level languages, and now large models that generate code may surpass that impact. He sees software moving from “algorithm” to “algorithm + data”, with parallelism reshaping our view of computing.

Automation and the Role of Teachers

Ullman discusses his startup Gradiance, which aims to automate homework by presenting short multiple‑choice questions that also provide instructional feedback. The system uses “root problems” with multiple correct answers to encourage students to show their solution process rather than merely guessing a single answer.

He observes that massive open online courses (MOOCs) cannot replace teachers for most learners; automation can assist but not eliminate the need for human guidance.

Prompt Engineering and Human‑Computer Interaction

The interview delves into prompt engineering, describing it as a new discipline akin to compiling, where high‑level intent is translated into prompts for large models. Ullman notes the lack of concrete principles for effective prompting and the challenges of adapting such tools for older users.

He emphasizes the importance of designing interfaces that consider the habits and preferences of senior users, who often find modern technology difficult to adopt.

Future Directions

Ullman mentions interest in quantum computing and general AI, while expressing skepticism about practical quantum breakthroughs. He reiterates that the introduction of parallelism and data‑centric software represents a profound shift comparable to the transition from machine code to Fortran.

Overall, the conversation highlights the continuous evolution of computer science, the need for adaptable education, and the emerging role of AI‑driven tools in software development and learning.