Assistant Professor Benjamin Lindquist focuses his academic research on historical development methodologies in computing. Specifically, the scholar examines the technical choices made during foundational periods of artificial intelligence development. As reported by MIT News, Lindquist "studies how engineers deliberately injected randomness into early AI" to shape computational behaviors.
These historical decisions continue to directly affect contemporary machine learning systems. The early injection of stochastic variables ultimately contributed to the foundational structures seen in modern architectures. According to the report, this intentional design choice played an essential role, "giving rise to deep-learning models that sometimes hallucinate." Consequently, the tendency of present-day systems to generate hallucinations reflects longstanding engineering traditions.