The LLM Paradigm Shift: From Determinism to General Intelligence

Please join us for the Orange County ACM Chapter's bimonthly evening program series. The LLM Paradigm Shift: From Determinism to General Intelligence Agenda 6:30 PM Doors Open & Networking 7:00 PM Announcements, Presentation, and Q&A 8:30 PM Meeting Adjourns Event Details The path to modern artificial intelligence is defined by a fundamental shift from hand-coded logic to large-scale emergent systems. This talk surveys that journey, beginning with a brief look at early deterministic AI from the 1950s through the 1980s: rule-based engines, expert systems, and symbolic logic. The focus shifts to deep learning in the early to mid-2010s, tracing how word vectors turned language into math, followed by recurrent neural networks and sequence-to-sequence models for language tasks. The defining pair of breakthroughs that unlocked modern AI arrived with the 2017 Transformer architecture and the late-2022 release of ChatGPT. We explore how Transformers introduced self-attention to process entire sequences at once, aligning with GPU scaling to train on massive internet data. We then examine how scaling gave rise to emergent intelligence, and how ChatGPT paired that architecture with instruction tuning and reinforcement learning from human feedback, turning raw text predictors into intuitive conversational assistants and driving mainstream adoption. The talk concludes with a brief forward look at the emerging frontier: how models are extending beyond chat interfaces toward autonomous reasoning and tool execution Speaker Johnny Lin is a Principal AI Scientist and Architect with over a decade of post-doctoral experience spanning applied statistics, data science, and industry research. He currently leads enterprise AI and machine learning initiatives at Constellation Energy, having previously served as a Senior Data Scientist at NRG Energy and a Research Scientist at Amazon. He holds a Ph.D. in Quantitative Psychology from UCLA and a bachelor's degree in…