Machine Learning Street Talk · Machine Learning Street Talk

The History of AI

·12 min·1 clip

AI started before powerful computers, with 1940s thinkers believing machines could mimic human intelligence.

Jaden Schaefer opens Machine Learning Street Talk with a solo historical survey aimed at listeners who spend enormous time and resources on AI but may not know the field's origins. He begins with the 1950s observation that even before powerful computers existed, philosophers and mathematicians were asking whether machines could think, rooted in the idea that reasoning is fundamentally reducible to mathematical logic. The 1956 Dartmouth Summer Research Project on Artificial Intelligence, convened by John McCarthy, Marvin Minsky, Claude Shannon, and others, is widely regarded as the birth of AI as a named field. Early researchers were wildly optimistic, predicting human-level reasoning within two decades and believing problems like vision and language were nearly solved. The reality was that rule-based systems, built on if-then logic, worked only in narrow, controlled domains. Chess programs played well by memorizing positions; they broke entirely when applied to real-world ambiguity. This gap between promise and capability led to the first major funding cuts, known as the first AI winter, following the Lighthill Report in the UK in 1973. A second wave came in the 1980s with expert systems, which encoded domain-specific knowledge in large rule databases and found commercial applications in medicine and finance. But these systems were brittle, expensive to maintain, and could not generalize. Another winter followed in the early 1990s. The episode covers the shift to statistical machine learning, where instead of encoding rules, researchers let algorithms find patterns in data, which scaled much better with computational power. The deep learning revolution, enabled by GPU acceleration and large datasets, finally delivered on many of the original 1950s promises. Schaefer frames AI history as a recurring cycle of hype, disappointment, and eventual breakthrough, suggesting awareness of this cycle is valuable for evaluating current claims.

As heard by us

A brisk tour of AI's shift from rules to neural nets.

Jaden Schaefer lays out the arc of AI from symbolic systems to machine learning, then on to the data, GPU, and training gains that made modern models workable.

Read the full review in PlayNext →

Why you'd press play

Want the backstory behind AI without getting buried in jargon?

Read the full recommendation in PlayNext →

Listen to the show on

Apple Podcasts Spotify

Play clips from this episode More from Machine Learning Street Talk