The fight between two approaches, Symbolic AI and Connectionism, has shaped the field of AI since the 50s. Is it possible that Symbolic AI held AI back by decades?
I learned about Neural Networks from a bad photocopy of 'Parallel Distributed Processing' Volumes and fell for them in ~1990s. The lecturer instructed us that the words "Neural Network" are forbidden. The 'cursed words' should not be uttered, and instead the term PDP (Parallel Distributed Processing) is to be used at all times. This was a tiny obscure faculty in an obscure country falling apart among much unrelated chaos. Only in retrospect I realised few things I witnessed, but didn't pay much attention at the time (had no other experiences to compare them with), were harbinger of things to come. Got reminded recently https://life.ieee.org/ieee-president-attends-plaque-unveiling-of-ieee-milestone-in-north-macedonia/. (Bozinovski was the Prof teaching PDP)
Silver lining - Chains-of-Reasoning or Chains-of-Thought (of currently latest-greatest Gpt-5) will be more along the lines of symbolics processing I presume? Somewhat fitting that 1/2, the Type-1 intelligence pattern matching is done by PDP NNs, while the other 1/2, for Type-2 intelligence - reasoning - it's looking like symbolics will have some role to play too. So it wasn't a total waste of time. To me symbolics always looked like top-down, while PDP looked bottom-up, and I was more sympathetic to the latter and disliked the former. (just random sentiment, contingent on personal circumstances)
Btw have you seen this delightful footage of McCulloch (of the 1943 McCulloch-Pitts neuron fame) https://m.youtube.com/watch?v=wawMjJUCMVw ? "A shirtless shoeless character of unflinching gaze looking like it's his last day on Earth, soaking a cigarette with gusto." would be a fair annotation by an AI. 😂 Born in the 19-th century no less!
I suspect the answer is - "Yes, it did".
I learned about Neural Networks from a bad photocopy of 'Parallel Distributed Processing' Volumes and fell for them in ~1990s. The lecturer instructed us that the words "Neural Network" are forbidden. The 'cursed words' should not be uttered, and instead the term PDP (Parallel Distributed Processing) is to be used at all times. This was a tiny obscure faculty in an obscure country falling apart among much unrelated chaos. Only in retrospect I realised few things I witnessed, but didn't pay much attention at the time (had no other experiences to compare them with), were harbinger of things to come. Got reminded recently https://life.ieee.org/ieee-president-attends-plaque-unveiling-of-ieee-milestone-in-north-macedonia/. (Bozinovski was the Prof teaching PDP)
Silver lining - Chains-of-Reasoning or Chains-of-Thought (of currently latest-greatest Gpt-5) will be more along the lines of symbolics processing I presume? Somewhat fitting that 1/2, the Type-1 intelligence pattern matching is done by PDP NNs, while the other 1/2, for Type-2 intelligence - reasoning - it's looking like symbolics will have some role to play too. So it wasn't a total waste of time. To me symbolics always looked like top-down, while PDP looked bottom-up, and I was more sympathetic to the latter and disliked the former. (just random sentiment, contingent on personal circumstances)
Btw have you seen this delightful footage of McCulloch (of the 1943 McCulloch-Pitts neuron fame) https://m.youtube.com/watch?v=wawMjJUCMVw ? "A shirtless shoeless character of unflinching gaze looking like it's his last day on Earth, soaking a cigarette with gusto." would be a fair annotation by an AI. 😂 Born in the 19-th century no less!