3 ms·
The errors, insights and lessons of famous AI predictions (2014) [pdf]
- olooney 8y ago> In Searle’s time, the dominant AI paradigm was GOFAI (Good Old-Fashioned Artificial Intelligence.) Russel and Norvig's book is probably the best introduction to "old fashioned" AI: https://www.amazon.com/Artificial-Intelligence-Modern-Approach-2nd/dp/0137903952/ https://www.amazon.com/Artificial-Intelligence-Modern-Approa... GOFAI may not have lead directly to true AI, but it produced a ton of useful algorithms such as A* and minimax. Although the attention has turned to machine learning algorithms (à la https://www.amazon.com/Elements-Statistical-Learning-Prediction-Statistics/dp/0387848576 https://www.amazon.com/Elements-Statistical-Learning-Predict...) the hybrid of GOFAI and ML has produced some extraordinary results, such as AlphaZero: https://deepmind.com/blog/alphago-zero-learning-scratch/ https://deepmind.com/blog/alphago-zero-learning-scratch/
- bcatanzaro 8y ago“Moore’s law seems much more counterfactually resilient: assuming that the current technological civilization endured, it’s hard to find any reliable ways of breaking the law.” That didn’t age well.