To be fair, if an industry consortium bought 1 trillion USD worth of semiconductor parts to prevent the AI winter of the 1970s, it would have looked like a foolhardy move unless some group of researchers happened to get drunk in the correct bars and be sober for the correct investor meetings, leading to early discovery of attention and leapfrogging even today's AI research to achieve several orders of magnitude efficiency gains in training so as to satisfy the eager compute capabilities of the 1970s.
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments