Well, it's not very nice to talk about people spending decades on failure, but I guess I can give some examples.
Cycorp still exists, from 1984. However, D. Lenat's approach to AI by ontology engineering has basically been completely infertile after about the early 90's.
Feigenbaum's expert systems stuff was a basic bust, it led to the Japanese just throwing that stuff away. People spent an incredible amount of effort and time systematizing expert knowledge and making expert systems and it was not a happy time. Much of that knowledge went into probabilistic forms, culminating in the Bayes net. The most famous application of Bayes net: Clippy (there are a lot more successful applications, but still...)
It was believed shortly after the AI conference that computer vision could be solved by a summer project in the 50's. That didn't happen.
Minsky and Papert gave a criticism on single-layer perceptrons in '86 where they proved that they could only make linear discriminators and therefore were useless for any real practical purposes harder than the XOR problem. They were wrong, given that we call multi-layer perceptrons neural networks.
Simon and Newell made their model and thought that models like theirs with production rules would point the way towards the way that humans could systematize thought. That didn't happen, although they had some cool papers.
People saw ELIZA and SHRDLU and thought that good NLP was coming in only a decade or so.... in the 60's.
Beveridge report. "The spirit was willing, but the flesh was weak" to "The vodka was good, but the meat was rotten." (that last one's a bit apocryphal, but still)
There was a huge and abiding torrent of neural net stuff that dealt with evolving topologies in late 90's. I see very little of it in any way shape or form in industry or academia today, because it's a lot of computation for basically no gain.
They thought that layerwise pretraining of neural nets was the way to go in 2006, before they realized that initializations, normalization, and better activations was the better way.
A disgusting amount of why Watson won Jeopardy was because it could buzz faster than Jennings and Rutter. Ain't that nice?
The skill-cap in Jeopardy is sort of low. The top players can all answer almost all questions, so victory comes down to the buzzer even between Jennings and Rutter.
The important thing is that Watson hit that skill cap. From there it wins on tie-breaks every time. I think we'll see this dynamic in many human/AI contests. If both competitors' skills are at the saturation point, the contest is decided either by luck, or some strategically unsatisfying thing like diligence or mechanics. I don't see why humans will ever have an advantage at this.
Is pretraining really all that much of a failure? I haven't really found an authoritative answer on whether or not pretraining is worth it these days. Hinton's 2012(?) Coursera course still focuses pretty deeply on generative/layer-by-layer pretraining with RBMs but I'm just not really sure if that's fallen by the wayside today. Or maybe it's still useful only in specific circumstances?
Saxe Ganguli McClelland, 2013, about linear nets and orthogonal initialization. But then, read Li Jiao Han Weissman 2017 (maybe preprint), "Demystifying ResNet", which makes a nice claim about the niceness being conditioning of Hessian at init.
Tldr: it's good conditioner but you can do better ab initio
I still wonder if the many who spent years on obtaining the Loebner would be considered a "failure" in this domain ... https://en.wikipedia.org/wiki/Loebner_Prize - I think no matter how much you read about AI or look into the rabbit hole you will always end up in some type of Chinese room argument ... https://en.wikipedia.org/wiki/Chinese_room
After listening to him for a bit, reading some of his books, annoying him a fair amount, I think that my opinion of J Searle is that he doesn't know jack shit about AI.
Gedankenexperiment as a methodology has had considerable success in physics and miserable, complete, ridiculous, awful failure in psychology and cognitive science.
