Like actual AI (not the expert systems everyone is misleadingly labeling as 'AI' for marketing reasons), fusion power has been 20 years away for the last 60 years.
Those estimates aren't particularly insightful. In the case of fusion, a group could make a manmade net-positive fusion reaction tomorrow and we could still be many years away from mass commercialisation. The reason '20 years' is given when the real answer is 'we don't know' is to encourage investment. 20 years is the carrot on a stick used to get funding for the fundamental research required at this stage. I'd suggest the figure is chosen as it's at the outside edge of something that'll make an impact within the average investor's lifetime, though this is speculation on my part.
Also, when it comes to fusion, I suspect we'd be much closer to fusion energy generation if the bulk of funding hadn't been put into a single basket (the tokomak approach). There are promising fusion approaches that are closer to achieving net energy from fusion than tokomaks that are massively underfunded in comparison.
> There are promising fusion approaches that are closer to achieving net energy
While I agree there should be more funding for non-tokomak projects, this is simply not true. JET has reached 70% of breakeven and JT60 in Japan reached "theoretical" breakeven (if it had tritium). Nothing else has ever gotten close.
"GIT-12 Z-Pinch Achieves 6 Joule Fusion Output with Deuterium Fuel"
"Z-pinch devices are closely related to plasma focus devices in that they also use the currents through the plasma itself, not external magnets, to produce the strong magnetic forces that confine hot plasma. In a z-pinch, however, the current flows between two electrodes in a line, rather than the plasma focus’s concentric electrodes. At the Prague conference, Russian and Eastern European physicists announced new experiments with the powerful GIT-12 z-pinch. They used a configuration in which the plasma was arranged in two concentric rings, puffed into the gap between the electrodes right before the discharge. Nearly 6 J of fusion energy (producing 6×1012 neutrons) were released with a 3 MA current and 3 MJ of input energy to the device. Since only 500 kJ of energy was released from the capacitor bank prior to the pinch, the ratio of 1.2 J of fusion per 100 kJ of input rivaled or slightly surpassed the best results obtained in either plasma focus devices or tokamaks, both of which are close to 1 J per 100 KJ input."
The implication is that both Z-Pinch and DPF devices are close to tokamaks in terms of energy efficiency, based on the experiments performed to date. Do you disagree with this? If so, why?
The years after the Dartmouth conference were an era of discovery, of sprinting across new ground. The programs that were developed during this time were, to most people, simply "astonishing": computers were solving algebra word problems, proving theorems in geometry and learning to speak English.
Few at the time would have believed that such "intelligent" behavior by machines was possible at all. Researchers expressed an intense optimism in private and in print, predicting that a fully intelligent machine would be built in less than 20 years.
That brings the earliest predictions to the range 1976-1994.
(Textual description of the part of their dataset that includes reasoning as to how the prediction was reached at http://aiimpacts.org/list-of-analyses-of-time-to-human-level.... That list is a bit less optimistic. Excel sheets for the full dataset also are available from that site)
In a way, yes. One popular belief was that you basically take world modelling and association tables and pile them onto a complicated logic system, until it magically understands everything and therefore becomes intelligent. This was a notion among actual researchers (especially in the 1950s/60s).
If you've just made the most complicated computer system (see e.g. the General Problem Solver), it's easy to believe that you just need to pile on some more complexity until you have a thinking thing.
