Replacing all of the lost jobs from the American auto industry I don't think is possible, namely because automobiles are both so incredibly common, are large-ticket items and the country's infrastructure is built for them. Henry Ford II said it well when he said to the United States Congress in 1975 that Ford was not the least bit concerned about the (then considered supposedly inevitable) transition from making automobiles to making mass transit vehicles, when he said that the country "has better things to do than to remake the country to suit 19th Century transportation." It's correct, as annoying as it is.
Anyways, technically the high point of the American auto industry in terms of market share after the introduction of imports was 1982, as the Big Three in 1982 controlled 86.9% of American auto sales, with 62.6% of that number in the hands of just General Motors and with Chrysler in the midst of the beginning of a major resurgence. The main problem was that the Japanese massively diversified in the 1980s, going from just small or smallish cars in the early 1980s to expanding their markets. Toyota was in the larger game with the Camry in 1983 and the Cressida in 1985, the Honda Accord steadily grew in size in the 1980s, Nissan introduced the Maxima in 1981 and Mazda, Mitsubishi and Subaru steadily followed suit, and that's before the Japanese luxury brands began rolling out - Acura in 1986, Lexus in 1989 and Infiniti in 1990. Slowing this is effectively impossible - Japan's growing prosperity of the 1970s and then the bubble economy of the 1980s made for all kinds of capital to undertake the development of such vehicles, and all of the luxury brands debuted at first with North American variants of their higher-end Japanese-market vehicles - the Lexus GS, for example, was a variant of the Japanese-market Toyota Aristo right from the start and stayed that way until Lexus got introduced to Japan in 2007. And that's before you get into the Koreans (who were horrible in 1980s but got better all the time, particularly after 1998 or so) or the Europeans, who were hit or miss - Volkswagen did just fine, Renault was looking good until their management decided dealing with AMC's problems wasn't worth the effort and the luxury brands all did well, but Peugeot-Citroen, Fiat and British Leyland all went from bad to worse, though Peugeot was savable even in the 1980s.
Marathag is more right than people know about the quality problems that afflicted GM. It wasn't merely them, either - everything about the system was meant to move cars out. The smart guys in that organization knew damn well that there was a lot to gain by putting more effort into vehicle quality - beyond the PR of it, better components and better assembly quality the first time would also reduce the warrant costs - but GM, whose board was dominated by financial men since Alfred Sloan ran the company in the 1920s, didn't think long-term, they wanted as much profit as humanly possible and they wanted it RIGHT FREAKING NOW, without any regard to what they were making. The engineers knew, but particularly after (as NothingNow points out) John DeLorean got wedged out of GM and Ed Cole retired, GM's top leadership truly stopped giving a damn about the products they were building, at a time when they desperately needed engineers to make the cars meet the targets they had to meet for emissions and fuel economy. When Soichiro Honda can show you up by designing cylinder heads to make somebody else's engine meet emissions standards that the company which built the car is fighting in court saying it can't meet them (true story, believe it or not), you look like a damn fool or somebody who isn't doing their job very well. The poor quality that all of the automakers had couldn't just be turned around by edicts or even better work by the assembly-line guys - if your components suck, better assembly in many cases won't make that much difference.
GeographyDude's idea of Deming's ideas getting seen early on in North America makes sense, but you'd have to get it working early on, far earlier than the late 1960s. What could really throw a curveball there (which I would have done had I decided to start back a little further with Land of Milk and Honey

) was to have AMC combine not just Nash and Hudson but also Packard and Studebaker, and have George Romney succeed George Mason as IOTL at AMC, and have a series of meetings between George Romney and Walter Reuther (the legendary UAW boss who led the union until he died in a 1970 plane crash) develop into AMC having a good relationship with the automaker, setting the standard for the automaker's future. Ed Cole does a good job developing GM cars in the 1950s and 1960s, and his management partner (DeLorean) does a similar deal with GM in the early 1960s, effectively forcing Ford and Chrysler eventually into line. AMC's better quality of products leads to better development of them, and as Ford and Chrysler have rather chaotic 1960s and 1970s AMC eventually overtakes Chrysler and draws level with Ford by the early 70s. Renault does deals with a strong AMC in the late 1970s (Chrysler follows in 1984-85) and both sides massively benefit from the deal, resulting in AMC-Renault being a true merger in 1987-88 being a true merger of equals, and the two of them together take over Nissan in 1998 to make a global automaker alliance. Using Deming's methods would be wise, but it requires a POD during a prosperous time, and doing so when the industry was still somewhat in flux in the 1950s makes the most sense. GM's immense size presents organizational issues, but in the 1950s GM was full of guys looking to make an impact both on the cars and the company. AMC shows the way, Cole and DeLorean see the light and organize the engineers and designers to take the company back from the bean counters and make fabulous cars. GM, having learned and applied such lessons, would effectively force Ford and Chrysler to keep up, but as both were totally controlled by either crusty old-school bosses or legions of short-term-thinking CPAs, the lesson would take some learning.
The idea of energy centers improving things is interesting, but it still relies on many orders and a big market. Its not real difficult to site nuclear power stations farther away from population centers, but the issue is getting the power to where it is needed. HVDC power transmission is very efficient at transmission of power long-distances, but is more expensive, relies on mercury-arc valves until the development of solid-state thyristor valves in the early 1970s and until the late 1970s isn't able to handle the 345kV power loads used on most main-grid transmission lines. Speed that up a little bit and you could put nuclear power stations well outside urban areas, but particularly in the industrial American Midwest which is dotted with hundreds of small-to-large cities this also has practical limits. Nuclear energy in America also has the problem of bad press (Three Mile Island and
The China Syndrome were the worst problems, but not by any means the first) and enormous cost overruns that bankrupted many power utilities and municipalities. Even if you get the ideal scenario - huge use of nuclear power in America and reactor makers and electrical producers (General Electric, Westinghouse, Combustion Engineering, Babcock and Wilcox, Foster-Wheeler, Bechtel, General Atomics) as major employers, you still don't get anything like the number of jobs produced as the auto industry, or its subsidiary industries in fields such as steel production. Doesn't mean its a bad idea, but its not gonna take over from a slumping American auto industry.
Driftless' point about the electronics and aviation also are valid, and I would add into it rail transport vehicles (America does alright in the freight world here, but in passenger vehicles the United States lags far, far behind the rest of the world) and shipbuilding (America had the shipyards for this all the way into the 1970s), as well as more radical ideas such as using segments of the knowledge economy to give big kicks to other industries. There is a reason most of the world's computer industry began in America and why so many American companies - Apple, Dell, Intel, Microsoft, Google, IBM, AMD, Cisco Systems, Oracle, Motorola and Hewlett-Packard, among others - dominate the world of telecommunications and computers. It's not that huge to have American companies like Atari dominate the world of video games instead of Sony, Sega and Nintendo, and companies like RCA, Zenith, Admiral and General Electric (or successors to them) could dominate the world of televisions and imaging technology. A lot of people don't know that Eastman Kodak was the developer of digital imaging technology, but they buried it for quite a while in an attempt to protect its sales of film. (Oops.) There are many ways of making America a dominant player in the world of electronics manufacturing, especially in more modern times as the processes become more automated and skilled technicians replace assembly lines owing to the increasingly-small size of computer chips. Shipbuilding I would say is also possible owing to the fact that American harbors are generally quite large (New York, Boston and San Francisco especially), the fact that the idea of using standard-size containers for shipping was an American idea and the fact that American steel mills have always been traditionally more focused on heavier grades of steel than are used in smaller-scale applications, but are ideal for shipbuilding projects.
HX87's point about gutting emissions controls IMO is a bit of a cover-up for American auto and oil companies. American fuel was (and still is) complete dog piss compared to what is available in Europe or many developed parts of Asia, and the primary reason for that is that American oil companies never had any need to make better refining processes to produce higher-quality gasoline fuel. The diesel units of many refineries have had to improve to make ultra-low-sulfur diesel fuel for emissions reasons, but the phasing out of lead in fuel (an absolute necessity owing to Lead's toxicity) simply forced Detroit's engines to deal with big drops in efficiency. Better-breathing engines and the ability to burn fuel more efficiently can be done with better cylinder head designs. Detroit didn't demand better fuel quality because they accepted the bigger-is-better idea, and for most cars using bigger engines to get around emissions-mandated reductions in compression is easily done. Hitting Detroit with both at the same time has the potential for greater effects, because improving the flow of a cylinder head can both improve power and emissions at the same time. Carburetors can be very imprecise, but marathag's point about the problems raised by early fuel injection systems is valid, too. If Detroit has a greater focus on efficiency early on and uses more four-cylinder engines, what might help there is individual carburetors tuned for the characteristics of each cylinder - fuel injection's big advantage was that you never had to worry about getting the right amount of fuel into the right cylinder at the right time. Individual carburetors can do the same thing. Its more complex, sure, but it's more likely to be reliable than many early fuel injection systems. There is a reason most European and Asian performance cars of the 1970s used multiple-carburetor setups. (The use of higher-pressure fuel systems could help, too - any American cars had low-pressure fuel systems until the late 1980s.) Having American refineries remove lead from fuel in the early 1970s as part of complete overhauls of their finishing units, which subsequently improve the standard of American regular unleaded fuel from 87 octane (the current standard) to the 92 octane or so standards of Europe (Japan's regular gas is about 93 octane) could also provide a major advance for efficiency by allowing higher compression. Combine that with multiple-carburetor setups and better cooling systems, you get the ability to improve compression ratios a sizable way towards leaded fuel numbers, improving efficiency. Another big curveball could be somebody in Detroit putting considerably more effort into turbocharging, but turbos through carburetors presents all kinds of issues with engine behaviour (particularly off-boost) and early turbos had real issues with light-switch power delivery. Solvable issues, but they take a considerable amount of engineering effort.