The underlying value of a stock is the net present value of its future paid-out earnings. This is just a fact of the financial system, as fundamental to it as the conservation of energy is to the physical universe. It should have been mentioned in the article.
The reason P/E ratios vary is that companies' earnings change over time and companies go bankrupt. If you buy a stock with a P/E ratio of 20 years and then the company's earnings stay the same forever, eventually those earnings will find their way into dividends or buybacks, and your return on investment will be 5% per year, forever. If the P/E ratio is 10 years, 10%. If the P/E ratio is 50 years, 2%.
(If the company invests the earnings in assets, it doesn't pay them out in dividends or buybacks that year. But then the depreciation of those assets is deducted from its earnings in following years, so to maintain the same earnings, it would need to have higher earnings-plus-depreciation. So ultimately it all balances out.)
So when people rationally buy a stock with a P/E ratio of over 20 years, either:
1. They're expecting the company's earnings to go up by enough to put it back below 20; or:
2. They're expecting less than 5% per year return on investment, which probably means they're treating the company as a very-low-risk investment like commercial paper rather than like an ordinary stock; or:
3. They're hoping to find a bigger fool to unload the stock on before its price returns to what its earnings can rationally justify.
If you buy and hold a stock whose P/E ratio is 350 years, like Tesla today, and those earnings never change, you're getting an 0.3% annual return, forever.
Which, and forgive me because I'm not an expert here, sucks.
(Apologies for including the correct units on P/E ratios. I'm not the kind of person who thinks "vega" is a letter of the Greek alphabet.)