The two main things are
1. You don't actually know your investing horizon and thus can't really match investments to actual liabilities (could be a lot shorter than you think for some money). It's not like you know how large of an emergency fund you'll ever need so personally I don't segregate out accounts and do AA based on all assets
2. You may not want to invest for the most likely outcome but hedge some based on a range of unlucky outcomes or adjust for the possibility of unfavorable structural changes
When people think about stock risks these days they typically imagine a progression like Scenario A, with high volatility:
But much worse would be Scenario B. Could well be worse than that long term, as seen in some of countries, even outside occupation/WWII. In Italy the last 50+ years, bonds have broadly outpaced stocks, for example.
There's a very wide cone of uncertainty with respect to the long-term outcome of stocks. In most, we think, the results range from decently positive to great. Historically in the US and in most countries that's what we've had, and assuming things kind of chug along and the market is pricing in long-term trends somewhat reasonably and there aren't huge, disruptive shocks, that seems very reasonable. No guarantees.
Of course, the opportunity cost of hedging stock returns by reallocating elsewhere may be more than you're willing to pay, and that's a reasonable position if you make it eyes wide open. Also, for all the uncertainties about stocks, bonds are not that good either. There's also direct real estate (among other things), though frictions are high and you can't readily diversify there.
In any case, even if you think stocks are the best long-term bet among asset classes, that doesn't mean the optimal bet for your purposes is to go all-in on the one best idea.
Now, you could also use leverage so you can diversify sources of returns while still aggressively seeking returns and keeping high equity exposure, but few like to go that route.