Though they may (or may not) be cheaper than usual, it's likely that the comment would be primarily along the line that they are the cheapest they have been relative to the valuation level of large caps. That changes the implied optimal investment strategy.
...
Are you saying they aren't cheaper than usual? How does it change the implied optimal strategy?

No, I was explicitly avoiding that question of whether small caps are cheaper than their own historical average. My point is that there are potentially different situations:
* If small caps are cheap (and as good an investment as they were historically) then they're probably a buy.
* If they are merely cheaper than large caps (and as good an investment as they were historically) but both are expensive, then probably neither is a good buy.

The problem with US small caps these days is that they aren't the same quality of company that they used to be. This is a very interesting article outlining the observation, though probably behind a paywall. Maybe try searching "Small stocks big problems".
Financial Times: Content Abfbf19e f963 4c1b b69e 7bef8896e8cd

Some observations from that: in 1995 only about 5% of small caps had negative forward earnings, now it's about a third.
The median ratio of gross profits to assets has fallen from around 29% to around 19%. (excludes healthcare, so as not to impose an undue penalty for more common profitless biotechs)
The balance sheet is worse: debt to earnings has risen from around 2x to about 3x.
There is a whole lot of small cap junk, very much more than was historically typical.

One possible narrative explanation is that all the reasonably decent-looking smaller firms have been snapped up by private equity or been acquired by bigger fish...only the less desirable up-and-comers have gone public.

I tend to think of the market as being made up of small caps, large caps, and a tiny number of truly gigantic firms. Ignoring the giants for a moment, the large caps are generally better businesses than the small caps, especially in recent years, so if the small caps are lagging them in market performance then that is not necessarily a buy signal for small caps. The giants (which dominate the S&P cap weight) are so few in number that they aren't statistically predictable one way or the other, even though they are the ones that determine whether the S&P 500 is leading or trailing lately. So any size comparisons with potential predictive power should probably leave those out. e.g., compare the Russell 2000 to the S&P 490 or to the S&P 500 Equal Weight.

Jim