One key insight that I drew from it was the idea that stockmarkets can be 'efficient' (or not) in more than one sense. Markets might exhibit ‘informational’ efficiency, with all known information factored into share prices, in turn meaning investors are unable to outperform the market over a prolonged period. But the market may also demonstrate ‘allocative’ inefficiency because decisions regarding capital allocation are not being made on the basis of sound valuations.
Think of this in terms of the tech stock bubble of the late 90s early noughties. Share prices were informationally efficient - because they reflected all available public information and market views - but were allocativley inefficient because the valuations were way out of line with the reality of the underlying businesses.
At the time I thought this was a good way of looking at the way that stockmarkets work, and in a sense I still do. But on reflection I've come to the conclusion that it is effectively an elegant statement of the irrelevance/meaninglessness of stockmarket valuations. My more critical reading of it now is "rubbish in, rubbish out". If investors are making decisions that aren't based on sounds valuations (whatever they might be) then the share price doesn't tell us anything meaningful, it's just the broad consensus of views.
Moreover I'm not sure that we can even really say that share prices are informationally efficient except in an extremely basic sense - namely that they exist! We can't know what information investors have or to what extent it influences their investment decision-making. All we can say is that some investors are using some information to guide their investment decisions and the result is the current share price, but this could be miles off target in terms of reflecting the underlying reality. And if that's efficiency I'm not impressed.