Agree with many others here that
talented early startup employees don't tend to make enough money to offset the opportunity cost of FAANG-like total comp. That point has been argued in the thread above more eloquently than I could.
If you accept that compensation is the issue, and want to fix it, and get our best and brightest back into startups, that has to be corrected somehow. I can only think of 4 sources:
A) Pay them more, shortening operating runway
B) Make their equity worth more, by making founders' worth less
C) Make their equity worth more, by making startup companies worth more
D) Make their equity worth more, by making investors' worth less
I'm going to argue as a current startup executive that (A) is already efficiently calibrated by the market and has yielded the current balance, so not likely to find more ground there.
The current supply/demand of founders would suggest (B) is not likely to make a dent either. If it could, that would suggest that there is an oversupply of (qualified) founders, such that we could do with fewer. I think most VCs would disagree with that world view, as most of them take 100s of meetings to do a single deal, and generally consider capital deployment to be their primary operating limitation.
If someone has serious ideas for (C) that can make a difference at scale, I think they would do well to share them here.
That leaves us with (D). Anybody who has negotiated a term sheet will tell you that this won't be easy. But I submit a humble suggestion for a cultural shift that I think could make a big difference here: make liquidation preferences unfashionable.
My reasoning is simple: in a modest outcome, the effect of the preference is much greater on employees (who lose out on their true equity value) than on investors. Most VCs have already hedged the downside risk within a given fund by diversifying over 10-20 other deals, and most of the time, at least one or two of them (if the fund is any good--and they won't be around long if not) will net a > 10x return. The value of the preference as downside protection is therefore quite limited.
On the flip side, it is a huge impact for an early employee, who might see an extra 50%+ dilution in a modest exit after the prefs get paid out. And let's face it... these modest exits are far more common than the unicorns.
Doing this would also free up better price competition among non-institutional investors who are happy to have non-controlling participation, especially in later growth rounds, and so could actually make the sector more attractive to a wider investor base.
This wouldn't produce an overnight change, but could result in more early employees having positive "EV" stories, which could eventually shift perception.