This is why we stopped giving employees equity pretty early on. We got to a pretty decent headcount before we sold (probably above the median for post-B-round YC companies). We just paid cash and bonus.

I don't often like Fred Wilson's compensation writing, but here I think he has the problem basically nailed down: employees (rationally) don't prize equity, and giving it to every employee is in the long run extremely expensive.

Two thoughts:

You don't have to give shares to every employee to give it to some of them. The first shares I ever got that were worth much (Secure Networks, back in 1998) were shares I had to opt in to, at the cost of some salary. I had coworkers who chose not to do that. We were all generally happy with the outcome.

People on this thread are making comparisons to 300k/yr Google salaries. For journeyman developers in SFBA, this is probably a reasonable comparison? (I mean, the median SFBA developer isn't making that, but you could argue that's because they're being underpriced in the market by the practice of issuing shares that will never be worth anything). But for employers, it's a pretty silly comparison. If you're a startup paying 300k/yr to someone who does work a CS graduate out of school can even theoretically do, you're doing it very wrong. You might have to adjust your hiring filters to get your cash compensation down to a reasonable numbers, but you should do that anyways, because filters that keep you stuck in the "competing with 300k/yr salaries" zone are, for almost every startup, a vanity.