"Part of the reason startups resonate with people is because the outcome is not guaranteed. If it were guaranteed, it would be boring... Whether or not you're the best person in the world at what you do, you're probably not going to alter Google's trajectory. But if you really want to benchmark yourself and see how much of a contribution and impact you can make--which is a really compelling prospect for a lot of the best people--a startup is a much better place to test that."

No smart person that you're hiring is under the illusion that you have a crystal ball into the future and that joining is a guaranteed thing. In fact, if you're telling them that and they select in, you shouldn't hire them because they didn't pass a basic intelligence test. I think it's important to tell them what's exciting and where you think the company can go. But also tell them where it will be hard and chart your best plan. And then tell them why their role can be instrumental

What I would discourage doing is whitewashing all of that. If people are joining your company because they want all of the certainty and safety of working at Google but also the perks of working at a small startup with lots of responsibility and transparency, that's a really negative sign.

Apparently in the early days of PayPal, Peter Thiel and Max Levchin would tell people after they interviewed all the reasons that the company would fail: "Visa and MasterCard want to kill us. We also might be doing something that's illegal. But if we succeed, we'll redefine payments."