You work hard because you're incentivized to do so: you're given raises if you do and fired if you don't. You're motivated almost entirely by the incentives the company provides you with. Because a lot of people at the company are the same way, at least a few of them will follow those incentives to the letter, benefiting themselves to the detriment of the company.
For example, if the company starts rewarding you for fixing bugs, someone will start introducing bugs in order to fix them, or focusing on bug fixes instead of their existing projects, or fixing bugs quickly but with unmaintainable hacks. If the company punishes people whose projects fail, people avoid risky projects. If the company rewards people who participate in meetings, work slows down while everyone frantically schedules meetings. I think it was Scott Adams who said "all incentives produce unintended behavior."
Modern large software companies mostly have their incentives sorted out by now, though. Systems of managers, performance reviews, peer reviews, project processes, and so on all align and balance each other out to induce a complex system of designers, managers, engineers, administrative assistants, operations engineers, and so on to work together for what amounts to the good of the company. I would compare it to an internal combustion engine: hundreds of parts and forces working around and against each other, but ultimately moving the vehicle forward at a steady pace.
For the individual piece in this machine, this has its perks. The piece doesn't have to care which way the company is driving- it just knows which parts it interacts with, and it needs only react to those. This piece might be an engineer who only needs to work as long as his performance incentives tell him to: if he's not penalized for going home after X PM, X PM is when he goes home. If he's not rewarded for working hard on something, he probably won't. Once he fulfills his obligation to his incentives, he goes home to his family free from concern. He wants the company to do well only in the vague way that one wishes a cashier "good afternoon." He'll work harder because it benefits him in terms of money or pride/recognition/etc, not because it furthers the company.
The downside, of course, is that he is subject to this barrage of incentives: rules, policies, programs, and review systems. A man is not a metal rod in an engine, and never fits quite right. Perhaps he'd be able to further the company more efficiently if he could ignore rules #34, #47, and #76. But rules are set up to govern everyone, and cannot be customized to every individual. The more people they have to cover, the more specific the rules must become.
When you work for a large company, you work with only your own interests in mind- this is both good and bad.
At a startup, you have one incentive: furthering the company. The smaller the company, the closer your incentives are to just "whatever you think is best." You could slack off, but if things go south, your equity and employment are materially affected. Further, unlike in a large company, your work can easily sway the success or failure of the company.
As compared to an engine, startups are a rocket*. There are a few different forces, but it's largely all parts pushing in the same direction. The complex balancing of parts is unnecessary. You shove the fuel of effort through the nozzle of product and try like hell to propel yourself upward.
The strength of this model, for the engineer, is that she can push as hard as she can. With few rules to limit her, she has the opportunity to figure out how she can best further the company. She is given the freedom to work as she wants.
The downside of this is that there are no limits. The freedom comes with responsibility. There's no policy telling you when to come in and when to leave, but when the site breaks at 2am, you need to fix it. There's no formal review process telling what you've done well and what you've done poorly, so you need to figure it out for yourself. You can work as hard as you want, so you work as hard as you can. Any less and you're not furthering the company, which devalues your equity and puts your job at risk if things go south.
At a big company, they'll force your interests to serve theirs, freeing you to pursue your own. At a startup, your interests become theirs, freeing you from restrictions. Neither path is right for everyone, and there is plenty of middle ground, but this is the dichotomy as I see it, and I can't imagine giving up the startup world.
* My understanding of rocket science comes from Estes rockets and the movie "October Sky", so let's consider a simplified rocket engine.