I don't want to work for a big company

What makes startups seem great is the survivor bias. The shitty ones with MBA-style management and closed allocation tend to faceplant. The good ones grow and become less good gradually.

Also, many of the big-company perks mentioned in the OP don't exist.

Big companies actually don't provide better benefits than startups. Banks, for just one example, have pretty shitty health insurance due to misguided shareholder activism and cost-cutting. Also, it takes forever to move up in large companies. The upshot of that, though, is that you'll never have a 25-year-old manager-favorite ladyboy hipster douche barking orders at you (which I have seen demolish a startup or two).

Finally, large companies are not great at long-term projects. Those tend to be cut first when there are cuts, and in most companies, continuous cutting is now in favor over press-making layoffs. The best people in large companies (excluding companies with a real research presence, and those aren't paying salaries that would impress anyone on HN) want to work on immediate P&L stuff, not long-term efforts that lack job security and short-term promotability.

There is exactly one reason to work for big companies, and it can be substantial: lateral mobility. Your compensation won't go up fast, but if you can move to more interesting work under a manager the ability to protect, it can be a pretty damn good deal, because with a good manager it doesn't matter that you're in a mediocre company. You'll get to learn a lot (even if you won't have a major impact) and your work will be interesting enough (much more interesting than what most of these startups do) to provide jumping off points into credible side projects and, possibly, startups. It can be a great way to get a 5-year stint (there are plenty of HR Boomer-o-saurs who are still stuck up about "job hoppers") and a lot of time to learn.

The problem is that most companies make it a lot harder to move internally than they advertise. They have ridiculous "headcount" issues and it's almost impossible to move in most firms without a top-20% performance record... but impossible to get a top-20% review unless your manager sees you as having staying power. So you have to play dishonest political games, in most companies, to get the one big-company benefit (lateral mobility) that actually matters.

In fact, most companies are so fucked up on the matter of internal mobility that the only genuine time at which transfer's possible is shortly after a promotion, but managers rarely promote people who seem to have any interest in mobility.

The three main criteria are "the product is software", "great colleagues" and "challenging problems". Small/large is not on the list, but maybe I should have added "having a high impact".

>In fact, most companies are so fucked up on the matter of internal mobility that the only genuine time at which transfer's possible is shortly after a promotion, but managers rarely promote people who seem to have any interest in mobility.

This was the driving reason I left my prior employer. I had no internal mobility without some sort of advocate, despite the large number of open reqs. At the end I got hit by the political double whammy of not having any support for a lateral transfer from my manager[1] and being on the 800lb gorilla project that was trying to draw people in from elsewhere at the time I was trying to leave. The company got away with shit like this, though, because most people thought that was just the way it was at large companies, and small companies were too risky and flaky.

[1] I even got labelled "disloyal" in the heat of the moment for expressing the desire once.

Here's probably more than you care to read on this topic. I just finished it this morning: http://michaelochurch.wordpress.com/2013/04/16/gervais-macle...

The short version, as pertains to what you experience, is that companies have (since Enron made it fashionable) included performance reviews in the transfer packet as a way of enabling managerial extortion (for the sake of project expediency, although it's unclear that it achieves that) and passive firing, reducing lawsuit risk. The passive-firing infrastructure is supposed to be there to make low-performers look objectively unwanted (lawsuit issues) but the headcount limitations and transfer blocks actually end up getting used against high-perfomers that their managers want to keep captive. Google (which has Enron-style performance reviews) has a well-documented problem with this.