Well it depends -- it's fairly simple math equation. The more paternalistic benefits you have, the more they cost, and the more people take advantage of more than one of them the more you risk hitting the value requivalent to the 3-5% promotion you have given that employee over 2 years.

Just my opinion, but if you were happy to take a 90k job because you live somewhere with a low cost of living for relatively less strenuous engineering work (let's be honest, Basecamp isn't pushing the boundaries of what can be done on the technical side, more on the product side), you're probably fine with waving off the extra 10k or whatever by thinking to yourself they'll pay for schooling if you ever do that (you probably won't).

Also, important to note that their 10% thing is performance linked ("profit sharing") -- so it pays for itself. If sales at the company grow 10% so they increase your pay 10% (and that isn't even necessarily what it is, they didn't clarify what the 10% is), the incentives are aligned and the company makes an outsize gain compared to each individual employee. It's a win-win-win, but most of the wins are on the side of the company, as usual. Compounding effects are much stronger than 10% yearly gain -- and what's even crazier is that you have to keep moving the company forward to keep getting the 10% -- essentially if you keep doing that, it's exponential improvement for the company, and each new person that comes in has to get on the hamster wheel. It's brilliant (which is why it's so common).