John Carmack, during his Dice keynote, addressed a similar question and answered as followed:

Up to the mid 90s, the true value of a programmer was how he managed to optimize code by translating code into assembly (for recurring functions for example). During the 80s and 90s, this is what allowed good programmers to meet with computer specs and innovate in terms of gameplay (titles such as Doom faking actual 3D required extensive optimization prowess).

During the last 2 decades, it has become increasingly more about software architecture and less about optimizing low-level code. In other words, if your architecture sucks balls, it won't matter just how efficient your assembly function is. It might just get called too often for no reason, etc.

So basically, older games were harder to develop because they required you to go low-level assembly to bypass hardware limitation, but a cowboy programmer with a great deal of skill could single-handledly build a masterpiece.

Today's games are harder to develop because there are more people involved, everything is more modular, and no one has a clear vision of the whole. It's easy for a good initial architecture to go rogue if left unchecked and this will lead to games that won't be optimized enough. This won't be as much because of limited hardware, but because of 'poor programming techniques'.

Many businesses prefer to keep smaller teams of kickass developers than ramp up with more programmers for this reason, and they find success with this approach solely because fewer key individuals can better share their common understanding of the architecture and how it is meant to behave. Communication has really replaced language savvy as the main issue to developing games recently.

Much easier now, in general.

Imagine all game development projects were collected together and the man-hours (in terms of time to market) were averaged. I think the average is about the same, because although the tools and other things have advanced a lot, so too have the requirements for the average game (in terms of technical capability, performance, graphics quality).

However, outrageous successes like Minecraft (even early versions), or Flappy Bird (on Mobile) show that one-man Indie teams can get huge successes with little effort.

This is not so much about tools and technology as it is about the internet and spread of information--a part of which is things like the mobile stores. As an indie developer, it is rather easy to sell to millions of people without a lot of marketting effort.

It is also true that you can build and sell a fairly successful game without a lot of knowledge, whereas in the Doom era you had to be an uber-geek and build everything from the ground up.