One of the things he talks about here, is his terrain. If you've ever seen a screenshot from his game, you'll notice that he has -tons- of hard edges, cliffs and caves. This is pretty hard to do using traditional terrain methods, so it's pretty cool to get the rundown on how he does it. The first thing that he does, is he stores not just one height value per vertex, but four, one for each of the connecting squares. This allows him to create sharp edges and drop each quad up or down. That's probably sufficient for most things, but he goes a step further, and adds in a boolean value if there's a 'cave' or whatnot underneath it, and then he stores another 4 height values. Probably not the paragon of memory efficiency, but it should probably work out just fine. It's also then limited to just one level of caves in the terrain, which while not a big concern, is still something that's there. You could definitely use that as a jumping point for some pretty stellar technology :)
As you can see here, cliffs and sharp drops are pretty characteristic of the game, and they look quite stunning
His ideas regarding creating content are also pretty fantastic, and something far more developers should pay attention to. His idea, in a nutshell, is to create less unique content, and create procedural ways to make it appear more unique. The less you have to do by hand, the faster it becomes to make things.
He also considers the people that -do- have to still make things by hand, (obviously, that will always be something that needs done) and says that they really shouldn't have to work hard to do the things that they do. Right now, the modeling tools industry is completely controlled by Autodesk, so innovation in modeling tools is at an all-time low. Maya and 3DS Max haven't changed all that much over the past decade or two (per Eskil's claims) still retaining the basic fundamentals. Where is the innovation? Where is the change?
His toolset is very much devoted to these ideas, he has an incredible modeling tool that takes advantage of Verse to streamline content management and rapid prototyping, and which also allows him to do things that so many other modeling applications can't. Instead of grids and buttons, he's simplified the user interface down to the mouse and a few small directional pop-up menus. Yet even these small options allow the content creator far more control over their model than any of the major modeling tools do.
Loq Airou is Eskil's own modeling tool, which is an incredibly fantastic tool
Now, while this in particular addresses issues specifically in game development, it's also an issue that should be considered outside of game development as well. With all this technology that we have, how much of it actually increases our productivity..? We need to make sure that as we progress into the future, the things that we create still drive us in a productive direction. Not enabling us to become lazy, but to simplify or abstract menial tasks so that we can direct our minds towards more productive items.