Vulkan

Vulkan is the OpenGL replacement that allows multicore CPU usage. I think this is pretty cool. Basically, before we could do something like this
glBegin(GL_TRIANGLES);
glVertex3f(0, 0, 0);
glVertex3f(1, 0, 0);
glVertex3f(0, 1, 0);
glEnd();

But with Vulkan, we can do this in parallel. Basically using following code:
vkBegin(VK_TRIANGLES);

....
thread1: vlVertex3f(0, 0, 0);
thread2: vkVertex3f(1, 0, 0);
thread3: vkVertex3f(0, 1, 0);
....

vkEnd();

So what do you think? Personally I think 2015 will be great for Linux gaming.

Vulkan is an iterated version of AMD's Mantle with the identifier names changed, the shader backend updated to a more recent version of Clang/LLVM using SPIR-V and with GLSL front-end support. The API spec for Mantle was pretty much created by Johan Andersson from DICE, and then it was shopped around until AMD noticed it and realized it would give them a competitive advantage. When Mantle was popularized, Apple also saw the advantage of it as a concept and created Metal. Then Valve, Unreal Engine, and Unity developers also got involved in refining the API, and glNext as an idea was born and through Khronos, they convinced the hardware vendors to adopt it.

http://www.pcworld.com/article/2894036/mantle-is-a-vulkan-amds-dead-graphics-api-rises-from-the-ashes-as-opengls-successor.html

And, it's not about parallelism per se, it's about removing driver abstraction overhead and giving you closer to the metal access to the underlying GPU hardware. The driver abstraction used by D3D11 and GL 4.x and earlier was getting in the way scalability, and of being able to generate draw commands directly on the GPU.

With D3D12 and Vulkan, it's up to the end developer to manage GPU memory and synchronization barriers between the CPU and GPU. The API no longer holds your hand. It's more difficult to program, but it's no longer a driver nightmare where you have to guess what hacks the drivers are doing, and the driver writers no longer have to guess and work around stupid things game developers are doing.

There aren't really any user-facing features that the gamers and common plebs are going to notice, except perhaps faster frame rates and more reliable/stable drivers across different platforms.

The real problem with the games industry right now (besides the infestation of SJWs and Marxists) are the content tools and the expectations of artists/modelers. The artists are used to just pushing polygon meshes and static textures created offline with the usual tools: 3DS Max, Maya, Photoshop, etc. They're averse to the idea of procedurally generated content. They don't know how to program or how to compose complex systems. They just want GUI sliders, radio buttons, and check boxes, and this interface is really limiting the type of content that could be possible.

That's why you sometimes see more revolutionary stuff in the indie scene, because the programmers are often also the artists, and aren't afraid to create particle systems and materials completely in shader code or do procedural content generation.

Now if someone could figure out AI to replace artists/modelers..

I'm the resident game/graphics programmer and I have access to DX12, Mantle, and we should be getting early access to Vulkan.