|
|
|
|
|
by xiaopingguo
3728 days ago
|
|
More people should definitely try out sam and 9term as examples of how much code/features you don't really need and can still get real work done. Personally, I wish sam had even less complexity and features than it currently does and thus be more understandable and hackable. Also, never going back to a normal terminal again, "dumb" terminals are awesome. And actually, this experience has lead to a philosophy of software where I think the "dumb"er a piece of software the better and more reliable/performant it is likely to be. So I'm also hugely skeptical of the current Deep AI efforts leading to anywhere good. |
|
In the case of deep learning, all current state-of-the-art models can be expressed as nothing more than matrix multiplication, addition, and clipping. That's all you need. It's so simple: multiply, add, clip. Multiply, add, clip.
But the model isn't transparent. It's clear what each step is doing (multiply! add! clip!), but it's not clear to the human why that's necessary.