12 votes

Introducing WebGPU support for llama.cpp

2 comments

  1. glesica
    Link
    This is pretty neat! I'm hopeful that we're moving toward a world where local models are the norm and remote models are mostly used for large, complex, or specialized tasks. I don't know how...

    This is pretty neat! I'm hopeful that we're moving toward a world where local models are the norm and remote models are mostly used for large, complex, or specialized tasks. I don't know how realistic that is, time will tell.

    5 votes
  2. cqns
    Link
    Nothing good about this. Just makes browser bloat even more of an issue, especially with folks - such as myself - that keep a dozen or so tabs on the wayside just in case. So glad I don't update...

    Nothing good about this. Just makes browser bloat even more of an issue, especially with folks - such as myself - that keep a dozen or so tabs on the wayside just in case. So glad I don't update anything. I don't feel the need to.