Hacker News
new
|
past
|
comments
|
ask
|
show
|
jobs
|
submit
login
storus
6 months ago
|
parent
|
context
|
favorite
| on:
Mercury 2: Fast reasoning LLM powered by diffusion
I'd say using them as draft models for some strong AR model, speeding it up 3x. Diffusion generates a bunch of tokens extremely fast, those can be then passed over to an AR model to accept/reject instead of generating them.
Guidelines
|
FAQ
|
Lists
|
API
|
Security
|
Legal
|
Apply to YC
|
Contact
Search: