Hacker News
The Inference Shift
18 points by swolpers ago
|
1 comments
perarneng
|
next
[-]
If it's that much faster then perhaps it can come up with 10 solutions and rank them and still have a total better performance than a larger LLM.
tmuhlestein
|
next
|
previous
[-]
[dead]
moveax3
|
previous
[-]
[dead]