[HN Gopher] A Year of Fast Apply - Our Path to 10k Tokens per Se...
___________________________________________________________________
A Year of Fast Apply - Our Path to 10k Tokens per Second
Author : eborgnia
Score : 41 points
Date : 2025-10-29 17:04 UTC (5 hours ago)
(HTM) web link (www.relace.ai)
(TXT) w3m dump (www.relace.ai)
| swyx wrote:
| > To streamline the process while maintaining quality, we built
| our own internal evaluation tool: a Git-style diff viewer with
| annotation tools for categorizing merge outcomes.
|
| vibecoding internal eval tools is the single best use case of ai
| accelerating ai i know of! nice to see
|
| (sorry if this gets asked a lot) - any philsophical/methodology
| differences to MorphLLM that you'd call out since you seem to be
| a direct alternative?
| eborgnia wrote:
| Hey, happy to answer! The manual evals we did showed that both
| morph-v3-fast and morph-v3-large had significantly more
| smoothing and hallucination behaviors.
|
| It's hard to know for sure because their methods aren't public,
| but my guess is the dataset they constructed pushes the Fast
| Apply model to more aggressively fix mistakes introduced by the
| frontier model in the edit snippet.
|
| This aligns with the fact that their flagship model
| (morph-v3-large) is 4x slower than ours -- the
| smoothings/hallucinations are not in the initial code or the
| edit snippet so they break speculative continuations more
| frequently. Their 2x faster model (morph-v3-fast) is likely
| quantized more aggressively (maybe fp4? and run on B200s?)
| because it exhibits very strange behaviors like hallucinating
| invalid characters at random points that make the code non-
| compilable.
|
| From an accuracy POV, auto-smoothing is helpful for fixing
| obvious mistakes in the edit snippet like missed imports from
| well known packages. However, it does increase the frequency of
| code breaking hallucinations like invalid local imports among
| other functional changes that you might not want a small apply
| model to perform.
___________________________________________________________________
(page generated 2025-10-29 23:01 UTC)