Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Launch HN: Morph (YC S23) – Apply AI code edits at 4,500 tokens/sec

news.ycombinator.com

Post reply on HN
Print view View original

41–50 of 196 posts

Re: Launch HN: Morph (YC S23) – Apply AI code edits at 4,500 tokens/sec

#41

Post by bhaktatejas922 »

treyd wrote: post #33
bhaktatejas922 wrote: post #10

Earlier quoted context omitted.

you can - but they dont work reliably in practice. Common issues include search match fails, missing commas in replaced items (model doesnt have surround context while replacing), and a few other error cases. This issues are much worse for scattered edits across a file from real world queries (ex: make this page look nicer). Patches tend to work fine for single line or extremely focused edits though - Cursor uses s&r…

I wonder if it'd be feasible to have a much smaller model that could go in and correct these meshing issues that require simpler reasoning?

hm maybe but correction/issue detection is a much harder task for models. If you pipe back the errors in it could work, but personally still see Fast Apply as the better approach

Re: Launch HN: Morph (YC S23) – Apply AI code edits at 4,500 tokens/sec

#43

Post by bhaktatejas922 »

seanw265 wrote: post #40

Last time I looked into Morph, I noticed you weren’t yet on OpenRouter. I see that’s changed, but it looks like only an older model is listed. Any plans to be more active there? Also, are there any benchmarks comparing your fast apply models to others like Relace or even Llama via Cerebras? I’m particularly interested in output accuracy.

the v2 model listed currently points to morph-v3-large. We're working with them to get v3-large and v3-fast listed

Re: Launch HN: Morph (YC S23) – Apply AI code edits at 4,500 tokens/sec

#45

Post by bhaktatejas922 »

ashwindharne wrote: post #31
deepdarkforest wrote: post #9

> 1) Raw inference speed matters more than incremental accuracy gains for dev UX—agree or disagree? I know you are trying to generate some controversy/visibility, but i think if we are being transparent here, you know this is wrong. People prefer using larger (or reasoning) models, with much bigger diff in tok/sec just for quality in coding, it comes first. Even if i have a big edit to apply, like 5k tokens, 200-300m…

I do find that having inference happen ~50% faster is much more valuable to my workflow than a single digit accuracy increase. If I'm going to have to check that the changes are correct anyways, getting more iterations in faster feels much better than incremental accuracy. There's definitely a tipping point though. If the accuracy gains are so high that I can check its work less carefully or less often, the benefits…

exactly. The point is that none of the users even realize a model is doing the apply - it should be so accurate and fast that it feels like its not there

Re: Launch HN: Morph (YC S23) – Apply AI code edits at 4,500 tokens/sec

#46

Post by bhaktatejas922 »

Workaccount2 wrote: post #44

Just for clarification here because I am a bit confused, Morph is a tool for integrating the output of other LLMs and not an LLM itself? It doesn't generate 4500 tok/sec, it can edit 4500 tok/sec?

Correct, but morph is a LLM as well. In practice its basically Big LLM using small LLM as a tool call

Re: Launch HN: Morph (YC S23) – Apply AI code edits at 4,500 tokens/sec

#47

Post by smrtinsert »

deepdarkforest wrote: post #9

> 1) Raw inference speed matters more than incremental accuracy gains for dev UX—agree or disagree? I know you are trying to generate some controversy/visibility, but i think if we are being transparent here, you know this is wrong. People prefer using larger (or reasoning) models, with much bigger diff in tok/sec just for quality in coding, it comes first. Even if i have a big edit to apply, like 5k tokens, 200-300m…

Slow is smooth and smooth is fast.

Re: Launch HN: Morph (YC S23) – Apply AI code edits at 4,500 tokens/sec

#48

Post by bhaktatejas922 »

smrtinsert wrote: post #47
deepdarkforest wrote: post #9

> 1) Raw inference speed matters more than incremental accuracy gains for dev UX—agree or disagree? I know you are trying to generate some controversy/visibility, but i think if we are being transparent here, you know this is wrong. People prefer using larger (or reasoning) models, with much bigger diff in tok/sec just for quality in coding, it comes first. Even if i have a big edit to apply, like 5k tokens, 200-300m…

Slow is smooth and smooth is fast.

and speculative edits is faster

Re: Launch HN: Morph (YC S23) – Apply AI code edits at 4,500 tokens/sec

#50

Post by klank »

candiddevmike wrote: post #24
bhaktatejas922 wrote: post #23

Earlier quoted context omitted.

how so? Is your view that flow state at all isnt a thing, or just with using LLMs?

Flow state is 100% a thing, it's just impossible with LLMs (at least, for me). I can't be blocked waiting on things during a flow state or my mind starts wondering to other places.

Time really is a flat circle. My software career started with me archaically flipping characters in a file I vaguely understood with long pauses waiting on magic compilers to give me my actual output.

Now it's dying in the same place. Thankfully I got to spend the brunt of my career working through the fun, intermediate years.

Post reply on HN