bhaktatejas922

5 ms·

31.
by bhaktatejas922 1y ago
A lot of this stems from friction in brittle process around applying edits or latency from full file rewrites. The bar to beat is time it takes for human intelligence + Cursor Tab to complete a task, but in practice the bar becomes "fe
32.
How Intelligence Self Organizes(tejasbhakta.com)
4 points by bhaktatejas922 1y ago|0 comments
33.
by bhaktatejas922 1y ago
Notably missing fast apply like Morph
34.
by bhaktatejas922 1y ago
mcp is out! https://morphllm.com/mcp
35.
by bhaktatejas922 1y ago
mcp is out! https://morphllm.com/mcp
36.
by bhaktatejas922 1y ago
I have never felt so disconnected from the findings of a study
37.
by bhaktatejas922 1y ago
Yeah, incremental updates are something models struggle with, which is what we're trying to solve.
38.
by bhaktatejas922 1y ago
Im not sure that it's "better". I still use Opus and it's better at coding but needs more steering to be less of a "Yes you're EXACTLY right" every time i suggest a new solution path. Purely anecdotal thou
39.
by bhaktatejas922 1y ago
for companies and startups we raise it considerably. For free tier there's rate limits + a limit of 100 req per month
40.
by bhaktatejas922 1y ago
This is a coding model, not a code merge model. What we do is 1 thing, extremely well - merging code edits into original files
41.
by bhaktatejas922 1y ago
we're quite a bit faster and specifically training for merging code edits. Google diffusion is a swing at a generalist model. Super cool work nonetheless
42.
by bhaktatejas922 1y ago
precisely. we work well on files up to 2k lines. It's hard to wrap your head around at first, but code merge has 100s of edge cases to deal with. Its the perfect application for a model
43.
by bhaktatejas922 1y ago
Thats for 0 data retention - on the Morph website its: 0.80 /1M token input, $1.20 /1M token output. We have discounts for large volumes/reserved instances as well
44.
by bhaktatejas922 1y ago
the claude link is hilarious, hahaha
45.
by bhaktatejas922 1y ago
What do we not complain about if we're being honest?
46.
by bhaktatejas922 1y ago
yep - great post!
47.
by bhaktatejas922 1y ago
if you make an account, we'll email you when we have something easier to try!
48.
by bhaktatejas922 1y ago
Claude Code uses find and replace, errors get piped back to the agent so you rarely feel them unless you hit an infinite error loop
49.
by bhaktatejas922 1y ago
say more! We're in the smart composer plugin, but the integration is out of date
50.
by bhaktatejas922 1y ago
thanks, let us know how it goes
51.
by bhaktatejas922 1y ago
Accuracy of course comes first - errors of any sort break flow state and are massively disruptive. Claude Code does search and replace. The failure rate of search and replace is rather high (over 11%) but claude can self correct them. Fas
52.
by bhaktatejas922 1y ago
if you've used cursor, you've probably felt how seamless fast apply can feel - fast apply is accurate and fast to the point where most don't even realize its a model
53.
by bhaktatejas922 1y ago
took a closer look, the update snippet provided is referencing css thats not defined. this is sort of a overreach example of the "semantic" nature of the edits - ie, if an update explicitly contains syntax, or minor reference err
54.
by bhaktatejas922 1y ago
You should give both a try! key difference is our models are faster and more accurate by a large margin
55.
by bhaktatejas922 1y ago
The shoggoth knows
56.
by bhaktatejas922 1y ago
factual
57.
by bhaktatejas922 1y ago
Its true - Cursor, Cline, and many others still use xml for tool calls. In JSON, the model needs to "focus" on escaping characters correctly while also sampling from a reduced token distribution. https://aider.chat&#x2F
58.
by bhaktatejas922 1y ago
"Creating models that learn developer biases" has a great ring to it - maybe we should make that our mission statement. Thats exactly what we're doing with our models. The Next-Edit completion model especially resonates with
59.
by bhaktatejas922 1y ago
nice catch. the html example was using a hardcoded snippet we forgot to uncomment. fixed
60.
by bhaktatejas922 1y ago
or lets be right faster

More ›