Dmitriy Kovalenko (@0xfflock) header
Dmitriy Kovalenko (@0xfflock) profile picture

Goose, James Goose. vi/vim Made fff, @fframes_rust, and other open source software you might already be using. Used to be @neogoose_btw

SF
Joined July 2018
Photo by Dmitriy Kovalenko (@0xfflock)Photo by Dmitriy Kovalenko (@0xfflock)Photo by Dmitriy Kovalenko (@0xfflock)Photo by Dmitriy Kovalenko (@0xfflock)Photo by Dmitriy Kovalenko (@0xfflock)Photo by Dmitriy Kovalenko (@0xfflock)Photo by Dmitriy Kovalenko (@0xfflock)Photo by Dmitriy Kovalenko (@0xfflock)
Dmitriy Kovalenko retweeted
Most of the best videos are not one-shot, by any means. Script -> storyboard -> lookdev -> iteration -> and pick the right tool for motion graphics, captioning, 3d rendering, etc. Fframes is 10x faster than BlenderMCP for rendering 3d models. That's a huge unlock.
I see a lot of confusion about the purpose of fframes. fframes is not a framework for one-shotting stunning videos, no. It is a framework to create and iterate on the videos using code aiming to make the AI iteration and *your* iteration faster and more pleasant.
1
1
11
912
Dmitriy Kovalenko retweeted
Replying to @neogoose_btw
Thank god for Fframes to let me build the video on time!
1
1
1
635
Dmitriy Kovalenko retweeted
Day 25/ We are doubling all usage limits for all Pro 100/200/500 tiers and Business Premium, effective today. We heard your kind feedback and won’t be reducing usage limits for Pro 200, so your Work and Codex usage will stay the same. We’re also going to continue offering extended limits for dots for now, and usage of your dot will still continue to be separate from your default usage. And we'll of course also keep shipping better models and other efficiency improvements that increase your effective usage over time.
7,387
2,701
3,648
39,159
7,553,201
Dmitriy Kovalenko retweeted
Lots of people should pay attention to what Dmitriy is building for agentic video
I've been doing some optimization to make fframes project spend less time on recompilation. In the next 1.3.1 recompilation time will go from 4 seconds to 0.69 seconds for trivial changes But we now also support blaze which brings compilation down to 69ms for the same change! if you have bazel installed fframes skill will automatically use it
1
1
5
869
Dmitriy Kovalenko retweeted
GPT-6.1 Terra: delivering fast and capable computer use at a lower price. GPT-6.1 Terra understands what’s on screen, takes action across apps, writes naturally, and is a more capable model for your whole workday with the efficiency to take on more of it.
🤖 Made with AI
2,162
2,551
2,328
28,757
8,106,871
Dmitriy Kovalenko retweeted
gosh I hate script based installers. hopefully bashka will help with that
I am tired of executing random scripts to install software and then not being able to uninstall it. Today I fixed it. Introducing bashka! It is a lightweight single-binary cli to manage bash madness. It statically analyzes the bash script that you are going to run for installations and then also keeps track of what you have installed
1
14
5,542
Dmitriy Kovalenko retweeted
This is not the flex you think it is.
In the next version of Claude Code Claude Code starts 20% faster
24
36
5
1,332
90,351
Dmitriy Kovalenko retweeted
Next release of Autolith (releasing very soon!) uses fff 0.10.3!
Had to wait 9 hours for CI but fff 0.10.3 is released and it’s fixed a lot of bugs and made fff more efficient in big directories plz update
1
20
1,609
Dmitriy Kovalenko retweeted
Autolith will bundle fff in the next update
I'm glad everyone sees that ripgrep's performance is a real bottleneck for coding agentic harnesses. It's awesome that OpenAI is working on Codex optimizations. But honestly ... it's not enough. Codex is perfectly capable of using a better file walker than ripgrep and skipping the need to parse those giant stdout blobs every time. Caching just treats the symptom -- it doesn't fix the actual problem.
1
1
1
16
2,155
Dmitriy Kovalenko retweeted
Replying to @neogoose_btw
you deserve it. FFF is a game changer, Freaking Fucking Fabulous
1
1
4
991
Dmitriy Kovalenko retweeted
This search truly rocks in @opencode
in the next version of @opencode, file search is powered by fff - files the agent opens start ranking higher - tool calls reuse the same search layer instead of starting cold - less wasted context - fast af here’s me searching around the linux repo, a ~40m line codebase
1
2
18
2,740
Dmitriy Kovalenko retweeted
I have a resume now (built with @webtui btw) need job pls, thx ironclad[.]sh/resume
I just got hit by Engineering layoffs My philosophy for frontend development is NOT tokenmaxxing to ship as much as possible, but rather beauty in simplicity while keeping complexity and chaos to a minimum. Who's hiring? --- I've worked with React Native for three years I'm very good at CSS and have a good eye for intuitive designs (creator of @webtui, btw) I can adapt to any web framework very quickly I'm very good at maintaining a clean and organized commit history as well as working with others in a VCS like git I'm not afraid to RTFM (read the friendly manual) and I'm good at writing documentation And lastly, I write most of my code by hand because I truly care about the software I'm building and how people will use it
14
11
2
135
9,914
Dmitriy Kovalenko retweeted
testing human+agent file search with fff from @neogoose_btw and it’s insanely fast 💨
your coding agent, anywhere.
1
1
6
3,078
Dmitriy Kovalenko retweeted
It is unbelievable how much better all software has become in the past 6 months! Like everything is working better! Fewer bugs. New features all the time! The apps I use every day have improved at least 10x from where they were just 6 months ago! Services I depend on never break. Oh wait. None of that happened. Almost everything is actually significantly worse. Not sure how to make sense of that given all the immense productivity gains. Actually I do. Regression to the mean has more gravity than ever. And that is not a good thing.
57
136
8
1,595
81,302
Dmitriy Kovalenko retweeted
Replying to @athasdev
8
13
1
208
14,342
Dmitriy Kovalenko retweeted
Fun fact: Lack of fidelity in Celsius has made certain cars more expensive. The F30 used dedicated seven-segment displays for temperature. Because half a degree of C is perceptible, the displays needed an extra digit after the decimal which impacted the BOM
Today I realized that there is an actual benefit of Fahrenheit - you can store temperature as u8 instead of f32
11
5
1
157
20,203
Dmitriy Kovalenko retweeted
This is codex right now xitter.cf/OGALANGLEY/status/2054…
Replying to @nicdunz
grep too apparently
1
1
710
Dmitriy Kovalenko retweeted
Replying to @neogoose_btw
@neogoose_btw's fff solves this btw
A good argument against “just use bash and grep for everything”: Grep doesn’t read your gitignore files and searches through unrelated garbage that forces you to truncate the search results.
1
1
11
2,148