Kirill Kirillov

@VictorTaelin I do some technical writing every now and then, and I never get to like the output from agents. haven't used any other tools, and probably won't because it's so much easier to prompt the clanker that already has access to whatever it needs

@VictorTaelin the first line my eyes landed on was "3.1 One File, Three Executors" and it's such a tell it was written by AI 😁 also, I've never seen 2 colons used in a sentence. it's my favorite writing tool, but my rhetoric teachers often didn't let me use one, let alone two

I casually noticed Fable formalized IEEE 754 in Bend2 !?? I just asked for float parsers but thanks I guess

@Bobby_Anguelov Kirill said on LinkedIn that he has recently completed one of his contracts. did you commission this work and open source it?

Regarding the new Esoterica Engine update. The renderer has been completely rewritten by Kirill Bazhenov. You can find more information about it here: docs.esotericaengine.com

@aneki @robaganrab @dr__madmax @thdxr skill issue

@kirusfg1 @robaganrab @dr__madmax @thdxr Haha this is not working on wayland

hex (how we do voice prompting) is coming to linux

@robaganrab @dr__madmax @thdxr installed dotool, set paste method to "clipboard (ctrl+shift+v). to trigger recording, I send a specific kill signal to handy that it treats as a recording signal since wayland doesn't support global shortcuts. I found all this here: #linux-notes" target="_blank" rel="nofollow noopener">github.com/cjpais/Handy#l…

@kirusfg1 @dr__madmax @thdxr How did you get it working? I only had issues with hotkeys, sound device, language and type back. Everything else (eg. the UI) is working fine. On Ubuntu, so not even compiling the system and such. Thanks for describing the tricks!

@thepanta82 @thdxr also, do you know for sure that hex is not local?

@kirusfg1 @thdxr handy seems focused on local only, meaning they can't do post-processing, meaning the output makes you sound like an idiot

@thepanta82 @thdxr occasional errors have never been a serious problem when using handy with harnesses. I don't use it for messaging

@dr__madmax @thdxr on wayland? I use it on wayland

@kirusfg1 @thdxr Handy doesn't work on wayanad.

@SebAaltonen any reason why yours is not X3D?

Zen6 seems promising. I will be buying the 24 core desktop version. +50% cores, +50% cache, 12 core CCD for gaming. If IPC/clock improvement is this good, it’s a massive uplift over my 7950X.

@flydotio @NousResearch am I the only one that's pretty sure they're a crypto scam?

Huge congrats to @NousResearch.

@thdxr what's that solution gonna look like? I'm genuinely curious, you even have a dm from me on it from a while back

@redshiftpress ohh yeah we're gonna provide a fix for this too

we landed on a pretty good workflow for doing parallel work in OpenCode this demo is with git worktrees but i also preview an alternative we're working on at the end this will be in 1.6.0

@SebAaltonen total graphics noob here. you had to use a different version of the scene because each object had its own mesh, right? is some sort of mesh deduplication pass practical or even feasible? I understand having sane assets is preferable and possible in games, but what if it's not?

Added auto-instancing to mini-engine renderer. 4x less draw calls in this scene. This is the Bistro scene with interior. Much more objects. GPU timing for G-buffer and shadows seem around 10%-20% faster. Instancing mostly helps CPU. Can even slow down GPU, since object data has non-uniform access path. But on a modern wide GPUs, it's good to have more work per draw. Less bubbles.

@kmdrfx ah, it was familiar somehow. loved that show. thanks

@kirusfg1 Atlas - Camo & Krooked & Christian Kolonovits - Red Bull Symphonic Johann Strauss 2025 Edition youtu.be/AZ7MN4WpZRI?si…

OpenTUI Keymap is a host agnostic key/cmd engine for DOM like apps. It allows extreme customisation of a generic core, from key stroke syntax parsing to to command resolution and dispatch behaviour. It composes layers of bindings and commands into a single adaptive dispatch model. Demo link in the replies. There are many such systems, most of which stop at the key binding part. Apps often are expected to implement their own command layer on top of these. This disconnects the bindings from the actual command registry. Looking for a proper solution usable in OpenCode I couldn't find anything that fulfilled the need of being extremely extensible to allow plugins to fully control key mapping and command behaviour. I wanted it to always be able to know exactly which keys and commands are reachable at any point in time, no matter from where and how the mappings are manipulated. The main driver was enabling a which-key like plugin and vim like bindings in a mostly declarative way. While any other plugin could extend the keymap even further with a custom config syntax for key strokes for example. So addons for it are mostly composable and it comes with a variety of addons providing common behaviour. It can also power other discovery features directly from live state. Help views, command palettes and graph/debug UIs can all ask the engine what is active, reachable, shadowed, pending, or dispatchable instead of rebuilding that knowledge separately. The core is intentionally very much generic: hosts adapt focus, hierarchy, input events, and lifecycle, while addons extend parsing, tokens, sequence patterns, command metadata, resolvers, interceptors, and event matching. The result is a keymap that plugins can compose with rather than work around. This could all be built directly into OpenCode and just be app specific. Working on OpenTUI though I want applications to have an out-of-the-box solution to build keyboard-first apps easily. Well, at least agents can. It is extremely over-engineered. And I love it. I hope some of you will too.

@__tinygrad__ is stuff like this just a side quest or does it contribute to the overall vision? would anyone need a petaflop over usb4 on a mac?

If you have a Thunderbolt or USB4 eGPU and a Mac, today is the day you've been waiting for! Apple finally approved our driver for both AMD and NVIDIA. It's so easy to install now a Qwen could do it, then it can run that Qwen...

@thdxr opencode opentui port when? (please)