I love dashboards metrics, but I hate creating them, I told myself: well I would love an MCP for Grafana, turns out they did it and released v1.0.0 last week! What a time to be alive! #grafana

Qwen 3.6 27B forever outlier of local inference! 🤞for 3.7 or 3.8

Go based GUI framework: practical, immediate mode, flexbox model github.com/hasenj/go-sh... I needed a quick native debug tool UI, I tasked an LLM to build it, it worked perfectly, this framework is created for humans and AI agents, it supports a headless renderer so it can output a screenshot 😍

GitHub - hasenj/go-shirei: Go based GUI framework: practical, immediate mode, flexbox model

Go based GUI framework: practical, immediate mode, flexbox model - hasenj/go-shirei

github.com

#duckdb Quack protocol is awesome, but it is missing some integrations for the enterprise, I've added JWT permissions management through a Go wrapper, give it a shot: github.com/akhenakh/qua...

GitHub - akhenakh/quackjwt: A DuckDB's Quack Remote Protocol server with JWT authentication and table-level authorization

A DuckDB's Quack Remote Protocol server with JWT authentication and table-level authorization - akhenakh/quackjwt

github.com

"Wait — actually" - GLM 5.2

Defer available in gcc and clang, using C.

Defer available in gcc and clang

About a year ago I posted about defer and that it would be available for everyone using gcc and/or clang soon. So it is probably time for an update. Two things have happened in the mean time: A tec…

gustedt.wordpress.com

I've made a Kubernetes that runs in the browser, and today we're open sourcing it. The hope here is to make super interactive and visual web content about Kubernetes much easier to write, and we have a whole bunch planned this year. Repo: github.com/ngrok/webern... Demo: webernetes-demo.ngrok.app

I really like the board from Numworks, board.numworks.com. Numworks develops a great physical calculator for students but also a great ecosystem around it.

Amd64 go compilation env GOAMD64=v4 micro architecture vs performance

How much do amd64 microarchitecture levels help in Go?

Our 64-bit Intel and AMD processors have evolved over decades. When you compile a Go program for a 64-bit Intel or AMD processor, the compiler targets, by default, a nearly 20-year-old instruction set...

lemire.me

Reasoning parameters directly in llamacpp UI, what a time to be alive!

Coding always was the easy part (CEOs got AI wrong). I have two different platforms integrated via their SDK, they can communicate, now I need a place to deploy it, make it redundant, restart automatically, monitoring... and so on.

Frozen datalake with no moving parts

Frozen DuckLakes for Multi-User, Serverless Data Access

We show how you can build high-performance data lakes with no moving parts.

ducklake.select

Experimenting with Cloudflare AI Gateway to use it as an inference provider (Kimi K2.6) so far so good. It's an okay price for a US provider. developers.cloudflare.com/workers-ai/p...

Pricing

Workers AI pricing is based on Neurons, with a free daily allocation and per-model rates.

developers.cloudflare.com

Reporting on llama.cpp MTP branch, on AMD XTX 7900, from 40t/s to 69t/s using Qwen3.6-27B-UD-Q4_K_XL-MTP github.com/ggml-org/lla...

llama + spec: MTP Support by am17an · Pull Request #22673 · ggml-org/llama.cpp

Overview This PR adds support for MTP (Multi Token Prediction) heads. I tested this on Qwen3.6 27B and Qwen3.6 35BA3B but in principle it should work for any MTP model. I've posted the detaile...

github.com

@unsloth.ai Hey Unsloth I think the MTP uploads are incorrect at least the Q4_K_XL, same build worked with havenoammo GGUF

Working on a terminal map renderer, using Kitty Terminal graphics protocol

Qwen did the impossible, I am testing it but Qwen 3.6 27B (dense) seems legit 😮 qwen.ai/blog?id=qwen...

The TinyGo team has announced a new version of its eponymous Go compiler targeting embedded and other resource-constrained environments, promising a wealth of new features — and support for the new Arduino UNO Q single-board computer.

TinyGo 0.41.0 Brings Wireless Support on Espressif's ESP32 Chips and Arduino UNO Q Compatibility

Fancy running your Go code on the Arduino UNO Q's STMicroelectronics STM32 coprocessor? Well, now you can.

hackster.io

Here is my take on a coding LLM analogy: LLM is a tool, it's like a smart shovel, if nobody is holding the shovel, it won't produce a hole. If the human holding the shovel does not know how to dig a hole, the smart shovel won't be able to produce a hole. It's a tool, just a tool, a great tool.

Qwen 3.6 35B-A3B MOE very close to the quality of Qwen 3.5 27B dense! What a time to be alive! Unsloth quants incoming huggingface.co/unsloth/Qwen...

unsloth/Qwen3.6-35B-A3B-GGUF · Hugging Face

We’re on a journey to advance and democratize artificial intelligence through open source and open science.

huggingface.co

Elegant atomic operation solution in #golang iampavel.dev/blog/atomic-...

Atomic Operations in Go • Asaduzzaman Pavel

How I actually use sync/atomic for lock-free programming when a mutex is too heavy, and the weird edges you hit along the way.

iampavel.dev

DaVinci Resolve 21 supports Lottie animation! www.blackmagicdesign.com/products/dav...

Interesting to see Synadia, the company behind NATS, talking about their cloud offer and the introduction of transformers using the Bloblang language. So they went with Bento! warpstreamlabs.github.io/bento/docs/g...

Transformers Tutorial: Process and Transform Data Streams | Beginner Connector Series

YouTube video by Synadia

youtube.com

Kitten ssh (kitty) is underrated, I would like to see it in Ghostty for example. It is part of my daily workflow to open new windows on remote hosts.

Truly convenient SSH

Automatic Shell integration on remote hosts, Easily clone local shell/editor config on remote hosts, Automatic re-use of existing connections to avoid connection setup latency, Make the kitten bina...

sw.kovidgoyal.net

Hot take, since we can't trust our own software developments anymore (because of LLMs), hosting will evolve into untrusted sandboxes like WASM, gVisor.

NATS Provisioner is a lightweight, dependency-free CLI tool for declaratively managing #NATS JetStream resources (Streams, Consumers, KeyValue stores, and ObjectStores).

GitHub - akhenakh/nats-provisioner: Dependency-free CLI tool for declaratively managing NATS JetStream resources (Streams, Consumers, KeyValue stores, and ObjectStores).

Dependency-free CLI tool for declaratively managing NATS JetStream resources (Streams, Consumers, KeyValue stores, and ObjectStores). - akhenakh/nats-provisioner

github.com

I've published the code for my Bento visual tool, it is already useful to me while creating new workflow.

An LSP and syntax validator for SQLite not « sql »

Split view are in Firefox already! (weird they hide it): Open Firefox (version 146.0 or later) Enter about:config in the address bar Search for browser.tabs.splitView.enabled Double-click the ‘false‘ value to set it to ‘true‘

Working on a stream visualizer for #Bento