iiinova — where your digital self keeps evolving

Reposted by Angel Munoz

Daniel Saewitz @saewitz.com · 26/09/2026

you think AI is getting better but then it writes this crap

4231

Reposted by Angel Munoz

Jordan Marr @jordanmarr.bsky.social · 26/09/2026

Fable.Lit.Dsl + Fable.Lit.Dsl.Shoelace v2 are out! • Growing lists and one-armed ifs no longer rebuild the parent, so scroll, focus & animations survive • New forKeyed for keyed lists • Text & attributes are bound: escaped, and updated in place github.com/JordanMarr/F... #fsharp

github.com

GitHub - JordanMarr/Fable.Lit.Dsl: Strongly‑typed UI DSLs for Fable.Lit — HTML, Shoelace, and more.

Strongly‑typed UI DSLs for Fable.Lit — HTML, Shoelace, and more. - JordanMarr/Fable.Lit.Dsl

052

Angel Munoz @tunaxor.me · 25/09/2026

I am worried that I'm building a templating engine to describe game maps e.g. map (36 20) { grass all columns (2 1) { meadow { rect (1 1 4 4 dirt) } hills { ... } } }

130

Reposted by Angel Munoz

Guy Buss @eluciusftw.bsky.social · 18/09/2026

We released a new version of SpectreCoff today (0.57.7), catching up with Spectre.Console again, including the newest features and api changes. #fsharp #dotnet Here's the changelog: github.com/EluciusFTW/S...

github.com

SpectreCoff/CHANGELOG.md at main · EluciusFTW/SpectreCoff

A thin, opinionated wrapper around Spectre.Console in F#. - EluciusFTW/SpectreCoff

061

Reposted by Angel Munoz

Maxime @mangelmaxime.bsky.social · 17/09/2026

Introducing Fable.Ripple: fine-grained reactivity for F# and Fable. A write updates the exact DOM nodes reading the signal. No virtual tree, no diff. Four focused packages for signals, DOM, forms, and routing. Try the beta: fable-hub.github.io/Fable.Ripple/ #fsharp

fable-hub.github.io

Fable.Ripple

Focused F# libraries for building reactive Fable applications

1164

Angel Munoz @tunaxor.me · 16/09/2026

I thought It was just the GLM-5.3 family the annoying one but it seems Deep Seek Flash is on the same train Before I can move on my games, I need to have hard design sessions to understand what the gaps are and where I can start tackling features that are share on both

100

Reposted by Angel Munoz

simoncropp.bsky.social @simoncropp.bsky.social · 10/09/2026

Polyfill now supports using c#15 Closed hierarchies on older runtimes github.com/SimonCropp/P... #dotnet

151

Angel Munoz @tunaxor.me · 09/09/2026

New package! FunGoo 0.1.0 is out, a very minimal library that adds some ease-of-use helpers to consume G# 's Goo GUI I ported an old project music player of mine using LibVLCSharp and Goo itself I guess I will be adding some other cool stuff in there #dotnet #fsharp #gsharp #libvlc

220

Reposted by Angel Munoz

Maxime @mangelmaxime.bsky.social · 03/09/2026

Fable 5.14.0, 5.15.0, and 5.16.0 have been released over the past few days! 🚀 These updates feature: - Improved support for Python, Rust, and Beam. - Added support for precompiled libraries in fable-standalone (stay tuned for a surprise coming in the next few days!). #fsharp #fablecompiler

182

Reposted by Angel Munoz

Maxime @mangelmaxime.bsky.social · 04/09/2026

Nacara is back! 🔥 A complete ground-up rewrite running natively on .NET. It's now fully pluggable and lets you turn F# code snippets into interactive playgrounds. 👀 See it live on Thoth.Json docs: thoth-org.github.io/Thoth.Json/ 🚀 Check out Nacara: mangelmaxime.github.io/Nacara/ #fsharp

0101

Angel Munoz @tunaxor.me · 04/09/2026

Today's dev log I need to figure out how to make it less chaotic in order to see it working nicely from the FPS perspective but other than that. I think I got the core mechanics I want (not sped up btw, just very late game)

010

Angel Munoz @tunaxor.me · 04/09/2026

Welp, life moves on What was I doing? Ah yeah, I keep tunning on this thing. It'll take long time, but I feel very comfortable working on this particular game I think it'll be worth it in the long run

000

Angel Munoz @tunaxor.me · 04/09/2026

Mixed feelings but today was my last day at work. It was a weird 5 year journey, from Angular to C# to Kotlin. My interests have changed over the years but I keep loving the trade If you're looking to hire someone familiar with Ai tools, the web and dotnet platforms let me know!

020

Angel Munoz @tunaxor.me · 03/09/2026

This thing is too tempting... I guess I can deviate the rest of the week from my current game dev stuff to make this shippable... I guess I'll be joining the Three.js demos this week github.com/AngelMunoz/M... github.com/AngelMunoz/M... github.com/AngelMunoz/M... #dotnet #fsharp #fablecompiler

github.com

Mibo.Fable/demo/Main.fs at main · AngelMunoz/Mibo.Fable

Contribute to AngelMunoz/Mibo.Fable development by creating an account on GitHub.

140

Angel Munoz @tunaxor.me · 03/09/2026

I am now in the rabbit hole of running Mibo (F#) in a G# project. As usual the F# code is not friendly to consume if not designed with that in mind. Looking at you FSharpFunc, FSharpList and friends it runs just like you'd expect, after all it's a dotnet project #fsharp #gsharp #dotnet

100

Angel Munoz @tunaxor.me · 01/09/2026

Niche #dotnet #fsharp #gsharp devs unite! We might get 24 users between the two languages!

172

Angel Munoz @tunaxor.me · 01/09/2026

Looks like I got sniped into working on a #fablecompiler version of Mibo.Core (headless runner mainly) this would enable adding backends such as threejs/babylonjs/phaser and others. Run the simulation headless from your F#->JS code and let the guys that know how to paint, paint

000

Angel Munoz @tunaxor.me · 01/09/2026

I was planning to record the next wave, but things went from zero to "oh god" #dotnet #monogame #fsharp

151

Angel Munoz @tunaxor.me · 01/09/2026

Would I be this productive if I switched to godot/unity? I'm not sure I feel quite happy with my own stuff (maybe because its mine?) I should eventually give it a shot though (headless mibo for the simulation though 😈)

100

Reposted by Angel Munoz

Kenney @kenney.nl · 31/08/2026

The updated City Kit: Industrial is now available! Adds solar/wind energy, water tower and much more. These game assets are free to use in any of your projects 🏭 kenney.nl/assets/city-...

117824

Angel Munoz @tunaxor.me · 31/08/2026

Kind of nervous it seems to work. Hopefully live editing will be a nice experience but surely this will be a long piece of work 😅

000

Angel Munoz @tunaxor.me · 31/08/2026

Before diving full into the map editor stuff (which I expect to make a userland feature) I wanted to give at least some jump in flow. As usual all you see is WIP and will indeed change #dotnet #fsharp #monogame

141

Angel Munoz @tunaxor.me · 30/08/2026

I'd say "touching grass" but it's a desert 🏜️ I'm not touching a cactus that's for sure

120

Angel Munoz @tunaxor.me · 30/08/2026

I've come to the conclusion now that I want to build something more complex (map editor, live edition experience in-game) That there's a point where the clanker is just unable to grasp how the code works. I guess this is where vibe-coding dies and the real dev process take on

110

Reposted by Angel Munoz

Cory LaViska @cory.laviska.com · 28/08/2026

“There is nothing preventing a GUI from being fully keyboard-driven4 just like — or even better than — a TUI.” Correct. Both can and should be keyboard accessible. ckardaris.com/blog/2026/08...

ckardaris.com

GUIs should be fully keyboard-driven

GUI applications should be fully driven by the keyboard using shortcuts

0316

Reposted by Angel Munoz

Unsloth AI @unsloth.ai · 28/08/2026

GLM-5.3 can now be run locally! The 2-bit model retains ~81% accuracy after we shrunk it from 1.51TB to 239GB (-83% size). Run on a 256GB Mac or RAM/VRAM setups. GLM-5.3 is the strongest open model to date. Guide: unsloth.ai/docs/models/... GGUF: huggingface.co/unsloth/GLM-...

0454

Reposted by Angel Munoz

Sung Kim @sungkim.bsky.social · 28/08/2026

GLM-5.3 is now open-weight. Their most capable model for agentic coding and cyber defense is now available to download, run, and customize. Weights: huggingface.co/zai-org/GLM-5.3 Tech blog: z.ai/blog/glm-5.3

huggingface.co

zai-org/GLM-5.3 · Hugging Face

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

113315

Reposted by Angel Munoz

Kenney @kenney.nl · 28/08/2026

Monday's update will add windmills, tanks, solar panels, shipping containers and a water tank to City Kit: Industrial! Will still be completely free, no premium or extended options. All free.

419021

Angel Munoz @tunaxor.me · 27/08/2026

re-working a bit the terrain generation Adding slopes, terrain variations and stuff like that decorations, buildings and stuff like that should be coming soon as well. These additions I've been doing have been also added to the source code I've added in tangled

010

Angel Munoz @tunaxor.me · 27/08/2026

My zcode usage over the last 30 days (No, I'm not trying to tokenmaxx to the contrary I try to stay slim) It has been hard to remove the deep seek flash footprint😅 I guess if I will spend any money this end of the year will be in zai anual plan. a bit expensive IMO but they deliver

000

Reposted by Angel Munoz

Unsloth AI @unsloth.ai · 26/08/2026

Qwen3.8-Flash can now be run locally! 🔥 The 125B MoE model outperforms Claude-Opus-4.6 (Max). Run on 75GB RAM via Unsloth GGUFs. Qwen3.8-Flash-Next enables CPU RAM / unified mem setups to deliver near VRAM speeds. Guide: unsloth.ai/docs/models/... GGUF: huggingface.co/unsloth/Qwen...

0274

Reposted by Angel Munoz

Unsloth AI @unsloth.ai · 26/08/2026

GLM-5.3-Flash, also known as ox-alpha, is out now! Zai's GLM-5.3-Flash is a new 320B parameter multimodal open model with 18B active parameters. GLM-5.3-Flash approaches Claude 4.8 Opus on coding and agentic benchmarks. GGUF coming soon!

111914

Reposted by Angel Munoz

mr. TIM @timkellogg.me · 26/08/2026

it’s official: GLM-5.3-Flash is Ox Alpha! z.ai/blog/glm-5.3...

Line graph titled "Agentic Coding Performance by Effort Level" plotting accuracy percentage against average output tokens per task for five AI models on Z.ai Code Bench v1.0:
* Claude Fable 5: Highest performance overall, scaling from ~28.8% accuracy at 33K tokens (Low effort) to ~39.4% accuracy at 114K tokens (Max effort).
* GLM-5.3: Second highest performance, rising sharply from ~24.6% (Low effort) to ~31.3% (High effort) near 49K tokens, peaking at ~34.5% accuracy at 74K tokens (Max effort).
* GLM-5.3-Flash: Scales from ~21.5% accuracy at 41K tokens (Low effort) to ~29.0% accuracy at 139K tokens (Max effort).
* Claude Opus 4.8: Scales from ~21.2% accuracy at 28K tokens (Low effort) to ~29.5% accuracy at 120K tokens (Max effort).
* GLM-5.2: Lowest overall performance, moving from ~19.4% accuracy (Non-Thinking effort at 42K tokens) to ~23.4% accuracy at 94K tokens (Max effort).

711712

Angel Munoz @tunaxor.me · 26/08/2026

The character progress and skill tree system is something I really didn't want to touch but ultimately, I have to do 😔 I'm not so interested in this part so I bet this will bite me in the future. At least I'm sure I have a clear idea of what I want to do with it

000

Angel Munoz @tunaxor.me · 25/08/2026

Wiring things up Placeholder character selection gui + initial build (you can see how the HP changes based on the stats selected) Those mages are still a menace though...

000

Angel Munoz @tunaxor.me · 25/08/2026

If you previously hesitated to add signature files to your #fsharp projects, I'd say make your clanker (if you're using one) do it! It helps them understand the public API surface and thankfully they stop slapping private/internal everywhere inline becomes easier to grok for them (when and what)

260

Reposted by Angel Munoz

Bentok @bencope.land · 17/08/2026

Here’s a a post I did relating to #fsharp, arguing that languages with algebraic type systems capable of rich domain modeling can guard business rules in code generation with compiler enforcement in a way that also reads like natural language. blog.bencope.land/where-the-co...

171

Reposted by Angel Munoz

Anton Moldovan @antyadev.bsky.social · 17/08/2026

🚀 NBomber 6.6.0 is out nbomber.com/blog/2026/08... - We reworked how NBomber calculates the scenario-level stats. - We added the calculation of the data transfer throughput. - We improved HTML report with a new data transfer chart. #dotnet #loadtesting #devtools #csharp #fsharp #testing #QA

031

Reposted by Angel Munoz

Jordan Marr @jordanmarr.bsky.social · 23/08/2026

Serde.FS v1.0.0-beta.6 🏎️ The full serialization + RPC pipeline now runs under Native AOT / trimming: every codec is emitted statically at compile time. No reflection, zero trim warnings. Plus: DI overloads for MapRpcApi. github.com/serde-fs/Ser... #fsharp

github.com

Release v1.0.0-beta.6 · serde-fs/Serde.FS

v1.0.0-beta.6 — Full Native AOT & trimming support 🏎️ Serde.FS now runs its entire serialization + RPC pipeline without reflection — publish with PublishAot=true or PublishTrimmed=true and everythi...

071