/dev/null
Sign in

Angel D. Munoz

@AngelMunoz

One of the 12 F# devs in the world. Sometimes here, sometimes there, always distracted

commits
92
following
29
followers
20

Commit activity

92 commits in the last 30 days

longest streak 16 days

1,847 / 2,500 XP653 XP to root

Pinned

Fri, Nov 6

Angel D. Munoz@AngelMunoz
Ugh, took some time but I had to make some massive refactors to keeep my agents in shape and not just bloat the app. DevNull.Client a Windows/Linux Desktop app has a new release! https://github.com/AngelMunoz/DevNull.Client/releases/tag/v0.10.0 Both scoop and flatpak should…
Attached image
GitHubRelease v0.10.0 · AngelMunoz/DevNull.ClientAdded Remotes: an approved commit can be shared into a direct remote from its card — pick a developer by name or handle (anyone out of reach says why and stays unpickable), see the commit you are ...

Tue, Nov 3

Angel D. Munoz@AngelMunoz
So, following up from this. I managed to run very simple (perhaps naive) evals from all of the models I use GLM-5.3 High, Max vs GLM-5.3-Flash High,Max, vs DeepSeek-Flash Max For this use case: GLM-5.3 high is the rules deriver (i.e. checks the constraints generates the…
QuotingAMAngel D. Munoz@AngelMunoz
The folks at /dev/null ship really fast, the Avalonia-F# Desktop app I made gets left behind with me in the loop. I'm exploring some codealta workflows to make refactors and feature applications loops and incremental. I know I should have looked into "loops" a while ago but…
Attached image
Show thread
Angel D. Munoz@AngelMunoz
The folks at /dev/null ship really fast, the Avalonia-F# Desktop app I made gets left behind with me in the loop. I'm exploring some codealta workflows to make refactors and feature applications loops and incremental. I know I should have looked into "loops" a while ago but…
Attached image
AM
Angel D. Munozquoted
So, following up from this. I managed to run very simple (perhaps naive) evals from all of the models I use GLM-5.3 High, Max vs GLM-5.3-Flash High,Max, vs DeepSeek-Flash Max For this use case: GLM-5.3 high is the rules deriver (i.e. checks the constraints generates the…
Show thread

Mon, Nov 2

Sun, Nov 1

Angel D. Munoz@AngelMunoz
Hi from /dev/null flatpak :) I just need to make sure to configure CI and the github page correctly and I think it should be possible to have automatic updates on this as well, for windows the scoop integration is done (requires a new release) So exciting! so close yet so far!
Attached image 1 of 2Attached image 2 of 2

Sat, Oct 31

Fri, Oct 30

Angel D. Munoz@AngelMunoz
So after the mini roadmap I have, I'll be looking to publish the desktop client I'm working via scoop in windows and flatpak for linux. Since the binaries I'm producing are AoT dotnet apps I do not expect distribution issues.

Thu, Oct 29

Angel D. Munoz@AngelMunoz
Sorry it took me some time :) If you're looking for a desktop client or /dev/null Here's one I worked on for fun, feel free to give it a shot if you have a chance! This particular release brings remotes alive and fixes a json serialization decoding issue that made the previous…
GitHubRelease v0.4.0 · AngelMunoz/DevNull.ClientAdded Client: the app now keeps a log that follows its flows. Every use case — sign-in, session restore, the feeds and threads, posting, the chat's polls and pushes, the image cache, the local fil...
Show thread

Wed, Oct 28

Angel D. Munoz@AngelMunoz
The repo is ~31K lines of F# across Core, UI, and two heads. Let me check what logging exists today, then fan out exploration.
Oh little glm-5.3-flash you have a good task ahead be patient and hopefully you'll get done in a couple of hours 😂
Show thread
Angel D. Munoz@AngelMunoz
Ugh, it's taking me a while to release the desktop client I'm working on due me rushing to add remotes in here, so if you were testing the desktop app for some reason and it's not working, I'll try to make another release later today with remotes included!

Tue, Oct 27

Mon, Oct 26

Angel D. Munoz@AngelMunoz
Any tools for agentic CAD? asking for a friend 😅
R
Ricardo
Free or service? There is a lot with the advent of physical AI. I can help with that if you want. Drop me a message. I've built some things that are not public to make custom parts for robots. can share my research (not the solution due to NDA, sorry)

Sun, Oct 25

Sat, Oct 24

Angel D. Munoz@AngelMunoz
Ugh, finally... Pasting "high" res images in in the composer brought this up to 400MB in Linux (after the supposed fix). Juggling between GLM-5.3 and DeepSeek Flash we've finally managed to get it back under 200MB in linux, time to touch some grass 😂
Attached image 1 of 2Attached image 2 of 2
QuotingAMAngel D. Munoz@AngelMunoz
pasting images open lots of layers to keep native memory around, I thought what I did earlier today fixed it, then I log into linux and "Hell naw" moment hits
Show thread
Angel D. Munoz@AngelMunoz
pasting images open lots of layers to keep native memory around, I thought what I did earlier today fixed it, then I log into linux and "Hell naw" moment hits