@EventModeling author, contributor to scaling #EventSourcing, CEO @AdaptechGroup, OS: @OmarchyLinux, #linux/#git nerd, YouTube youtube.com/@eventmodeling
Vancouver, British Columbia
- Tweets 121,664
- Following 658
- Followers 7,974
- Likes 143,165
Pinned Tweet
The success that @EventModeling and #eventsourcing enable is incredible. Please feel free to share this with others if they are not convinced that they can improve their projects this way. Thank you to Jason, Fran and others at @ShedSuite for an incredible journey together!
7
35
226
66,259
Adam Dymitruk retweeted
retweeting this because I'm actually flabbergasted at how stupid fast this thing is at voice processing. My custom built home assistant responds SO damn fast now.
We release Whistle: speech to text in a 16.9MB file, running on the CPU in the same engine as Needle. It mostly beats Whisper base with 9x less file size and 6x speed. Whistle supports English, German, French, Spanish, Italian, Dutch and Polish.
cactuscompute.com/blog/whist…
14
44
1,601
76,954
Adam Dymitruk retweeted
Wrong. They produce better results with better specs. That doesn't mean one unit test at a time is the ultimate spec like the tdd mafia wants you to believe.
1
155
Adam Dymitruk retweeted
In this current market of agentic mania, apathy and hubris appear to be the primary competition for most anything involving solid engineering.
Apathy: we don't need no stinkin' rigor, because we can iterate so quickly
Hubris: give me a swam of agents and enough tokens to stand on and I can move the world
27
45
303
12,374
HUAWEI launches the Mate90 Series, equipping all models with flagship τ chips, including Kirin 9050 Pro, the world's first chip with LogicFolding architecture. Watch Huawei Consumer BG CEO He Gang unbox and enjoy the beauty of photography with HUAWEI Mate90 Pro Max and XMAGE Z10!
16
46
263
968,570
Adam Dymitruk retweeted
Google phones have Android desktop mode, Samsung has Dex, and Apple has nothing.
3
7
62
1,922
Adam Dymitruk retweeted
Regularly scheduled reminder that it’s time to dump Windows. Cachy is a great choice if you want excellent gaming performance.
#1 🐧🎮
In the September Steam Hardware Survey, CachyOS is now the most-used Linux distro on Steam, at 18.12% of Linux gamers (+2.65% this month).
Huge thanks to everyone who installed, tested, reported bugs & spread the word. We're just getting started. 💙
59
67
2,007
52,424
Adam Dymitruk retweeted
After nearly 17 years, it has come time for me to move on from @JetBrains. I wish JetBrains and all of my friends and colleagues the best success. Thank you so much for everything. linkedin.com/posts/hadi-hari…
52
16
542
33,826
Adam Dymitruk retweeted
After nearly 7 years at Booking, today is my first day as Head of Infrastructure for Omarchy.
Somehow I got here through personal curiosity that started with Linux in Sarajevo more than twenty years ago.
Entire story below :)
81
42
1,352
46,118
Adam Dymitruk retweeted
6:30 AM. Need to go to sleep, but I was excited because this finally got to the point where it's rendering and not crashing. Gears of War 3 decomp, running native on linux, using rexglue + full custom backend graphics renderer. Yes, I am using AI to assist.
The xenia graphics backend did work, but I wanted as close to a full recomp as I could get and better control and performance over the graphics, so I asked it to port the graphics backend used in the Sonic Unleashed decomp, then extend it, using the xenia backend as a reference.
Finally it goes in game and doesn't crash and mostly renders well enough. Got some performance and rendering bugs to still fix but holy balls its working!
12
5
148
6,445
Adam Dymitruk retweeted
It’s artificial intelligence, Sundar.
And in the history of computing, this puerile attempt to rename it simply because of the simpleminded whims of an incompetent and corrupt narcissist will be seen as a failure of leadership, an abandonment of intellectual integrity, and an embarrassment for any who join with him.
2
3
44
3,744
Adam Dymitruk retweeted
True.
For example, I have a Python script for our pipe parts.
I can generate any variant in seconds, and no one needs a 3D model of a piece of pipe or an elbow.
Everyone at our company knows what it looks like.
Manually designed and maintained 3D models are a waste of time and resources when it comes to creating sheet metal parts.
Just give me the result.
AI takes this even further.
i'm guessing mechanical engineers won't see this one coming because it's difficult for me to wrap my mind around it as a software developer, but the real innovation in CAD is going on here:
classcad.ai
part of thing that makes it hard to reason about is that it's a "headless" CAD engine (that can run in the browser) called ClassCAD built on an obscure kernel called smlib from nvidia (stayed tuned on that)
but guess who doesn't mind headless. guess who likes parametric CADing in code instead of point and click.
1
1
5
501
i don't care how you get there, but the day you secure 2x DGX Spark is the endgame of local ai. you feel frontier intelligence right in front of your desk.
if a single RTX 3090 shows you the way, 2x DGX Spark will enlighten you. secure the boxes and you'll rarely miss closed models like claude or chatgpt.
it's been about 3 months with my 2x DGX Spark and i'm not making assumptions, i've run Qwen 3.8 Flash Next, DeepSeek V4 Flash and now GLM 5.3 Flash on them, and i'm still sitting here surprised by what these two nodes can do.
you can do some crazy shit with these boxes and they sip power, the two GPUs pull about 73W together while GLM decodes.
i'm really happy with these two boxes, and i hope i can add a 3rd one if nvidia shows grace.
loading GLM-5.3 Flash on my 2x DGX Spark, 256GB of unified memory across two boxes, NVIDIA's NVFP4 weights split over the ConnectX cable on stock vLLM
numbers so far:
> 320B total, 18B active, 204GB weights
> decode, MTP off: 15 tok/s, flat to 128K
> decode, MTP on (k=3): 19.5-31.7 tok/s
> code + math, MTP on: 26.7-31.7 tok/s
> 120K context, MTP on: 30.4 tok/s
> prefill, MTP off: 1,350-1,580 tok/s
> 128K prompt: first token in 84s
> thinking on, temp 1.0, live server
working on improving it now, testing quality and all, i've been waiting to try this model, moved home, got my wifi just yesterday and today it's loaded.
98
47
682
100,873
