Konstantinos Zagoris (@kzagoris) header
Konstantinos Zagoris (@kzagoris) profile picture

Full-stack engineer & founder, Zagor Solutions. I build or rescue software. C#/.NET, TypeScript, PhD. Blog: https://xitter.gq/t.co/0PUN7J1BwZ · Father of 3.

Paphos, Cyprus
Joined October 2013
Photo by Konstantinos Zagoris (@kzagoris)Photo by Konstantinos Zagoris (@kzagoris)Photo by Konstantinos Zagoris (@kzagoris)Photo by Konstantinos Zagoris (@kzagoris)Photo by Konstantinos Zagoris (@kzagoris)Photo by Konstantinos Zagoris (@kzagoris)Photo by Konstantinos Zagoris (@kzagoris)Photo by Konstantinos Zagoris (@kzagoris)Photo by Konstantinos Zagoris (@kzagoris)Photo by Konstantinos Zagoris (@kzagoris)
Konstantinos Zagoris retweeted
Passenger v0.4.0: .NET MAUI and mobile apps. Run, debug and Hot Reload on Android, an iOS simulator or your Mac, with a live XAML preview. Plus screen readers, themes and Vim mode. Still free and MIT. passenger-ide.net/?v=0.4.0
3
10
1
64
32,954
Konstantinos Zagoris retweeted
Starlink saves lives
BREAKING: Greece’s government is using Starlink to support a railway safety system that warns of potential train collisions. Starlink sends live train data to the control center, including through mountainous areas without mobile coverage. • Monitoring equipment installed on 134 locomotives. • GNSS-HEPOS positioning pinpoints trains with accuracy of up to 5 centimeters. • Location updates every second. • A 24/7 control center receives automatic collision-risk alerts so operators can intervene.
938
1,727
125
13,225
1,381,872
Konstantinos Zagoris retweeted
i built gpt-2 in pure assembly. 124m params. zero libc calls. the tokenizer, matmuls, attention, softmax, even exp(): all arm64 asm. ~100 tok/s on a single m1 core. 4 kb of machine code. output matches pytorch token for token.
84
89
16
1,366
74,554
Konstantinos Zagoris retweeted
A lot of people don't realize that Windows has a built-in feature called Windows "Sandbox" that boots up a brand new virtual Windows machine in seconds. So I made an MCP server so you can test your apps inside a throwaway machine in seconds. github.com/shanselman/sandbo…
95
301
23
1,912
89,727
Konstantinos Zagoris retweeted
This work is simple, clear, compelling, but embellished. I first found it shocking, the notion that a model can natively understand neural network weights and productively edit them! That's really not what's happening, & many here are annoyed. BUT, it's still notable & worth understanding. I've done a bit of analysis & auditing, here are some thoughts: 1/ This paper does *not* show an agent can edit its weights ala RSI. It does *not* show a model can edit in new capabilities. It DOES show that a frontier model can remove "bugs" (reasoning loops), using the same ablation technique researchers have used to remove safety refusals (Exp 8 - Qwen). So yes, renewed concern about frontier LLMs getting access to their own weights & editing safety/CBRNE/alignment features. 2/ This paper also does *not* show that an agent can "natively" read, deeply understand, and fix model weights. What is happening is that the model can cleverly deduce the same things that mech-inturp researchers do and fix them in mostly ham-fisted ways. Still cool, but not as groundbreaking. Evidence in thread below...
We asked GPT-6 Astra to edit Qwen3.5-0.8B weights to improve math. It directly changed only 0.016% of weights, fixing 0.8B's reasoning loops & improving math up to 6x and even instruction following 3x, without gradients or training. We call this Weight-Space Programming: frontier LLMs can now natively read/write neural network weights! 🧵 Blog: rikiyat.github.io/blog/weigh…
Readers added context they thought people might want to know
LLM was given weights explicitly as text in prompt and used code/forward passes (per authors' logs) to ablate Qwen looping direction, a known thinking-mode issue per model card. Primarily fixed repetition in that mode, not general 'natively' improving weights/math. rikiyat.github.io/blog/weight-sp… github.com/RikiyaT/lm_und… huggingface.co/Qwen/Qwen3.5-0… huggingface.co/RikiyaT/Qwen3.…
12
50
12
379
34,571
Konstantinos Zagoris retweeted
There's a lot going on in my GPU. I'm making a GPU Taskman to get insights to where my GPU memory is going. PRs welcome
34
47
11
553
32,293
Konstantinos Zagoris retweeted
We asked GPT-6 Astra to edit Qwen3.5-0.8B weights to improve math. It directly changed only 0.016% of weights, fixing 0.8B's reasoning loops & improving math up to 6x and even instruction following 3x, without gradients or training. We call this Weight-Space Programming: frontier LLMs can now natively read/write neural network weights! 🧵 Blog: rikiyat.github.io/blog/weigh…
Readers added context they thought people might want to know
LLM was given weights explicitly as text in prompt and used code/forward passes (per authors' logs) to ablate Qwen looping direction, a known thinking-mode issue per model card. Primarily fixed repetition in that mode, not general 'natively' improving weights/math. rikiyat.github.io/blog/weight-sp… github.com/RikiyaT/lm_und… huggingface.co/Qwen/Qwen3.5-0… huggingface.co/RikiyaT/Qwen3.…
2,368
7,258
5,755
55,917
10,084,941
Konstantinos Zagoris retweeted
📢 New Positions at @SignalGeneriX ! 1️⃣ Firmware/FPGA Engineers 2️⃣ EU R&D Project Managers 3️⃣ Purchasing & Supply Chain Officer 📍 Limassol, Cyprus | Full-time 📩 Apply now: [email protected]
2
16
1
56
167,883
Konstantinos Zagoris retweeted
Introducing terminal-code: VS Code inside the terminal - VS Code compatible CLI - works over ssh - syncs with your terminal theme terminal-code.com
514
616
529
7,619
2,635,926
Konstantinos Zagoris retweeted
Pretty intuitive animation explaining seasons, equinox, solstice, length of day and night. Today marks the summer solstice in the Northern Hemisphere. [🎞️ German Valencia Garcia]
29
736
16
2,812
150,748
Konstantinos Zagoris retweeted
SharpLsp is live. This is how you add a NuGet package
1
5
1
12
1,853
Konstantinos Zagoris retweeted
We've added a catalog of developer productivity tools to our website. Get in contact with us if you'd like to find out more
1
2
1
3
369
Konstantinos Zagoris retweeted
We've shipped more than a thousand versions of Zed, but all of them began with zero. Today, that changes. zed.dev/blog/zed-1-0
285
845
249
8,212
682,381
Konstantinos Zagoris retweeted
Πάσχα Easter Greece 1950's #oldgreekphotos
4
28
1
187
4,420
Konstantinos Zagoris retweeted
A sacrifice must be made. Nominated for the Palme d'Or in 1977, Iphigenia (dir. Michael Cacoyannis) adapts Euripides’ tragedy set before the Trojan War. Out soon in the UK and US/CA, secure your copy now.
19
275
51
2,037
326,163
Konstantinos Zagoris retweeted
You can now run 70B LLMs on a 4GB GPU. AirLLM just made massive models usable on low-memory hardware. 𝗪𝗵𝗮𝘁 𝗷𝘂𝘀𝘁 𝗵𝗮𝗽𝗽𝗲𝗻𝗲𝗱 AirLLM released memory-optimized inference for large language models. It runs 70B models on 4GB VRAM. It can even run 405B Llama 3.1 on 8GB VRAM. 𝗛𝗼𝘄 𝗶𝘁 𝘄𝗼𝗿𝗸𝘀 AirLLM loads models one layer at a time. Instead of loading everything: → Load a layer → Run computation → Free memory → Load the next layer This keeps GPU memory usage extremely low. 𝗞𝗲𝘆 𝗱𝗲𝘁𝗮𝗶𝗹𝘀 • No quantization required by default • Optional 4-bit or 8-bit weight compression • Same API as Hugging Face Transformers • Supports CPU and GPU inference • Works on Linux and macOS Apple Silicon 𝗪𝗵𝗮𝘁 𝘆𝗼𝘂 𝗰𝗮𝗻 𝗱𝗼 • Run Llama, Qwen, Mistral, Mixtral locally • Test large models without cloud GPUs • Prototype agents on cheap hardware
360
1,140
147
10,859
640,394
Konstantinos Zagoris retweeted
Skills.sh is an open ecosystem for finding and sharing agent skills. Add a skill to any agent with: ▲ ~/ npx skills add <owner/repo>
148
418
141
3,277
648,772
Konstantinos Zagoris retweeted
My entire childhood has led me to this moment... I built AgentCraft - orchestrate your agents with your favorite RTS interface! ⚔️ Coming soon 👀
millennial gamers are the best prepared generation for agentic work, they've been training for 25 years
405
369
173
5,028
792,507