George Georgovassilis writes about information technology architecture
I ve found the vscode copilot error No lowest priority node found to be about a context overflow somewhere in the chain inference engine - LLM - copilot. Some/all of the workarounds you can try when that happens: Of course, that will only work if the inference engine and model support larger windows.
When trying to enable the discrete NVIDIA GPU of the Acer Nitro ANV15-41[ANV], the graphical login screen hangs : the OS works fine, but won t accept keyboard or touchpad input, which makes it unusable. It took me a day of wrong paths taken to realise that X11 picks the wrong input devices. The fix is to Continue reading Acer Nitro V15 login hangs
In a previous post [GM4] we ran gemma4 12B on a 16GB RAM budget; today we ll push the envelop and run deepseek-r1:14b-qwen-distill-q4_K_M on the same budget. While last time we got away by fiddling with the model definition, today we ll have to reconfigure the OS itself (in addition to fiddling with the model definition). While Continue reading Running even larger LLMs on small hardware
Suddenly the Flatpak version of VLC on Ubuntu stopped playing video files directly from a network fileshare (SMB/windows share) mounted via Nautilus. The error is about VLC not being able to play smb:// files. That s annoying because Nautilus is supposed to hide the fact that this is a network share and create then illusion of Continue reading VLC Flatpak won t play videos from network shares
The Nitro V15 supports battery protection via min/max charge levels, but Ubuntu doesn t talk the hardware protocol out of the box. There s a kernel module which enables charge levels. I m a bit in a hurry, so the TLD:
Basic idea Project capabilities This project [DEM] packages OmniParser into a reusable, isolated Docker [DOC] container exposing a standard web server endpoint. It abstracts away complex machine learning dependencies so the tool can be integrated into any local project. It demonstrates an end-to-end use case by wiring this vision engine to a headless Chromium browser Continue reading Visual…
4 years ago I wrote about my Android phone work profile setup [WPS]. As companies become more security conscientious, the (imho useful) topic of work profiles has diminished and employees are just handed a second device, managed 100% by the company. But Android is still an interesting ecosystem which has evolved over time, so I Continue reading My android phone setup 2026 edition
Google released Gemma 4 12B today. I m a huge fan of the Gemma model family, they have improved with each iteration and consistently perform on par with larger models. It didn t run at first because it needs more VRAM that my laptop has, but there s a workaround. Here s a short instruction for how to run Continue reading Running large LLMs on small hardware: Gemma 4 12B on a VRAM-constrained…
Independent Component Analysis (ICA) fascinates me [DEMO]. It s a neat method for isolating signals from a signal mix, as long as multiple recordings of the signal mix from different perspectives are available. An application would be isolating speakers in a conference room with many people talking (over each other), while a set of different microphones Continue reading OPAW: Independent Component…
This paper by Martin Farach-Colton, Andrew Krapivin and William Kuszmaul shook the computer science interwebs last year as it proposed a much faster implementation for hash maps. An important disclaimer: unless you are way more bewandered in data structures than me, you re forgiven to believe that this paper revolutionises mainstream hash map implementations like the Continue reading OPAW: Optimal…