gav.ilicio.us

After years of delays, Tesla says the second-generation Roadster will finally be unveiled on October 1. Flying cars! The future envisaged in my youth is finally arriving ...

Oracle has started another round of job cuts, and once again the news arrived as a 6am email from "Oracle leadership". My heart goes out to the affected employees - a 6am email is a very cold way to part ways. I wrote about this pattern back in March in AI Layoffs in Australia and the Contractor Anxiety They Create, and nothing about it has become more humane since.

After a few weeks listening to David Sinclair's podcast, I'm tempted to subscribe to Lifespan magazine. The future of longevity research is looking extraordinary.

A practical Magit guide to Git worktrees—the neat trick for keeping parallel agent sessions isolated without cloning the whole repository.

Might be fun to spin this up!

Web-LLM is a high-performance in-browser LLM inference engine. Running models locally in the browser removes the need to send prompts to a server and opens up offline or privacy-sensitive use cases.

This article introduces ImHex by reverse engineering FEZ's save file format. ImHex has become one of the best hex editors I have found recently, and the pattern matching features are particularly useful for decomposing binary files.

This video really encapsulates my life as a Linux admin at this point in time.

I think I'm going to buy this! A walkable ASCII cyberpunk city rendered in a single HTML file.

This was a really progressive resolution, you can't please everyone - the smartest people I know lean into this technology. I think this will ultimately accelerate Debian and FOSS in general.

Bob might be getting an upgrade ...

binvis.io is a browser-based tool for visualizing binary files. It maps bytes to colors using space-filling curves and entropy visualization, making it easier to spot structure, compressed sections, and encrypted regions without uploading data to a server.

Tailcat is a netcat-style utility that runs over Tailscale's encrypted data plane without needing Tailscale's control plane. It is useful for point-to-point connectivity in environments where the full Tailscale coordination server is not practical.

Nvidia has reportedly agreed to buy Hugging Face, the open-source AI model hub, for $12.9 billion. The acquisition would extend Nvidia's influence from silicon into the platform layer and give it a stronger position in cloud services.

Z.ai is releasing its Ox Alpha reasoning model under an open-source license as GLM-5.3-Flash. The move puts another high-performance open-weight model into circulation as the release cadence from Chinese labs and their partners continues to accelerate.

Same base model, better training: GLM-5.3 turns post-training alone into a claimed 50% coding boost—and an unsettling leap in end-to-end cyber exploitation.

Every IT professional should learn PostgreSQL: the original, and still the best.

The race to the bottom accelerates: US labs are fighting back with smarter optimisation, pushing AI-powered analytics below half a cent per answer.

Look out, Claude! Another multi-trillion-parameter open-weight model from China is landing, this one geared for coding.

A sharp defence of PKM: tools like Obsidian don't create knowledge; they give people a place to do it.

A dizzying vision of agent swarms, self-rewriting workflows and post-human CI/CD. I found it thoroughly confusing, but there may be gold here once it clicks.

Locally hosted Kimi K3 at a blistering 0.5 tokens per second!

A Microsoft Research visualization language that helps AI agents turn compact, human-editable specifications into polished charts across Vega-Lite, ECharts and Chart.js.

A great blog post that mirrors some of my feelings about AI export controls and how they relate to U.S. cryptographic export controls from 25 years ago.

Amazing map of castles located around the world.

A cozy command-line clubhouse for computer people.

A clean, visual guide to local, remote and dynamic SSH port forwarding.

HTOP Explained.