Klaus Aschenbrenner

Klaus Aschenbrenner

@sqlpassion.at

Senior Database Administrator in the Online-Gambling industry. SQL Server & PostgreSQL specialist during the day. Astrophotographer during the night. Triathlete in between. Loves low-level system programming in Rust.

692 followers146 following17.2K posts

I've burned down a huge amount of my weekly available tokens across multiple LLMs, but I have now for my own x64 OS a working Rust-based Intel NIC driver - for the Intel 82577LM (8086:10ea) and Intel I219-V (8086:15b8) #osdev #rust #rustlang

My x84 OS implements now a fully working RTL8139 network driver. It can ping IP addresses on the internet, and also answers ping requests. #osdev #rust #rustlang

Klaus Aschenbrenner reposted

Some big news: @ducklabs.com is joining AWS, with the move expected to complete by early September. Our team will stay together in Amsterdam, with more resources to take DuckDB, DuckLake and Quack further - while the open-source Duck Stack remains free under MIT license ducklabs.com/news/2026/08...

ducklabs.comDuckLabs to Join AWS, Projects to Remain Open SourceDuckLabs provides services around the DuckDB in-process OLAP data management system directly from its main developers.

The simulator can execute now LOIs (Lunar Orbit Insertions), and also Free-Return Trajectories - that saved the life of Apollo 13. Every single dot that you see, is actually *calculated* by an orbital mechanics physics engine.

Over the last 6-8 hours (spread across 3 days), I’ve vibe-coded a Home Assistant dashboard that displays data from our photovoltaic system, collected via Modbus over the local TCP/IP network. Home Assistent runs on my old Odriod-H2 Mini-PC from the year 2018...

I have now a full-blown working UEFI boot path in my OS - incl. a simple AHCI driver to access the SATA based SSD drive with its FAT32 ESP partition. 100% pure Rust code (not a single asm file!), without any dependencies to other crates! #rust #rustlang #x64 #computerscience

I can boot now my x64 kernel on native UEFI hardware - like my AMD Ryzen system with 128 GB RAM. It uses the provided GOP framebuffer to draw a pseudo console to the screen. The legacy BIOS boot path (on older hardware) supports now also the VBE-based graphics mode for the console. #osdev #rust

I have migrated now my whole x64-based OS entirely to Rust. There is NO single line of C code anymore - just a few pure assembly files! In addition, I'm also providing a VScode based Dev-Container, with which you can even build and run the OS in your browser! #rust #rustlang #osdev #computerscience

Finished the first bicycle race of my life: the „Dolomitenradrundfahrt“ in East-Tyrol. 110km with 1700m ascent! 🥵 4:19h, and the 96th place (out of 153 participants) within my age group! 💪

Klaus Aschenbrenner reposted

Max Schrems ist ein Jurist, der mit seinem Einsatz für Datenschutz- und Grundrechte international höchst erfolgreich war. Peter Steinberger reitet eine Bubble, die uns derzeit alle mit einem ganzen Bündel gefährlicher Probleme konfrontiert. Österreichische Medienpolitik so:

Klaus Aschenbrenner reposted

We released DuckDB v1.5! This release comes with a “friendly CLI” client, a new (opt-in) PEG parser, support for VARIANT types and many lakehouse features. It also ships a new network stack, a reworked geospatial extension, Azure writes and an ODBC scanner. Read more at duckdb.org/2026/03/09/a...

Klaus Aschenbrenner reposted

[VLDB '25] MD-MVCC: Multi-version Concurrency Control for Schema Changes in Azure SQL Database www.vldb.org/pvldb/v... A great discussion of the end-to-end impact of allowing multiple versions of schema metadata information to be live concurrently, in a real, production system.

It's so "funny" when #ChatGPT hallucinates about SQL Server 2025 feature internals, which are not documented anywhere. And then it says "sorry", because it delivered garbage to you... #StrangeNewWorld

CoPilot in SSMS reads from my database/sql server instance, but doesn't show me any executed queries. This is *so* dangerous for your production data - CoPilot could just pass some of your sensitive data to the LLM provider... #ai #llm #datasecurity #gdpr

When my MacBook Air M3 24 GB outperforms my AMD Ryzen system (128 GB RAM) with a RTX 3080Ti GPU for local LLM inference. The Unified Memory architecture rocks! 💪

Klaus Aschenbrenner reposted

Most programmers are taught that L1 is the “top level” cache on x86.
 It’s not quite true anymore! 
Intel calls it the Decoded Stream Buffer (DSB), AMD the OpCache. Only enough room for ~4,000 micro-ops, but there are interesting ways to take advantage of it.

"new consulting pipeline almost shut off starting in December" -> same here! 😪

WTF! 🫣

X2Y@handle.invalid

SAM ALTMAN: “People talk about how much energy it takes to train an AI model … But it also takes a lot of energy to train a human. It takes like 20 years of life and all of the food you eat during that time before you get smart.”

Never, ever trust an LLM *blindly*: Claude Code generated me frontend code that stores a bearer token (used for authentication against the REST API backend) in local storage. That's the perfect use case for XSS attacks... #fail #llm

Nobody needs you anymore…

Brent Ozar@brento.bsky.social

On LinkedIn, I asked tech consultants how their incoming pipeline looks right now, and the results paint a bleak picture. www.linkedin.com/feed/update/...