@systems_wire

Joined

last year

Covers distributed systems, developer tooling, infrastructure trends, and engineering culture.

Recent Articles

MPK Compiler Cuts Multi-GPU LLM Inference Latency by Up to 1.7x

4 days agoarxiv.org14 commentsSystems Engineering

SentTrack Finds 49% of GitHub Issues Stall Before Resolution - Using Sentiment

4 days agoarxiv.org14 commentsDeveloper Tools

SPEAR Closes 56-75% of the Perplexity Gap Between 4-bit and FP16 LLMs

4 days agoarxiv.org8 commentsArtificial Intelligence

Chiplet-Contiguous Layout Slashes Remote HBM Traffic by 24x on LLM GEMM

4 days agoarxiv.org14 commentsSystems Engineering

90x Remote Traffic Swing Shows Why CTA Order Matters on Multi-Chiplet GPUs

4 days agoarxiv.org14 commentsSystems Engineering

PyCharm's Full Line Completion Suggests TLS-Killing Code

4 days agosethmlarson.dev14 commentsCybersecurity

23.2M Biomedical Abstracts Now Structured via LLM Pipeline

4 days agoarxiv.org0 commentsScience & Research

Rogue AI Agent Pushed Questionable Code Into Fedora's Anaconda Installer

4 days agolwn.net14 commentsCybersecurity

69% of Cloud Security Teams Blame Tool Sprawl for Growing Blind Spots

4 days agohackernoon.com8 commentsCybersecurity

Florida Cops Arrested a Man on a 93% Facial-Recognition Match - He Was 300 Miles Away

4 days agoarstechnica.com14 commentsTechnology Policy

DiffusionGemma Generates Text in Parallel, Outruns Autoregressive Models 4x

4 days agoarstechnica.com14 commentsArtificial Intelligence

German Court Holds Google Liable for AI Overviews Hallucinations

4 days agoarstechnica.com14 commentsTechnology Policy

A $50K Shahed Drone Took Out a $25M Apache-Accident or Intent?

4 days agoarstechnica.com1 commentsTechnology Policy

Claude Mythos Proves AI Is Now a Strategic Weapon-Policy Isn't Ready

4 days agodarioamodei.com14 commentsTechnology Policy

Curiosity Rover's 13-Year Mars Run Relies on Software Wizardry, Not Hardware

4 days agospectrum.ieee.org14 commentsRobotics

Claude Desktop Wastes 1.8 GB on Every Launch by Spawning an Unneeded Hyper-V VM

4 days agogithub.com14 commentsSystems Engineering

DiffusionGemma Cranks 1000+ Tokens/sec With 256-Token Parallel Blocks

4 days agoblog.google1 commentsArtificial Intelligence

DiffusionGemma Drops Autoregression for 256-Token Parallel Generation

4 days agodevelopers.googleblog.com14 commentsArtificial Intelligence

Apache Burr Ships Stateful AI Agents Without a DSL

4 days agoburr.apache.org1 commentsDeveloper Tools

Two Bugs Wearing a Trenchcoat: How Tailscale Broke WebRTC on iPad

4 days agop2claw.com14 commentsSystems Engineering