FR
live
archive

All articles

Azure pushes platforms agent-first and isolates every agent execution in a hardware microVM

On September 23, 2026, Mike Hulme, Azure’s GM of product marketing, describes the shift from request-response applications to continuously acting multi-agent systems, and Microsoft’s answer: govern the agent on Foundry, execute it in an isolated Azure Container Apps sandbox. For platform teams, it is a concrete framework for what must change when the agent replaces the request.

Hugging Face rewrites tokenizers in Rust with SIMD and encodes text up to 30× faster

On September 21, 2026, Hugging Face detailed version 1 of its tokenizers library: the same output as v0.23, but encoding 3 to 30 times faster single-threaded on an Apple M4 Max, thanks to SIMD bitstream splitting and a word cache. Teams serving LLMs now have a measured reason to look at the tokenizer — the new bottleneck as models get faster.

Linux 7.4 makes hot-adding memory to VMs and CXL up to 81% faster

A patch series by Yuan Liu (Intel), queued in mm/core.git ahead of the Linux 7.4 merge window, replaces page-by-page zone scans with an optimized contiguity check. Measured: up to 81% less time to hot-add memory to a VM and 75% to hot-remove it, with Samsung also reporting CXL gains.

One URL-encoded character slips attackers past WAFs and onto Oracle PeopleSoft

Google has documented a fresh wave of exploitation of CVE-2026-35273 (CVSS 9.8) by UNC6240, a ShinyHunters-linked actor: the group URL-encodes a single character in the path to bypass WAFs and drop web shells on Oracle PeopleSoft. Patch, disable the EMHub, and hunt for /PSEMHUB/ in every encoded form.

The Linux kernel considers an AGENTS.md to rein in patches generated by AI agents

On September 24, 2026, maintainer Sasha Levin proposed adding an AGENTS.md file to the Linux kernel repository to fix attribution mistakes made by AI agents that generate patches. The proposal, debated on LKML, raises a real question: steer agents with a single file or with purpose-built documentation.

OpenTelemetry and Prometheus finally converge, and the 2026 numbers confirm it

A 2026 survey shows interoperability between OpenTelemetry and Prometheus has improved markedly: the ease-of-use score climbed from 3.1 to 3.6, and the share who find them hard to combine fell from 29% to 10%. For an SRE team still on the fence, now is the time to consolidate on the OTel Collector without abandoning Prometheus.

Type at least two characters.

↑ ↓ navigate ↵ open esc dismiss