Google — The Keyword / AI / Research / DeepMind / Developers
TurboQuant: Redefining AI efficiency with extreme compression
Algorithms & Theory
Tuesday, 24 March 2026
Everything this site carried that day, newest first. Stories are kept for a month, and the widely covered ones for a year.
Wiz Research
Three’s a Crowd: TeamPCP trojanizes LiteLLM in Continuation of Campaign
LiteLLM is the latest victim of TeamPCP’s open-source attack spree. Malicious versions 1.82.7 and 1.82.8 abuse Python’s .pth mechanism for stealthy persistence. The malware exfiltrates cloud credentials, CI/CD secrets, and keys to attacker-controlled domains.
Google — The Keyword / AI / Research / DeepMind / Developers
Mapping the modern world: How S2Vec learns the language of our cities
Algorithms & Theory
Vercel (blog + changelog)
Chat SDK now supports concurrent message handling
Chat SDK concurrency now lets you control what happens when a new message arrives before a previous one finishes processing, with the newoption for the Chat class. Four strategies are available: documentation Read theto get started. Read more drop (default): discards incoming messages queue…
Vercel (blog + changelog)
Chat SDK now supports scheduled Slack messages
Chat SDK now supports scheduled messages on Slack, allowing you to deliver a message at a future time. thread.schedule() postAt Useand pass your message and adate, like: documentation Slack adapter Read theto get started, and learn more about the. Read more
Huntress
ITDR for Google Workspace | Huntress Managed ITDR
Huntress now delivers ITDR for Google Workspace to protect identities against BEC, inbox rule manipulation, and account takeover, all with a 24/7 SOC-led response.
OpenAI News
Helping developers build safer AI experiences for teens
OpenAI releases prompt-based teen safety policies for developers using gpt-oss-safeguard, helping moderate age-specific risks in AI systems.
Language blogs: Rust / Go / TypeScript / Swift / Kotlin / Java / .NET / Ruby / PHP / Elixir / ISO C++
Swift 6.3 Released
Swift is designed to be the language you reach for at every layer of the software stack. Whether you’re building embedded firmware, internet-scale services, or full-featured mobile apps, Swift delivers strong safety guarantees, performance control when you need it, and expressive language features…
OpenAI News
Powering product discovery in ChatGPT
ChatGPT introduces richer, visually immersive shopping powered by the Agentic Commerce Protocol, enabling product discovery, side-by-side comparisons, and merchant integration.
OpenAI News
Update on the OpenAI Foundation
The OpenAI Foundation announces plans to invest at least $1 billion in curing diseases, economic opportunity, AI resilience, and community programs.
Vercel (blog + changelog)
Elastic build machines now available in beta
Elastic build machines are now available in beta for all paid plans, giving teams control over build performance without project-level micromanagement. You can configure elastic builds at the team or project level. Standard Enhanced Turbo Rather than a one-size-fits-all approach, Vercel evaluates…
Databricks / ClickHouse / DuckDB / Supabase / Neon / PlanetScale
Intelligent security at ClickHouse speed: How Cogent Security built an AI-native vulnerability management platform
"The key thing that ClickHouse unlocks for AI is speed. Everything we do needs to be done at machine speed to counter AI-enabled attackers.” Karan Gugle, Founding Engineer
Hugging Face Blog
A New Framework for Evaluating Voice Agents (EVA)
Databricks / ClickHouse / DuckDB / Supabase / Neon / PlanetScale
Building high-performance full-text search for object storage
High-performance full-text search, even on object storage. This deep dive from the ClickHouse engineers explains the new text index design and how it keeps queries fast at scale.
Language blogs: Rust / Go / TypeScript / Swift / Kotlin / Java / .NET / Ruby / PHP / Elixir / ISO C++
Type Construction and Cycle Detection
Go 1.26 simplifies type construction and enhances cycle detection for certain kinds of recursive types.
Databricks / ClickHouse / DuckDB / Supabase / Neon / PlanetScale
Smarter Auto-Scaling for ClickHouse: The Two-Window Approach
How ClickHouse Cloud's two-window recommender and target-tracking CPU algorithm cut scale-down latency from 30 hours to 3 hours while eliminating oscillations and reducing infrastructure costs.