Urgent.News

What's breaking now, across thousands of outlets.

Tuesday, 29 September 2026 Reading Room

Tech

Meta Releases OpenZL 0.3 With Decompression Speed 144% Faster Than Zstd

Just shy of one year ago Meta announced OpenZL as a format-ware compression framework focused on delivering both high compression ratios and high speed compression/decompression. OpenZL has continued making more optimizations and out today is OpenZL v0.3 and advertising some huge gains over Zstandard (Zstd) that is also worked on by Meta engineers...

phoronix.com ·

We haven't written up this one. Phoronix has the full story — the link below goes straight to it.

Read the original at phoronix.com →

More in Tech

I Built a Support Agent That Remembers Customers Between Chats

A support agent can give the right answer and still make a customer repeat themselves. I built this system around a simple idea: use long-term memory to carry useful context between conversations, but…

  • Memory layer uses Hindsight with unique bank per customer for retrieval.
  • Retrieval is tailored to current query, preventing random reiteration of stored information.

·dev.to

We Didn’t Have a Product Problem. We Had a Memory Problem.

Building PRATHIDHWANI: A Memory-Powered Customer Feedback Intelligence System Using Hindsight A few months ago, I was sitting in a product discussion when someone asked a deceptively simple question…

  • Team discovered lack of clear memory of past decisions hindered addressing customer feedback
  • PRATHIDHWANI system connects customer feedback, product changes, and historical context
  • Dashboard provides visual overview of feedback trends and historical signals

·dev.to

How Many Tokens Is That Elasticsearch Hit? A Reproducible RAG Compression Benchmark

This is a follow-up to my earlier post introducing jtoken . That one was the pitch. This one is the actual measurement — the benchmark I run before claiming any savings number, and how you can run it…

  • jtoken compression library evaluated for token savings in RAG pipeline
  • Benchmark uses real tokenizers (tiktoken) to measure compression impact
  • jtoken achieves 19.1% reduction for MongoDB docs, 13.2% for nested API events

·dev.to

More from Tuesday 29 September →