7111
How We Found 7 TiB of Memory Just Sitting Around
How we saved 7 TiB of memory across our Kubernetes clusters by disabling namespace listwatching in Vector, reducing daemonset overhead and API server load at scale.
New
The latest blog posts to take off on Hacker News. 80,826 posts.
7112
Learning from failure to tackle hard problems
This blog post is based on the work BaNEL: Exploration Posteriors for Generative Modeling Using Only Negative Rewards. Tackling Very Hard Problems The ultimate aim of machine learning research is to push machines beyond human limits in critical applications, including the next generation of th
7113
How the cochlea computes (2024)
Sensory coding 2: electric boogaloo
7114
Signs of introspection in large language models
Research on the ability of large language models to introspect on their own internal states.
7115
Qt Creator 18 Released
Qt Creator 18 highlights development containers to ease kit configuration, a convenient Overview tab, and various UX improvements.
7116
Dating: A mysterious constellation of facts
Are speed dating and network effects contradictory?
7117
SPy: An interpreter and compiler for a fast statically typed variant of Python
Antonio Cuni's blog
7118
PlanetScale Offering $5 Databases
Introducing the $5 PlanetScale plan.
7119
Free software scares normal people
80% of the people only need 20% of the features.
7120
RISC-V takes first step toward international ISO/IEC standardization
At RISC-V Summit North America 2025, Andrea Gallo, CEO RISC-V International, and Phil Wennblom, Chair of the ISO/IEC Joint Technical Committee (JTC 1)., announced that RISC-V International has been approved as a PAS Submitter by the ISO/IEC JTC1.
7121
One year with Next.js App Router and why we're moving on
A critique of React Server Components and Next.js 15.
7122
7123
Crunchyroll is destroying its subtitles
With the Fall 2025 anime season, Crunchyroll demonstrates zero respect for anime as a medium as the presentation quality of its subtitles reach an all-time low.
7124
A century of reforestation helped keep the eastern US cool (2024)
Throughout the 20th century, as most of the world warmed, the eastern US remained mysteriously cool. Now, scientists think they have an explanation: trees.
7125
Llamafile Returns
Mozilla.ai is adopting llamafile to advance open, local, privacy-first AI—and we’re inviting the community to help shape its future.
7126
How to Obsessively Tune WezTerm
An extremely detailed guide on how to configure WezTerm
7127
Independently verifying Go's reproducible builds
Introducing Source Spotter, a Go Checksum Database auditor and Go toolchain reproducer
7128
Uv is the best thing to happen to the Python ecosystem in a decade
Released in 2024, uv is hands-down the best tool for managing Python installations and dependencies. Here's why.
7129
Dithering – Part 1
Understanding how dithering works, visually.
7130
The Internet runs on free and open source software and so does the DNS
The ICANN Security and Stability Advisory Committee recently published SAC132: The Domain Name System Runs on Free and Open-Source Software.
7131
Developers are choosing older AI models
At Augment Code, we run multiple frontier models side by side in production. This gives us a unique vantage point into how different models behave in real…
7132
The Green Tea Garbage Collector
Go 1.25 includes a new experimental garbage collector, Green Tea.
7133
Tailscale Peer Relays
Tailscale Peer Relays provide predictable, performant access, entirely within your own tailnet.
7134
7135
Composer: Building a fast frontier model with RL
Composer is our new agent model designed for software engineering intelligence and speed.
7136
Replacing EBS and Rethinking Postgres Storage from First Principles
Fluid Storage reimagines block storage like EBS with zero-copy forks, true elasticity, and sync replication. Built for the age of agents.
7137
I made a 10¢ MCU Talk
I taught a 10¢ CH32V003 RISC‑V MCU to talk. By bit-banging PWM as a DAC and using a tiny 2‑bit ADPCM decoder, I squeezed ~6–7 seconds of recognizable audio into 16 KB of flash. Then I went full ret...
7138
Kafka is Fast – I'll use Postgres
Why you should just use Postgres instead of Kafka for small-scale message queuing and pub-sub patterns. Benchmarks and practical tests included.
7139
From VS Code to Helix
I want less Big Tech in my life. After years of dismissing Helix as too fringe and complicated, I've tried (and loved) it.
7140
Zig's New Async I/O
In celebration of the std.Io introduction patchset landing today , here is the text version of a short, interactive demo I gave at Zigtoberfest 2025 .