Noxu DB, a Rust port of Berkeley DB Java Edition

Search engine internals: how to win "Search Benchmark, The Game"

Back in March, we entered the Search Benchmark, The Game with a search engine we’ve been working on called IResearch. It’s a C++ native alternative to Lucene or Tantivy that isn’t widely known yet, but we actually ended up winning it. The cool part for us was that the Tantivy maintainers validated the results and approved the commit themselves. We appreciate this competition and treat it as a way to contribute our findings back to the search and information retrieval community, so we’...

Nim-Presto – REST API Framework for Nim Language

New accessibility features powered by Apple Intelligence

Stilta raises $10.5M from a16z and YC to help companies rediscover the patents they forgot they had

Annie Ghazali: PostgreSQL’s Growing Role in AI Infrastructure

PostgreSQL, often through platforms like Supabase, is increasingly becoming part of the default stack for many AI applications. That level of adoption says something important about where engineering teams are placing their trust.Supabase has become one of the most common starting points for AI products. Most AI frameworks support PostgreSQL and pgvector directly. For many teams, PostgreSQL is already part of the stack before the AI layer is even introduced.  ...

Gaussian Splat of a Strawberry

Show HN: Id-agent – Token efficient UUID alternative for AI agents

The Windows DLL loader lock: how a Rust thread can hang your JVM

I Found Ultra-Pure Quantum Crystals in an Abandoned Mine in the Atacama Desert

Vibe Coding with SoundCloud: Announcing New Docs for your Agents

Have you ever had a brilliant idea for a music app, a custom playlist curator, or a unique way to discover underground tracks on SoundCloud…

Even More Tagged Union Subsets with Comptime

Built a DBMS from scratch in C to study buffer pool behavior on real SQL workloads

I’m a third-year CS student and over the past year I’ve been building minidbms — a database engine written from scratch in C and Python — to study buffer pool replacement policies experimentally. Current features: - slotted-page heap storage - direct pread/pwrite I/O - LRU / Clock / NoCache / OPT - trace-based telemetry replay - benchmark + sweep analysis tools - interactive cache inspector - B+ tree indexes (in progress) Some interesting results so far: - Bélády-related behavior re...

My domain got abused on GitHub Pages

A new EDIT tool for LLM agents

Right now I'm working to an agent for my DS4 project. Local inference is token-poor, it's a battlefield where optimizations count. I was quite surprised by the fact the EDIT tool everybody is using right now forces the LLM to emit the old version of the text verbatim. This CAS (check and set) mode of operation, where I say EDIT old="foo" new="bar", is needed because there are often colliding edits (the user is editing as well, or checked out a different branch, and so forth) and because the L...

The lasting influence of Netscape Time

Show HN: I built a sovereign OS, L1 blockchain, AI agent, and language

Humble Tech Book Bundle: The Ultimate Functional Programming

Theo Baker spent four years investigating Stanford. Before he leaves, here’s what he found.

Cursor Introduces Composer 2.5

Show HN: Clark-Browser – Stealth Chromium

LLMCap – A proxy that hard-stops LLM API calls when you hit a dollar cap

Codex-Maxxing

Show HN: Hsrs – Type-Safe Haskell Bindings Generator for Rust

Sieve – scans Cursor/Claude chat history for leaked API keys

As of April 2026: Iran has destroyed 42 U.S. Military Aircraft in Op: Epic Fury

AgentCRM – Headless CRM for Claude Code

Your fridge could be a threat to national security

On the floor of HumanX, Ryan is joined by Adam Meyers,  Senior VP of Counter Adversary Operations at Crowdstrike, for a deep dive on their latest Global Threat Report that tracks over 281 adversaries across nation states, e-crime, and hacktivist organizations.

PyTorch Landscape

D$^2$Evo: Dual Difficulty-Aware Self-Evolution for Data-Efficient Reinforcement Learning

arXiv:2605.17037v1 Announce Type: new Abstract: Reinforcement learning (RL) has demonstrated potential for enhancing reasoning in large language models (LLMs). However, effective RL training, which requires medium-difficulty training samples, faces two fundamental challenges: Effective Data Scarcity and Dynamic Difficulty Shifts, where medium-difficulty samples are scarce and become trivial as models improve. Existing methods mitigate this scarcity to some extent by generating training sampl...

Reliability and Effectiveness of Autonomous AI Agents in Supply Chain Management

arXiv:2605.17036v1 Announce Type: new Abstract: This paper studies autonomous generative AI agents in multi-echelon supply chains using the MIT Beer Game. We identify four inference-time levers that shape performance: model selection, policies and guardrails, centralized data sharing, and prompt engineering. Model capability is the dominant factor: an out-of-the-box reasoning model exceeds human-level performance, and optimized reasoning models reduce costs by up to 67% relative to human tea...

Privacy Policy Enforcement Guardrails for Data-Sensitive Retrieval-Augmented Generation

arXiv:2605.17034v1 Announce Type: new Abstract: Standard PII filters often miss contextual data leakage in RAG systems, such as non-regulated attribute clusters that collectively identify individuals. We introduce a Privacy Policy Enforcement (PPE) framework using dual one-class density estimators with fused text embeddings and a calibrated abstain region for out-of-distribution inputs. Using an axis-stratified, multi-LLM synthetic data pipeline across medicine, finance, and law, we found th...

Generalizable and Actionable Parts Pose Estimation with Symmetry Annotation-Free Learning Strategy

arXiv:2605.17033v1 Announce Type: new Abstract: Urgently needed generalizable robot object interaction and manipulation requires high-quality Cross-Category object perception. As a pioneer of this area, Generalizable and Actionable Parts (GAParts) understanding has attracted increasing attention from relevant researchers. However, most recent works either have insufficient design regarding the symmetry issue or require rich symmetry annotation, which severely impedes precise GAPart pose esti...

Task Abstention for Large Language Models in Code Generation

arXiv:2605.17029v1 Announce Type: new Abstract: Large language models (LLMs) have revolutionized automated code generation. One serious concern, however, is the so-called ``hallucination'', i.e., LLMs may generate seemingly plausible but functionally incorrect code. In this paper, we study the task abstention problem, i.e., determining whether a given LLM should abstain from performing a specific code generation task to avoid likely hallucination. Our approach features a calibrated abstentio...

PARALLAX: Separating Genuine Hallucination Detection from Benchmark Construction Artifacts

arXiv:2605.17028v1 Announce Type: new Abstract: Large language models (LLMs) hallucinate with confidence: their outputs can be fluent, authoritative, and simply wrong. In medical, legal, and scientific applications this failure causes direct harm, and detecting it from internal model states offers a path to safer deployment. A growing body of work reports that this problem is increasingly tractable, with recent methods achieving high detection performance on widely used benchmarks. We show, ...

Why Do Reasoning Models Lose Coverage? The Role of Data and Forks in the Road

arXiv:2605.17026v1 Announce Type: new Abstract: Recent progress in large language models has led to the emergence of reasoning models, which have shown strong performance on complex tasks through specialized fine-tuning procedures. While these methods reliably improve pass@1 accuracy, prior works have observed that they show a coverage shrinkage behavior, where pass@k degrades relative to the base model. In this paper, we investigate the reasoning shrinkage arise under SFT-based post-trainin...

Intermediate Constacyclic Codes and Scalar-Residue Reed--Muller Layers

arXiv:2605.17022v1 Announce Type: new Abstract: A 2024 paper of Sun, Ding and Wang introduced a second class of constacyclic codes over finite fields, denoted $C(q,m,r,\ell)$, with length $(q^m-1)/r$, where $r\mid(q-1)$ and the defining monomials have total $q$-ary degree congruent to $r-1$ modulo $r$. In the non-projective intermediate range $2

ASMTP – SMTP for Agents

War Game Exposed U.S. Vulnerability to Low-Tech Warfare

PgBackRest Will Continue

The last six months in LLMs in five minutes

Data Center Waste Heat as an Emerging Urban Thermal Hazard

.at TLD is top on CF Radar

Regex Chess: A 2-ply minimax chess engine in 84,688 regular expressions

Leaving the Sea of Nodes in the V8 JIT

Meta Reassigns 7k Employees to Focus on A.I

SIMD-accelerated integer-to-string conversion

When can the C++ compiler devirtualize a call?

What political censorship looks like inside an LLM's weights (Qwen 3.5)

Casuarina Linux

Microsoft surprises with its first server Linux distribution: Azure Linux 4.0

Christophe Pettus: All Your GUCs in a Row: bonjour and bonjour_name

PostgreSQL's Bonjour parameters let you advertise a server on the local network via Apple's service-discovery protocol—a clever 2002 idea that hasn't aged well.

Here’s why Elon Musk lost his suit against OpenAI

On Monday, the jury in Musk v. Altman dealt Elon Musk a major blow—reaching a unanimous advisory verdict that Musk sued OpenAI too late and, as a result, his claims are barred by the applicable statutes of limitations. US District Judge Yvonne Gonzalez Rogers immediately accepted it.  Musk announced on X that he will be…

Bytecode Interpretation From Scratch in C

Coding on Paper

Anthropic co-founder to present AI encyclical alongside Pope Leo XIV

Canceled by Hinge

Introducing Casuarina Linux: A glibc-Based Chimera Linux Derivative

What Do Gödel's Incompleteness Theorems Mean?

Mexican government breached by solo user with Claude, 150 GB exfiltrated

CISA Admin Leaked AWS GovCloud Keys on Github

Alignment pretraining: AI discourse creates self-fulfilling (mis)alignment

Who will buy your services if you fire us all?

OSHA probing worker death at SpaceX’s Starbase site

US bill proposes new national EV tax, while some push to slash gas tax to zero

The Enshittification of History

SandboxAQ brings its drug discovery models to Claude — no PhD in computing required

Designing an FPGA Calculator from Scratch

Vibhor Kumar: Beyond Vector Search: Why PostgreSQL Could Become the Memory Layer for Enterprise AI Systems

The conversation around AI infrastructure today is heavily focused on models, GPUs, inference speed, and vector databases. These are important building blocks, but they often distract from a deeper architectural challenge that is beginning to emerge as enterprises move from experimentation toward operational AI systems. The challenge is memory. Not memory in the simplistic sense of storing chat history or embeddings, but memory in the broader sense of maintaining durable context, operat...

X accounts are limited to 50 posts and 200 replies a day unless they pay

The tyranny of single page apps

No More JetBrains Products for Me

Crypto industry lobbies to evade AML/CTF rules

Understanding Singleflight in Go

cargo-crap: Finding Untested Complexity in AI-Generated Rust Code

Show HN: We missed Winamp, so we built an audio player for macOS

Hyperpolyglot Lisp: Common Lisp, Racket, Clojure, Emacs Lisp

We let AIs run radio stations

Aether: High-performance, formally-verified storage engine written in Rust

Show HN: Auto-identity-remove – Automated data broker opt-out runner for macOS