Micropatching Brings the Abandoned Equation Editor Back to Life (2018)

Self-Driving bus in Sweden crashes with tram on first day of passenger service

The User Is Visibly Frustrated

Notes on Pope Leo XIV's Encyclical on AI

Accuracy Analysis of the Proxy Point Method with Applications to Some Toeplitz Matrices

arXiv:2605.24231v1 Announce Type: new Abstract: For some kernel matrices, low-rank approximations can be quickly obtained via analytic techniques. One important class of analytic methods that has received attention in recent years is based on the use of proxy points. Accuracy analysis for various proxy point methods has often been heuristic in nature, other than for certain special kernels. For more general cases, the methods lack an explicit number or location of proxy points required to yi...

How Well Do Models Follow Their Constitutions?

arXiv:2605.24229v1 Announce Type: new Abstract: Frontier AI developers now train models against long written behavioral specifications, such as Anthropic's constitution (Anthropic, 2025a) and OpenAI's Model Spec (OpenAI, 2025a), integrated into post-training via methods like character training (Anthropic, 2024) and deliberative alignment (Guan et al., 2024). These documents serve a governance function, but it is unclear how well models actually follow them under adversarial, multi-turn press...

ECo-MoE: Embodiment-Conditioned Mixture of Experts Increases the Evolvability of Robots

arXiv:2605.24225v1 Announce Type: new Abstract: In this paper, we introduce a model of evolution and learning in robots that co-optimizes a distribution of latent design vectors (genotypes) and a mixture of control experts (neural modules), which are gated by the latent coordinates of each decoded design (phenotype). This provides a scalable alternative to co-design algorithms that either train an individual policy for every robot, which is inefficient, or a monolithic universal controller f...

Sketch Bug: Using Sketch-Based Input for Interactive Code Debugging

arXiv:2605.24228v1 Announce Type: new Abstract: We investigate sketch-like pen input as an alternative way to support execution control in interactive debugging. In our interface, programmers draw lightweight marks to set breakpoints, use symbolic strokes to control execution, and extend strokes into spirals to repeat traversal actions. The prototype combines gesture recognition with Python execution tracing in a conventional editor interface. In a controlled study with 24 programmers, we co...

Analyzing the Effects of Two-Stage Peer Evaluation

arXiv:2605.24222v1 Announce Type: new Abstract: Peer-evaluation and selection systems are used when sets of agents evaluate each other in order to select the best $k$ among them. These are commonly used in real-world settings, including academic conferences where those reviewing papers are often the set of submitters. Conferences have attempted to better allocate their reviewing resources by moving to a two-stage mechanism, in which some papers are eliminated after a first stage of review an...

Polar: Agentic RL on Any Harness at Scale

arXiv:2605.24220v1 Announce Type: new Abstract: Reinforcement learning for language agents increasingly depends on custom harnesses that manage long-running context, multi-turn tool use and multi-agent orchestration. However, porting these harnesses into RL environment interfaces remains difficult and often loses important training signals. We bridge this gap with polar, a rollout framework for scalable asynchronous RL over arbitrary agent harnesses. Polar treats the agent harness as a black...

Beyond Final Answers: Auditing Trajectory-Level Hallucinations in Multi-Agent Industrial Workflows

arXiv:2605.24219v1 Announce Type: new Abstract: Large Language Models (LLMs) are increasingly deployed as autonomous agents that reason, use tools, and act over multiple steps. Yet most hallucination benchmarks still evaluate only the final output, missing failures that originate in intermediate Thought-Action-Observation steps. We present Trajel, a dataset and evaluation framework for auditing trajectory-level hallucinations in multi-agent industrial workflows. Trajel introduces a five-type...

QUEST: Training Frontier Deep Research Agents with Fully Synthetic Tasks

arXiv:2605.24218v1 Announce Type: new Abstract: Deep research agents extend the role of search engines from retrieving keyword-matched pages to synthesizing knowledge, fundamentally changing how humans interact with information. However, frontier systems remain proprietary, while existing open agents often generalize poorly across different task types, leaving unclear how to train a broadly capable deep research agent. We release QUEST, a family of open models (ranging from 2B to 35B) that s...

Identifying and Mitigating Systemic Measurement Bias in Production LLM Inference Benchmarks

arXiv:2605.24217v1 Announce Type: new Abstract: As Large Language Models (LLMs) transition from research environments to production deployments, evaluating their performance against strict Service Level Objectives (SLOs) has become critical. However, current evaluation methodologies suffer from severe measurement bias at scale. We demonstrate that widely used benchmarking utilities rely on single-process, asyncio-driven architectures that introduce fundamental client-side queuing bottlenecks...

Do you have what it takes to run AI in production?

From the floor of HumanX, Ryan Donovan is joined by Peter Salanki, CTO and co-founder of CoreWeave, to chat about what it really takes to run AI in production; the growing importance of observability, utilization, and scheduling; and Peter’s advice for avoiding the trap of over-architecting too early.

Why the Smart Home Bubble Popped

A Comma and a Question Mark

Maia-3: free and open source

Designing for and against the manufactured normalcy field (2012)

Mathematical Patterns in African American HAIRSTYLEs

TP-7 Field Recorder

Browser-based file encryption tool using WebCrypto

Does Anybody Actually Like React?

Every Frontier AI Is INTJ

Performance of Rust Language [pdf]

Christophe Pettus: All Your GUCs in a Row: client_min_messages

`client_min_messages` controls what your session sees, not what the server logs—a confusion that spawns most of its trouble.

Building an Enterprise Agent Platform: Enforcing Identity, Data, and API Governance

While enterprises deploy AI agents at a rapid pace, their governance strategies often remain fragmented. Most organizations enforce identity, data access, and API security in separate silos, which creates dangerous gaps as agents move across systems. Salesforce Engineering addresses this by shifting the focus from securing individual agents to enforcing governance at the platform level. […] The post Building an Enterprise Agent Platform: Enforcing Identity, Data, and API Governance appeared...

Show HN: Fungible – A local personal finance app in the terminal

Agentic Patterns

How Shamir's Secret Sharing Works

The Vatican-Anthropic relationship that's reshaping the AI ethics debate

Using AI to write better code more slowly

CVE-2026-28952: Apple macOS 26.5 Kernel Vuln found by Claude

Garden Grove chemical tank cools, risk of catastrophic explosion eliminated

Cox Media fined after bragging it spied on users through their phones

Show HN: TryPost – open-source Social Media Scheduler

Google is its own worst enemy

Reticulum: Source-privacy claim vs. routing metadata

A demonstration deployment of Haskell on AWS Lambda

Citing Gandalf, Pope Leo says we must "disarm" AI

Nobody Cracks Open a Programming Book Anymore

Secure TUI to monitor docker containers

Chatbot Has a Long Memory. That Isn't Always a Good Thing

Show HN: OpenBrief – Local-first video downloader/summarizer

Building an AWS Lambda-Like Runtime with Firecracker MicroVMs

Microsoft Copilot Cowork Exfiltrates Files

Hacker News front page as a site

Ninth Circuit Panel Goes Out of Its Way to Question Section 230–DOE vs. Meta

Extremely simple internet radio controlled via IRC

Riscrithm – An intuitive RISC-V assembler and optimizer coded in Go

Anthropic to release Mythos-class models to the public

Ferrari Luce, Maranello's first ever electric car

Yoti age checks share facial photos and device fingerprints with third parties

Jensen–Shannon Divergence

Norway's 2 petabytes of Huawei flash storage and LLM training

Richard Yen: Foreign Tables and Materialized Views: A Dynamic Duo

Introduction I recently wrote a post about WAL log shipping and how a standby built on log shipping is a great way to give data analysts production data without putting the primary at risk. Having access to the production data in this way is great, but it’s read-only. How can we create views of this data for better analytics work? I want to make the case today that Foreign Data Wrappers and Materialized Views can make a great solution – not only in accessing production Postgres data, ...

Splitting Konsole views from Helix to run tools

AWS Weekly Roundup: AWS Local Zones in Istanbul, open-source ExtendDB, Kiro Web, and more (May 25, 2026)

There’s something genuinely energizing about working with startups — something I’ve been doing intensely for more than two years now. Startups operate at a different frequency: the urgency is real, the constraints are tight, and the stakes are personal. Helping them navigate the challenge of proving their business model requires not just technical depth but […]

Where’s that comment?

Hello Tumblr! @jubs here, today, to tell you about a feature we’re building: Comments by author!Comments (or replies) have been around for a while, but they have always been tucked away from your regular feed, and so we’re working on a few ways to integrate them into the rest of our experience.Soon, blogs will have a Comments tab alongside Posts, Likes, and Following. There, you will be able to see posts other people have been commenting on, and, more importantly, find your own comments...

Japan's New Hypersonic Engine Could Make 2-Hour Flights to the US a Reality

Anthropic Cofounder Chris Olah's Remarks on Pope Leo XIV's "Magnifica Humanitas"

Show HN: Volt – front end tooling for Phoenix that runs inside the BEAM

The bootstrapper's EU stack for under €10 per month

Performance of Rust language

Blind Waymo Users Revel in the Joy of Riding Alone

Google facing court for retaliation against Gaza whistleblower

Exit IP VPN servers mitigation rollout

California moves to exempt Linux from its age-verification law after backlash

Introducing USB4STREAM Protocol for Linux – Opening Up Some Nifty Uses for USB4

CPPL: A Circuit Prompt Programming Language

The Coming Layoffs and the Revenge of the Measurers

The analog computer museum's online library

Ubers COO says its getting harder to justify the money spent on AI tokenmaxxing

I manage teams without a single call

What ClickUp’s mass layoff tells us about the future of work

Launch HN: Chert (YC P26) – Twilio for iMessage

Gnutella: A Protocol Outliving the World That Created It

The pope’s AI encyclical isn’t really about AI

The pitch trick that helped an eSports startup raise $20M when VCs only wanted AI

Distributing LLM inference in DwarfStar

High end NVIDIA cards, and the server and power needed to run them, cost a lot of money, especially if you plan to reach enough VRAM to run massive models. The alternative, so far, has been Apple hardware, or the DGX Spark that even if severely limited because of memory bandwidth allows to run LLMs prompt processing (prefill) fast enough. The Mac Studio provided up to 512GB unified memory, a solution with modest memory bandwidth (but much better than the Spark) and compute at a price that was...

Startup Battlefield 200 applications close in days: Apply before May 27