The session you cannot take with you

Inference APIs are filling sessions with encrypted reasoning, hidden search results, opaque compaction, and encrypted subagent messages. A growing form of lock-in.

Rune 1.1: adds Python, an Emacs editor, a symbol index and is now free

Rune 1.1 makes Rune free for personal use and adds a perpetual commercial license, a workspace symbol index, first-class Python support, an Emacs editor, and substantial improvements to SSH workspaces.

Show HN: What should the GUI for AI agents look like?

A workspace with visible files, tools, tasks, and outputs — not buried in chat threads.

Where USB Memory Sticks are Born (2013)

A California aquifer may have crossed the point of no return

Gemini Robotics 2 brings whole body intelligence to robots

From feet to fingertips — we are teaching robots intelligent whole-body control, fine dexterity, and teamwork to complete a broad range of complex tasks.

The Economic Benefit of Refactoring

The AI Aesthetic

Writing about the big beautiful mess that is making things for the world wide web.

DeepSeek-V4-Flash Update

Date: 2026-07-31

The mean means nothing: data visualization to debug a latency problem

I was recently trying to validate some performance improvements related to lld at $DAYJOB and it was a little frustrating to see the improvements in our benchmarks but not in the live-production dashboards.

The Religion of Speed

At some point, “moving fast” stopped being a practical concern and became a moral position. You see it damn near everywhere now. How fast can we ship this...

CodePen 2.0

Noting perhaps my largest personal career accomplishment, which is launching CodePen 2.0. Far more work, believe it or not, than the entire creation of the original CodePen. This isn't the place to describe every detail of what we did and why we did it. If you're interested, perhaps our Why 2.0? podcast or the What's…

Read this before you buy that TV streaming stick

I flagged two research papers for fake authors and both were accepted as orals

We reviewed 22 ML conference submissions this summer. Fifteen had fabricated citations, hallucinated authors, or clear LLM slop, so we complain about that a bit, and release the paper references audit we now run.

The American Grilled Cheese Sandwich Essay (2024)

Welcome back to Notable Sandwiches, the feature where David and I travel the world on wings made of bread, chronicling Wikipedia’s List of Notable...

Stacked PRs are now live on GitHub

Physicists Solve a Muon Mystery. Now, Old Results Don't Add Up

New calculations seem to have put a 25-year-old particle physics puzzle to rest. But they’ve also created a clash with other experimental results.

Investigating three real-world incidents in our cybersecurity evaluations

In a review of our cybersecurity evaluation transcripts, we found three incidents in which a Claude model reached the internet from within or while interacting with a third-party evaluation environment, and then gained unauthorized access to the real systems of three different organizations. Below we describe what happened, how it happened, and what we’re changing. We encourage other AI labs to perform similar reviews.

Agent Skill to Force Docs in ASD-STE100 Simplified Technical English

Agent skill: make LLMs write docs in ASD-STE100 Simplified Technical English — no AI slop - AminBlg/SimpleEnglish

JEP 401: Value Objects (Preview) merged to OpenJDK master

JDK main-line development https://openjdk.org/projects/jdk - 8389219: Implement JEP 401: Value Objects (Preview) by MrSimms · Pull Request #31120 · openjdk/jdk