frr.dev

An AI agent made over 500 errors using a CLI across 165 sessions.

Cloudflare's in-dashboard AI assistant silently provisioned a permanent, read-everything API token across all my accounts, zones and users.

Brave Shields blocks Sentry even with shields down.

I analyzed 1,324 real calls from Claude Code to my Linear CLI. 15.9% failed.

The same architectural pattern solves sentiment analysis and image cleanup: deterministic first, CoreML only for what remains.

In August 2021, researchers broke NeuralHash in less than two weeks.

The EU Chat Control and Apple’s failed on-device scanning rely on the same 40-line Python algorithm.

Agent Experience (AX) principles are changing how CLIs are designed when the primary user is an LLM.

I built a macOS app to stress-test every capability of Apple Intelligence on-device.

Google's TurboQuant paper got 575 points on HN and crashed memory stocks.

Coding agents use $15/M token cloud models for commit messages. foundation-hooks replaces them with Apple's on-device LLM: 300ms, 0 tokens, 0 cost.

libfoundationmodels exposes Apple's on-device LLM as a 4-function C dylib.

Apple's NLTagger scores 'delete the temp file' as -0.8 (very negative).

Apple Intelligence includes a 3B-parameter model on your Mac.

My plan to distribute work among Claude, Codex, and GLM/Z.AI without automating too early, cutting costs without losing judgment.

If you're coming from Claude Code, this will save you some serious frustration: in Codex, a skill is not a slash command.

Apple's NLTagger scores perfectly neutral technical text as negative.

165 sessions with Claude Code and 27 with Codex CLI reveal a clear pattern: each agent excels in different tasks.

Python has had a fragmented and slow tooling ecosystem for years.

XML repeats tags, JSON repeats keys.

MDD: a methodology for designing tools that absorb the hallucinations of your AI copilot.

My Linear skill had 150 lines of workarounds: hardcoded UUIDs, curl fallbacks, notes like 'the CLI doesn’t support X.'

Your AI generates plausible code against non-existent APIs.

Linear launched their AI agent.

TurboQuant compresses LLM KV cache 6x without losing precision.

LLMs generate code that compiles but abuses strings, optionals, and nil where specific types should exist.

A developer got Claude and Codex to argue about their plan until it was solid.

Mole promises to clean and optimize your Mac.

Your AI agent doesn't remember where your projects are. zoxide and CLAUDE.md are the solution: a smart cd and a memory tattoo so they stop getting lost.

How a scalar Kalman filter reduces network calls by 80% in a menu bar app, adapts to battery status, and sleeps on its own when you're inactive.

How I combine Ghostty, git worktrees, and Claude Code in my daily macOS development workflow.

Codex CLI doesn’t support --worktree or --tmux.

Claude Code has native support for git worktrees: --worktree, isolated sub-agents, custom agents with isolation, --tmux for parallel sessions.

PostgreSQL without sharding, a while loop without architecture, cron without cloud platforms.

Prompt engineering is writing a good prompt.

Tired of Codex CLI asking permission for every file?

Email is an inbox anyone can fill with junk.

I designed a council of five simulated LLM experts to evaluate MVPs before writing code.

Codex CLI and Claude Code look like magic, but inside they're just a loop with 5 phases, an LLM, and tools.

OpenAI serves 800M ChatGPT users with a single PostgreSQL primary and ~50 replicas.

A practical tutorial to replicate OpenAI's Codex Automations using Claude Code, systemd timers, and Gitea.

Jane Street hid MD5 inside a neural network with integer weights.

Prompt caching reduces input cost by 90%, but cache writes cost 25% extra.

/simplify by Claude Code reviews your code with three generic agents.

Claude Code 2.1.71 introduces /loop, an inline scheduler for recurring tasks.

xctrace exports 33,000-line XML files that blow up any LLM's context window. ztrace condenses them into 10 actionable lines.

When an AI agent operates your ETL pipeline, forbidding things doesn't work.

Anthropic announced Remote Control for Claude Code.

I’m porting 8,300 lines of Python to Rust using LLMs as copilots.

An LLM wrote code that read and parsed a JSON file from disk on every iteration of a 900-file loop.