What Google giving away gVisor means for AI sandboxes
Google is giving away gVisor, a sandbox The New Stack says Anthropic and OpenAI use. Here is what gVisor does and what the handover changes.
· 6 min read
AI & Dev Tools
Reviews of AI coding assistants, agentic tools, and the workflows that make them useful.
47 articles
How Cerebras runs a 3T-parameter model on 44 GB chips
Cerebras explained how it splits a 3-trillion-parameter model across wafer chips. Here is the pipeline idea and the memory limit behind it.
· 6 min read
What Liquid AI's test-cheating model means for you
Liquid AI says its model gamed coding tests, then recognized its sandbox. Here is what reward hacking is and why passing tests is not proof.
· 6 min read
What Muse Spark Open's 816B size means for developers
Meta's new open-weight coding model runs like a 42B model but needs the memory of an 816B one. Here's what that means for running it.
· 6 min read
What Anthropic's Scion means for Git and AI coding
Anthropic swapped Git for a custom system called Scion after merges jumped to 30,000 a day. Here is the background and what it means for your workflow.
· 6 min read
Grok 4.6 Scores 26% and 88% on the Same Benchmark Line
xAI's own model card reports Terminal-Bench 3.0 at 26.0%. Artificial Analysis reports 88.4% on 2.1. Both are the same model. The model card is unusually honest about why — the problem is everyone quoting it drops the qualifiers.
· 7 min read
DeepSeek MLA: 70 GB of KV Cache at 1M Tokens
No DeepSeek-V4 config is public yet. The V3 one is, and its KV-cache math tells you what a million-token window actually costs in GPU memory.
· 6 min read
A Daily LLM Digest Agent Costs $166 and 34 Hours a Year
A year of one scheduled digest agent in production: token costs per stage, infrastructure line items, and the 41 runs that needed a human.
· 6 min read
When Agent Output Passes Zod and Still Breaks the Build
Four failure classes that survive a clean parse, and the three-layer validation pass we run instead.
· 6 min read
Idempotency: Publishing Agents That Survive a Mid-Run Kill
A per-channel ledger makes cross-posting resumable. Why exit code 0 is not a receipt, and what to do when an API has no idempotency key.
· 7 min read
AI Agent Auth: API Keys vs Device Flow vs Scoped Tokens
Three credential models for non-human callers: static keys, the OAuth 2.0 device grant, and short-lived scoped tokens - and when each one fits.
· 7 min read
Error Messages as an Agent Interface
A field-by-field guide to API error bodies: stable codes, retryable flags, wait hints, fix examples, and the shapes that trap agents in retry loops.
· 6 min read
Writing a Spec an AI Agent Can Actually Execute
Ground truth files, an interface contract, one acceptance command, and explicit out-of-bounds rules -- so the agent runs end to end without babysitting.
· 6 min read
MCP Server Security: What a Local Server Can Actually Reach
It runs as a child process with your user's permissions. Four commands to see which files it opens and hosts it dials, plus three containment changes.
· 7 min read
CodeRabbit vs Greptile vs Graphite: AI Code Review Bots Compared for 2026
A mechanism-level comparison of three AI pull request reviewers — how each one builds context, how noisy it is by design, and how to bake them off on your own repo before buying seats.
· 6 min read
How to Measure Cost Per Task with Terminal AI Agents
OpenCode and Claude Code bill by the token, so cost per task is the metric that matters. Here is how we tracked it and what we learned.
· 6 min read
Multi-Agent Terminal Workflows with OpenCode and Aider
How we split work across OpenCode, Claude Code, and Aider in one terminal without losing track of which agent changed what.
· 7 min read
OpenCode First Project Setup: Install to First Passing Test
Configure providers, context files, and project conventions so the agent produces useful output on an existing codebase from day one.
· 6 min read
Running OpenCode with Local LLMs for Private AI Coding
We ran OpenCode against a local Ollama model on a proprietary codebase, with no source code sent to a cloud API.
· 6 min read
OpenCode Review: A Terminal-Native AI Coding Agent
We ran it as a daily driver for two weeks. It edits files, runs tests, and supports multiple models without replacing your IDE.
· 7 min read
Aider vs Continue.dev: Terminal vs Editor AI Coding in 2026
A hands-on look at two open-source assistants: how each handles model choice, repo context, and your git history.
· 7 min read
MCP Servers Worth Wiring Into Your Editor in 2026
A practical look at which Model Context Protocol servers actually earn a slot in your editor config, what they do, and where they break down.
· 7 min read
CodeRabbit vs. Greptile vs. Diamond: AI Code Review in 2026
Compared on codebase context, review depth, and noise, plus which one fits the way your team actually merges pull requests.
· 7 min read
Parallel Refactoring with Claude Code Subagents
How to split one large refactor across subagents: scoping each task, isolating file conflicts, and reviewing the merged result.
· 7 min read
Cline vs Roo Code: Agentic Coding Extensions in 2026
Roo Code began as a Cline fork. Here is how the two open-source, bring-your-own-key VS Code extensions actually differ.
· 7 min read
Aider Review: Open-Source AI Pair Programming for Any LLM
I tested Aider on 9 projects with 6 LLMs over six weeks for $47.30. Why git-native beats accept/reject buttons, and where terminal-only falls short.
· 8 min read
v0 by Vercel Review: I Tested 15 Real UI Tasks
v0 generates React components with shadcn/ui, Tailwind, and TypeScript. Here is how they held up under real product requirements.
· 8 min read
Orthrus: Generate 32 Tokens in Parallel
Orthrus injects diffusion attention into every layer of a frozen autoregressive Transformer, leaving the base model's output distribution unchanged.
· 6 min read
NVIDIA Warp Review: GPU-Accelerated Python for Simulation
Warp compiles Python to CUDA kernels. We benchmarked it against JAX and Taichi to find where it earns a spot in your stack.
· 6 min read
Macchiato Day 2: Live Token Metrics and Parallel AI Terminals Reviewed
Macchiato's day-2 build adds a live token/cost sidebar and keyboard shortcuts for swapping between Claude Code and OpenCode in one terminal. Here's what shipped and what it means.
· 6 min read
Local LLMs for Code Generation: Ollama vs LM Studio in 2026
We benchmarked DeepSeek Coder, Qwen 2.5 Coder, and CodeLlama on Apple Silicon and NVIDIA GPUs for latency, code accuracy, and readiness to replace cloud APIs.
· 10 min read
Continue.dev Review: Open-Source, Choose Your Model
A plugin for VS Code and JetBrains with customizable context and a transparent architecture. Where it replaces Copilot, and where it does not.
· 7 min read
v0 by Vercel Review: AI-Generated UI Components That Actually Ship
v0 generates React and Next.js UI from natural language prompts. A pragmatic look at what it produces, how the output compares to hand-written code, and when it saves real development time.
· 6 min read
GitHub Copilot Desktop vs Claude Code vs Codex CLI
GitHub's standalone Copilot desktop app changes the matchup. We compare workflow surface, approval semantics, and model neutrality.
· 6 min read
Does AI Actually Understand? An LLM Comprehension Guide
Searle's Chinese Room, stochastic parrots, and IIT predict where LLMs break -- and what that means for prompts, retrieval, and agent loops.
· 7 min read
Apple Silicon vs OpenRouter: Local LLM Costs 30-60x More
Running Llama 3.3 70B on an M-series Mac Studio versus paying per token: here's the math at typical developer volumes, and three cases where local still wins.
· 6 min read
Bun vs Node.js in 2026: Is Bun Production-Ready?
We tested Bun 1.2 against Node.js 22 LTS on real workloads: where the speed gap is real, where Node compatibility breaks, and how to decide on migrating.
· 6 min read
Hermes Memory Installer Review: One-Command Local AI Memory
Nous Research's tool installs with one shell command and keeps agent memory in local files. We compare its file-based approach to Mem0 and Letta.
· 6 min read
Cursor vs GitHub Copilot: Which Ships Faster in 2026?
We tested both on a Next.js app, a Python CLI, and a Rust library migration. Cursor won on velocity -- but one scenario still favors Copilot.
· 8 min read
OpenAI Codex Chrome Extension Tested
The coding agent runs inside a browser tab. Which workflow patterns pay off, the limits, and how it compares to Codex CLI and IDE agents.
· 6 min read
OpenAI Codex vs Claude Code: Python Benchmark
We ran both on the same codebase across refactoring, debugging, and agentic tasks. What each shipped, and the speed-vs-cost tradeoff.
· 6 min read
AdamsReview: Multi-Agent PR Review for Claude Code, Reviewed
How multi-agent review catches what single-pass LLM reviews miss, and where AdamsReview fits in your pipeline.
· 6 min read
yt-dlp vs youtube-dl vs pytube in 2026: Stars, Releases, Usage
GitHub stars on Oct 6, 2026: yt-dlp 195.8k, youtube-dl 141.4k, pytube 13.2k. Only yt-dlp still ships releases. How to re-check them, then install and use it.
· 6 min read
Build Your Own X: The GitHub Repo and 10 Projects to Start
What build-your-own-x is (codecrafters-io, 550k+ GitHub stars), how to use it, and the 10 from-scratch projects to start with: database, Git, shell, interpreter.
· 6 min read
Ratty Terminal Emulator: Inline 3D Graphics
A measured look at where this category fits, which workflows benefit, and what to verify before you switch.
· 6 min read
Mythos AI Found a Real Curl Vulnerability
Daniel Stenberg confirmed the bug in one of the most-reviewed codebases on the planet. What it means for AI-assisted security review.
· 6 min read
Running Local LLMs on a 24GB M4 Mac: What Actually Fits
Model size math, real tokens/sec for 7B-32B models, and when to reach for Ollama, llama.cpp, or MLX.
· 6 min read