Archive
21 posts
Everything so far
Every post, newest first, grouped by year. Looking for a topic instead? Try the tags.
2026
-
Codex Product Design: Design Website and App UI
Two methods for better UI with Codex: use the Product Design plugin to explore three directions for a website or app, then polish it with the award-tier prompt.
-
Why OpenAI Is Sprinting: Inside the AI Bubble's Capex Trap
OpenAI's 28-day sprint, Anthropic's $2T IPO bid and $800B of hyperscaler capex define the AI bubble's most dangerous phase: real tech, fragile cash flows.
-
Apple AI Strategy: Who Controls the Next Action?
Apple’s AI strategy connects Siri, Foundation Models and App Intents. What its partnerships and agent research establish, and what still needs testing.
-
OpenAI Dots and the Economic Paradox of Personal Agents
OpenAI Dots and GPT-6.1 Sol show the personal agent war has moved from model IQ to inference economics. Why cross-subsidization will decide the First Layer.
-
Why AI Agents Need a Mac: The Hidden Virtualization Limits
Why AI coding agents from Cursor to Devin need real Macs for iOS builds, and the hidden limits: a 2-VM kernel cap, blocked snapshots, and 24-hour leasing rules.
-
September 2026 AI Model Rankings: Reading the Chart
Read the September 2026 AI model rankings as a dated tier chart, with its attribution, missing methodology, and limits made clear before choosing a model.
-
OpenClaw Isn't Dead: It Lives in Big Tech's Answers for Agents
Why OpenClaw's early architecture—persistent environments, computer use, and local runtimes—now defines how Google, OpenAI, Anthropic, and Meta build AI agents.
-
Meta Muse and Grok Bot: Are Cloud-Hosted Agents the Next Form?
Meta Muse and Grok Bot give each agent a persistent cloud computer. Are always-on VM agents the next form? Plus usemuse.lol, built entirely inside Muse.
-
How Chinese Apps Detect Mobile VPNs: Signals and Limits
How Chinese apps can detect mobile VPN signals through network APIs, proxy settings, DNS behavior, and location mismatches—and what those signals cannot prove.
-
How Windsurf Split Into Google Antigravity and Devin
How Windsurf split into Google Antigravity and Devin Desktop—and why Cognition's quiet SWE-2 launch makes that strange AI coding history matter again.
-
Why ChatGPT Can Feel Less Capable: What to Check
When ChatGPT feels less capable, check model selection, reasoning effort, context, and repeatable results before treating a bad answer as a hidden downgrade.
-
GPT Image 2.5: Flare, Sunburst, and 12 Prompts
Explore GPT Image 2.5 Flare and Sunburst, with 12 reusable prompts, eight generated examples, and a practical look at speed, image quality, and editing.
-
AI Workflows: How to Choose Models and Agents
A practical AI workflows guide for choosing models, designing agents, controlling costs, and evaluating real work without trusting benchmarks alone.
-
GPT-5.6 Sol vs. Terra vs. Luna: Which Should You Use?
Fresh CodexRadar data shows when GPT-5.6 Sol, Terra, or Luna makes sense, separating repeatable task coverage from latest-run cost and speed.
-
DeepSeek V4 Flash + GPT-5.6 Luna: A $0.42 Sol Alternative?
DeepSeek V4 Flash and GPT-5.6 Luna theoretically match Sol xhigh's coverage for about $0.42, but only if a verifier can identify the correct patch.
-
How to Use GPT-5.6 Luna Max as a Codex Subagent
GPT-5.6 Luna is now 80% cheaper in the OpenAI API, making Luna Max practical for bounded Codex subagent tasks while Sol handles planning and review.
-
GPT-5.6 Sol Max vs. xhigh: Why Max Isn't My Default
GPT-5.6 Sol passed fewer repository tasks at max than xhigh while costing 46% more. This benchmark maps the full Codex effort curve across 112 tasks.
-
GPT-5.6 Juice Values for Sol, Terra, and Luna
Personal GPT-5.6 Juice value observations for Sol, Terra, and Luna across reasoning effort levels in Codex, plus prompts used to reproduce them.
-
How to Build a Custom LLM Evaluation You Can Trust
Build a custom LLM evaluation with hidden tests, mechanical scoring, blind review, and checks for variance, leakage, and framework confounders.
-
Astro Blog SEO Audit: What Actually Mattered
An Astro blog SEO audit uncovered missing social cards, 12 MB of images, inconsistent canonical URLs, and weak archive and tag discovery.
-
Claude Opus 5 vs. Claude Fable 5: Eight Tests, Same Prompts
Claude Opus 5 and Claude Fable 5 tied on most of eight tests. Fable won the decision memo, while Opus delivered the same objective score at half the price.