Is Gemini 4 Argon Free? Plans, API Price and Limits
Gemini 4 Argon isn’t free: you need Google AI Pro, Ultra or the paid API. I hit the Pro cap after about 20 turns, and here’s when Argon came back.
AI
Claude Code, Codex and Antigravity break in specific ways when you use them on real projects. I’m a backend developer who runs them at work and on my own code, plus local models on an RTX 4090, and I write down each error and the setting that fixed it.
Local LLM
- Local LLM Too Slow? 8B vs 30B on My RTX 4090
- vLLM vs llama.cpp on a Single RTX 4090: Why I Went Back
- Docker Model Runner vs Ollama: Should You Switch? (RTX 4090)
Agent permissions and safety
- Antigravity CLI Always Proceed Still Asks? Permissions Tested
- Codex CLI npm install Fails? It’s Sandbox Network Access
- Antigravity CLI Remote Control: Setup, Fixes, Approvals
- AI Agent Ran git commit –no-verify? Make CI the Real Gate
- Stop Claude Code, Codex, Antigravity Asking for Permission
- Meta Muse Permissions: What to Allow for Email and Payments
Plans, models and context
- Codex “Ran Out of Room in the Model’s Context Window” Fix
- Claude Outage Fallback: Request Failed, Files Still Changed
- ChatGPT Pro 500 vs Pro 200 for Codex: Why I Went Back
- GPT-6 Sol vs Luna: When Luna Is Enough (60 Coding Runs)
When agents fail
- AI Agent Calling the Wrong Tool? I Cut It from 6/20 to 1/20
- RAG Retrieval vs Generation Failure: How I Tell Them Apart
- x402 Agent Spending Limits: The SDK Cap and the Rest
- Claude Code Custom Commands to Skills: Why /project: Fails
- Cursor Rules Not Working? Check the Glob and Description
- Claude Code Not Following CLAUDE.md? I Tested It 57 Times
AI apps
- Gemini Chats Not Loading or Missing? Check 3 Places First
- Is Gemini 4 Argon Free? Plans, API Price and Limits
All posts, newest first
GPT-6 Sol vs Luna: When Luna Is Enough (60 Coding Runs)
GPT-6 Luna costs 1/20 of Sol per token. In 60 runs on real coding tasks it was mostly enough for small edits but slipped on repo-wide work. Here are my escalation rules.
ChatGPT Pro 500 vs Pro 200 for Codex: Why I Went Back
Pro 500 vs Pro 200 for Codex after the October 30 cut: what each plan gives, how Ultrafast burns 8x usage, and why I moved back to Pro 200.
OpenAI Dots Permissions: What I Allowed, What Needs Approval
On my first OpenAI dot, reading stayed open, doc edits and sending needed approval, and payments and local Mac access stayed off. Here’s the split.
Meta Muse Permissions: What to Allow for Email and Payments
Which Meta Muse permissions to allow, which to approve one at a time, and what one-time cards and Sentinel can’t stop for your email and payments.
Gemini Chats Not Loading or Missing? Check 3 Places First
Gemini app not loading your chats? Check My Activity first. Some of mine were gone for good: the default 18-month auto-delete had removed them.
Stop Claude Code, Codex, Antigravity Asking for Permission
Switch Claude Code to acceptEdits, Codex to workspace-write with on-request. My 85-run test shows what still asks, and what slips past a deny rule.
Claude Pro vs Max: Why I Haven’t Upgraded for Claude Code
I hit Claude Pro limits in Claude Code, but waiting and Codex cover the gap. I’d move to Max at 3+ delayed days or over 10 hours of waiting a month.
Local LLM Hardware and Runtime Guide: My RTX 4090 Notes
Which GPU, Mac or runtime for a local LLM? A decision table and the numbers I measured on my RTX 4090 24GB, with links to each full write-up.
Local LLM Too Slow? 8B vs 30B on My RTX 4090
On my RTX 4090 (24GB), 7–8B models answer instantly; a 30B-class model took 20–30 s to start. Why big models slow down, how to check it, and what to pick.