Reading up on Opus
7 deep · digging since jan 29
- Fable's judgement
The author learned to let Claude Code's Fable model use its own judgement to delegate coding tasks to lower-power models, improving efficiency and token usage.
- My Thoughts on AI, Part 2: Agent Setup, Workflow, and Tools
The author details his personal AI development setup and workflow using OpenCode agent, CodeNomad GUI, Opus 4.6 model, and a structured orchestrator-subtask methodology.
- Cybersecurity looks like proof of work now
AISI research shows spending more tokens on LLM vulnerability scanning yields no diminishing returns, making cybersecurity a proof-of-work arms race.
- Why I believe in SOTA models over custom ones
The future of AI is powerful general models becoming cheaper, not narrow custom models, because specialized tasks still benefit from broad intelligence.
Takes
the killer AI coding combo for me:opus 4.6 w/ max 20x ($200/mo) — primary workhorsegpt 5.3 codex w/ plus ($20/mo) — review pipeline + weird bugsall via CLI and @conductor_build
@Shpigford
here's some tips for prompting Opus 4.6 + taking advantage of some new API features (e.g., adaptive thinking, 1M context, context compaction, + more). https://t.co/5Lucl0g73n
@RLanceMartin
Since most of my work now involves reviewing plans in Markdown files, I had Opus create a native macOS Markdown viewer for me. Features include: table of contents, file auto-reload, scroll position persistence, and native macOS tabs.I'm too busy to release this officially… pic.twitter.com/6RPdiGV4Fs
@frankdilo