Campbell Duncan

building

[Campbell Duncan @camsduncan]

@thsottiaux Merge with the OS.

[Campbell Duncan @camsduncan]

@DavidSojevic Nice! Do you plan on launching something soon? Im building the same thing to solve a personal frustration with current aggregators, but haven't found the time to build it out fully

[Campbell Duncan @camsduncan]

@ideacasino How long to get App Store approval?

[Fernando Borretti @zetalyrae]

I wrote about how accumulating capital won't save you from being disempowered by superintelligent AI.

[Campbell Duncan @camsduncan]

@ankurnagpal sounds like you're doing wash sales? which is illegal

[Campbell Duncan @camsduncan]

@jordanhknight_ why spread fake news?

[Campbell Duncan @camsduncan]

@kushalbyatnal credit pricing is ambiguous. you should price on # of pages.

[Campbell Duncan @camsduncan]

@michaelfrazis dilution? or feeling that it's toppy?

[Campbell Duncan @camsduncan]

@ArrowInvestor incentivises leveraging up to your eyeballs on ppor
nothing else makes financial sense now

[Campbell Duncan @camsduncan]

Or use the skill: https://t.co/ewvDSHBqDV

[Campbell Duncan @camsduncan]

Claude version:

Look back over my recent work from the last 30 days, or all available history if shorter, and identify repeated manual workflows worth packaging.

Use available evidence in this order:
- Recent Claude Code session transcripts and their summaries (e.g. ~/.claude/projects/, and the history surfaced by `claude --resume`).
- CLAUDE.md memory files (project and user) to find patterns and conventions repeated across sessions.
- Shell history, `git log` / `git reflog`, and commit messages to spot repeated work outside Claude Code. Use these for discovery only; confirm important details in the relevant source system or repo when possible.
- Existing skills (.claude/skills/), subagents (.claude/agents/), slash commands, hooks, scheduled tasks, and configured MCP servers, so you reuse or extend what already exists instead of duplicating it.

Look broadly for work that is repeated, time-consuming, error-prone, context-heavy, or benefits from a consistent process. Include workflows across coding, research, writing, planning, communication, operations, analysis, and personal administration.

Only act on a candidate when it:
- occurred at least twice, or is clearly likely to recur and costly to repeat;
- has stable inputs, a repeatable procedure, and a clear output or stopping condition;
- would materially improve speed, quality, consistency, or reliability;
- is not already adequately covered.

Choose the smallest appropriate form:
- Skill: a reusable workflow or playbook as a SKILL.md (auto-invoked when relevant, or run explicitly as /command).
- Subagent: a bounded specialist role or investigation task suitable for delegation in an isolated context (.claude/agents/).
- Automation: a hook for event-driven checks tied to tool/lifecycle events, or a scheduled headless `claude -p` job (via cron or GitHub Actions) for recurring reports, reminders, or monitors.
- Skip: work that is too one-off, ambiguous, sensitive, or poorly evidenced to package.

First produce a compact shortlist with:
- repeated workflow
- supporting evidence and dates
- frequency/confidence
- recommended form: skill, subagent, automation, extend existing, or skip
- why it is or is not worth creating

Then create only the high-confidence missing items. Keep them narrow, practical, source-aware, and easy to validate. Do not create speculative, overlapping, or overly broad assets.

Finish with:
- what you created or extended
- what you deliberately skipped
- what needs more evidence before packaging

[Greg Brockman @gdb]

self improvement prompt for codex

[Campbell Duncan @camsduncan]

@michaelfrazis Could you cover $IREN in the next pod?

[Campbell Duncan @camsduncan]

@chalaska I'm building https://t.co/Lb04jSB6xF

[Campbell Duncan @camsduncan]

@LimitingThe You covered Talga in some of your earlier videos, and mentioned being invested in them. Are you still?

[Peter McCrory @PeterMcCrory]

I'm in Australia this week, and we're releasing a new country report from the Anthropic Economic Index. Here's what we found about how Australians use Claude. 🧵

[Campbell Duncan @camsduncan]

@noahzweben Are you going to support browser agents in the cloud to unlock e2e testing? It's stopping me from using the cloud environment.

[Riley Walz @rtwlz]

made my computer dramatically play BBC news music before every meeting

[Campbell Duncan @camsduncan]

@noahzweben Can you get your local skills working on cloud sessions?