CHAPTERS
- 0:00 – 1:01
Cold open: time scarcity, automation test, and “Claude learns by doing”
Hilary previews why her time is more precious than ever and shares the key heuristic she uses to decide what to automate. Claire frames automation as starting from a clear problem statement, not heavyweight engineering.
- •New-mom attention fragmentation drives urgency to reclaim time
- •10x impact test: if being 10x better wouldn’t matter, automate it
- •Automation can start with a problem statement rather than code
- •Claude improves iteratively by observing real behavior
- •Goal is reducing system maintenance cost to near-zero
- 1:01 – 1:32
Show setup: Hilary’s return + the “anti-system system” promise
Claire reintroduces Hilary as a repeat guest with new constraints: entrepreneurship and motherhood. They tee up a practical, low-maintenance approach to using Claude Code for personal operations.
- •Hilary’s life changed: entrepreneur + new baby
- •Upgrading from custom GPTs to Claude Code workflows
- •Focus on day management: to-dos, calendar, and execution
- •Positioning: powerful automation without high setup overhead
- 1:32 – 3:03
Sponsor break: WorkOS and enterprise-ready AI apps
Claire explains why AI products face enterprise security scrutiny and how WorkOS provides drop-in building blocks. The segment emphasizes auth, access controls, and auditability as table stakes.
- •AI tools need deep access to company systems to be useful
- •Enterprise buyers require strong security controls and auditability
- •Building enterprise features from scratch is costly
- •WorkOS as “Stripe for enterprise features”
- •Examples of companies using WorkOS
- 3:03 – 4:14
Life update: why opportunity cost is highest—and attention is most fractured
Hilary describes the reality of higher time value paired with more logistics and less sleep. The goal isn’t hyper-optimization; it’s reclaiming time for meaningful work and life enjoyment.
- •Opportunity cost spikes with parenting + running a business
- •More life admin/logistics than before; historically not her strength
- •Less sleep and more cognitive load increase drop rate
- •Automation as “reclaim time” for work she loves and leisure
- •“Two wolves”: work and picnic; neither is life admin
- 4:14 – 8:07
The anti-system system: complexity must earn its keep
They contrast “Notion-heavy” systems with lightweight approaches. Hilary argues most AI productivity demos fail because they require too much setup and ongoing maintenance.
- •Many AI systems require upfront configuration and ongoing organization
- •Hilary’s bar: minimal setup, minimal maintenance
- •Claire’s framing: spectrum from structured Notion to chaotic desktop
- •Principle: start simple, avoid tools that create new work
- 8:07 – 9:31
Al Gore’s chaotic desktop: permission to be messy—and still effective
Hilary shares a memorable photo of Al Gore’s cluttered desktop as a metaphor for real-life cognitive clutter. The takeaway: you can build effective support systems without appearing perfectly organized internally.
- •Anecdote: Al Gore’s desktop as an honest look behind the scenes
- •Feeling “seen” by visible messiness in high performers
- •Reframe: chaos is common; systems should work with it, not fight it
- •New organizing spectrum: “Notion boy to Al Gore desktop”
- 9:31 – 11:46
Fast task capture with iPhone Back Tap shortcuts (no AI required)
Hilary shows her lowest-friction input method: double-tap the back of the phone to dictate a task into an inbox. This prevents ideas and obligations from slipping through a “leaky sieve” brain.
- •Lock-screen capture via Back Tap + dictation box
- •Tasks like “reschedule pediatrician appointment” captured instantly
- •Setup basics: Shortcuts app + Accessibility → Touch → Back Tap
- •Philosophy: capture first, optimize later
- •Sharing setup steps via a reference guide/Figma file
- 11:46 – 16:17
Claude Code basics + “Plan My Day” command: day planning from the terminal
Hilary demonstrates opening Claude Code and running a simple “Plan My Day” prompt. Claude pulls reminders and organizes them into a markdown file, operating inside a dedicated folder of working context.
- •Quick Claude Code install/run flow (terminal, install line, start)
- •Single command: “Plan My Day” triggers a reusable instruction set
- •Claude reads Reminders and maintains a categorized markdown task doc
- •Context lives in a dedicated folder of markdown files
- •Hilary uses Obsidian for writing, but rarely opens the system files
- 16:17 – 19:34
Preference learning + calendar scheduling: make the plan match reality
Claude uses observed preferences (pumping, childcare constraints, energy patterns) to propose a schedule. Hilary emphasizes preferences should reflect actual behavior, not aspirational plans.
- •Claude factors constraints: pumping, feeding, childcare coverage
- •Preferences are learned by observation over time, not manual setup
- •Avoid “fantasy preferences” (e.g., daily 60-minute walk that never happens)
- •Claude pulls calendar and proposes an actionable daily plan
- •Calendar as the execution surface: commitments made visible
- 19:34 – 23:22
Beating procrastination: break big tasks into the first tiny step
Hilary explains why overdue life admin persists: tasks are too big to fit into the margins of the day. Claude helps by identifying the smallest next action and timeboxing it into an available slot.
- •Procrastination driver: tasks feel too large and multi-stage
- •Claude reframes: “You’re not doing the passport, just booking the appointment”
- •Tiny steps fit into 10–15 minute gaps and build momentum
- •Calendar timeboxing reduces avoidance and cognitive load
- •Benefit: progress without overwhelming “all-at-once” execution
- 23:22 – 30:50
The yappers API + daily logs: lightweight observability and feedback loops
Instead of wiring every integration, Hilary “talks” to Claude as she works and lets it log what happened. Claude produces daily notes with schedules, activity logs, and pattern detection to improve future planning.
- •“Yappers API”: narrate what you’re doing instead of building integrations
- •Claude creates daily markdown notes with schedule + activity log
- •End-of-day reflection prompts exist but can be optional
- •Pattern mining: where time goes vs. stated priorities
- •Iterate by doing: daily feedback improves the system without heavy maintenance
- 30:50 – 32:49
Screenshots as a universal interface: low-friction context sharing at work and home
Claire and Hilary advocate screenshots as a practical bridge when permissions or integrations are hard. The approach enables quick proof-of-concepts before investing in more complex connections.
- •Screenshots can replace deep integrations for early workflows
- •“Held together with screenshots and prayers” can still prove value
- •Use cases: calendar snapshots, notes from appointments, on-screen context
- •Work constraint workaround: prove value before asking for data access
- •Principle reinforced: complexity must earn its keep over time
- 32:49 – 39:16
Deciding what to automate: the 10x impact framework (and career-stage nuance)
Hilary introduces her decision rule: automate tasks where getting 10x better wouldn’t yield 10x results, and invest in tasks where improvement compounds. They add an important caveat: what to automate depends on where someone is on a learning curve.
- •Core test: would being 10x better create 10x impact?
- •Automate low-leverage busywork; protect high-leverage thinking/creative work
- •Extend to life: protect joy, not just productivity
- •Within-task decomposition: keep “uniquely you” parts, automate the rest
- •Learning curve caveat for managers: juniors may need reps on ‘busy’ tasks
- 39:16 – 48:36
“Recording on” anonymization mode + building a returns-tracking skill from scratch
Hilary shows how she demos real workflows without doxxing herself by toggling an anonymization skill. She then illustrates her first-principles method for creating new Claude Code skills (example: returns tracking) by starting with frustration, clarifying questions, and incremental integration.
- •Recording mode: automatically anonymize names/identifiers while demoing
- •Avoids maintaining separate “demo” environments (reduces double complexity)
- •Product insight: AI changes the altitude of problem-solving (new paradigms)
- •Skill-building workflow: describe the pain, let Claude ask a few questions, then build
- •Returns skill example: email access, return deadlines, and retailer return instructions encoded into a reusable slash command
- 48:36 – 51:43
Adoption advice + closing: build the muscle memory to reach for AI
Hilary’s final guidance is habit-based: do one thing with Claude Code daily until it becomes second nature. The episode closes with where to find her work and resources.
- •Practice loop: “one thing every day” to rewire reflexes
- •From terminal-averse to terminal-native via repeated exposure
- •Claude Code’s power grows once the interface fear fades
- •Hilary’s home base: Substack (hills.substack.com) + manager AI course/events
- •Show wrap: like/subscribe, podcast platforms, and episode hub
