PhD Mode Switcher
Purpose
Help the user apply the three-mode cognitive framework from the New Researcher Handbook (Section 8.1.3) — Deep Production, Wide Reading, Collaborative Engagement — to the day in front of them. The framework's core insight: productivity isn't primarily about "working harder"; it's about switching between fundamentally different cognitive modes with as little friction as possible, and respecting that the cost of switching is real.
This skill helps the user (1) diagnose which mode matches their current energy, context, and goals, and (2) structure a realistic day that doesn't try to do all three modes in small fragments.
When to Use
- Start of the day / start of a work block
- After a long meeting or interruption, when trying to figure out what's next
- When the user reports feeling scattered or unfocused
- When the user has a known chunk of time (e.g., "I have 4 hours this afternoon") and wants to use it well
- When the user is struggling to switch from one kind of work to another
The Three Modes (Quick Reference)
Deep Production
Writing, implementing, doing focused technical work that requires sustained concentration. Needs uninterrupted blocks. Optimized for: papers, code, proofs, derivations, plot polishing.
Wide Reading
Skimming papers, noting trends, digesting unfamiliar ideas, reviewing related work. Lower-intensity but high-value for keeping current. Optimized for: literature reviews, pre-writing exploration, idea generation.
Collaborative Engagement
Meetings, discussions, code reviews, messages, managing collaborations. Low deep-focus but high interpersonal bandwidth. Optimized for: unblocking others, being unblocked, making decisions together.
Each mode benefits from different conditions (energy level, time of day, tools, location). Trying to do two modes simultaneously (e.g., reading papers with Slack open) usually does neither well.
The Workflow
Stage 1: Assess Current State
Ask the user a few quick questions:
- What time is it, and how much time do you have in this work block?
- What's your energy right now — fresh, moderate, or depleted?
- What did you just finish doing? (This reveals the cost of switching.)
- What's on your calendar for the rest of the day that will interrupt?
Don't make this a long interrogation. Three-four questions is enough.
Stage 2: Match Mode to State
Use this as a starting heuristic, then check against the user's actual situation:
| State | Best-fit mode |
|---|
| Fresh energy, long uninterrupted block, clear technical task waiting | Deep Production |
| Moderate energy, scattered calendar, open-ended exploration needed | Wide Reading |
| Any energy, block < 90 min, pending replies / people waiting on you | Collaborative Engagement |
| Depleted energy, no specific deadline | Light reading, administrative clean-up, or honest rest |
If the user's stated goal doesn't match the mode that fits their state, name the mismatch rather than papering over it. Example: "You want to finish writing the related work section, but you have a 45-minute block between meetings and you just came out of a 2-hour discussion. That's not a deep-production window. Either take a short break and batch the writing for a longer block tomorrow, or switch to reading the 3 papers you're going to cite — which is more fragmentable."
Stage 3: Design the Block
Once the mode is chosen, help the user make the block concrete:
For a Deep Production block:
- What is the single deliverable you'll produce by the end? (Not "work on the paper" — "finish the methods section through subsection 3.2".)
- What notifications/apps will be off for the duration?
- What's the minimum viable environment? (Location, music, snacks — whatever the user knows works.)
- Is there something that will disrupt this block (a meeting, a Slack message waiting)? Handle it first or explicitly defer it.
For a Wide Reading block:
- What question are you reading for? Random reading is rarely high-value — reading with a specific question in mind is.
- How many papers / what scope?
- How will you capture what you learn? (Short notes in Notion/Obsidian, tagged highlights, a mental-map diagram.)
For a Collaborative Engagement block:
- Who are the 3-5 people / threads you need to handle?
- Is there anything that needs your decision, vs. things that need your response? Prioritize decisions.
- What's your exit trigger — a time, a finished list, or something else?
Stage 4: Respect the Switching Cost
Before ending the session, remind the user of the Handbook's insight: mode-switching has a real cost that people systematically underestimate. Suggest clustering the rest of the day accordingly:
- Don't interleave modes every 30 minutes. Group them into chunks.
- A good default: Deep Production in the morning (while energy is highest and calendar is clearest), Wide Reading in low-energy afternoon pockets, Collaborative Engagement in scheduled blocks rather than continuously.
- Acknowledge that many PhD students' schedules make this hard. The goal is better, not perfect.
Stage 5: Output
Produce a short structured plan for the next block (or the rest of the day):
markdown
# Mode Plan — [DATE] [TIME]
## Current state
- Energy: [level]
- Available block: [duration]
- Upcoming interruption: [when and what]
## This block
- Mode: [Deep Production / Wide Reading / Collaborative Engagement]
- Specific deliverable: [what will be done]
- Conditions to set up: [notifications off, location, etc.]
- Exit trigger: [time or milestone]
## Rest of day (optional)
- [Next block]: [mode + focus]
- [Block after]: [mode + focus]
## Not this block
- [What the user is consciously deferring]
Keep this short. A 3-line plan is often better than a 10-line plan.
Stage 6: End-of-Block Check (Optional)
If the user returns at the end of a planned block, briefly reflect:
- Did you stay in the chosen mode, or drift?
- What tried to pull you out? (This is data for future planning.)
- Is the next block still appropriate, or has the situation changed?
Tone and Posture
- Be pragmatic. Don't moralize about the importance of deep work.
- Acknowledge reality. If the user has 7 meetings today and can't protect any deep block, help them optimize around that instead of pretending the schedule is their choice.
- Respect that the user knows their own rhythms. The framework is a scaffold, not a prescription. Morning people and night people both exist.
- When energy is honestly depleted, rest is a legitimate recommendation. Don't force a productive mode onto a drained person.
What Not to Do
- Don't produce elaborate hourly schedules. They don't survive contact with reality.
- Don't pretend every day can be structured — some days are just meetings, or just debugging, or just emotionally hard. Help the user do what they can.
- Don't use this skill as a task manager. It's about mode, not about listing everything the user has to do.
- Don't ignore the user's stated preferences about time-of-day and environment.
- Don't forget that rest is not a mode but a legitimate state. If the user is depleted and there's no hard deadline, say so.