Course 08 Implementation Spec: The 10-Year Bet

1. Title and source files used

Course title: The 10-Year Bet

Owned output file: 08-ten-year-bet/08_ten_year_bet_course_content.md

Source files used as base truth:

• 08-ten-year-bet/curriculum.md
• 08-ten-year-bet/website-prompt.md

Purpose of this document: Translate the base curriculum into a facilitation-ready, assessment-ready, implementation-ready course-content specification for instructors, operators, and AI-assisted evaluators.

Source-of-truth commitments for this spec:

• Preserve the core philosophy from the curriculum: long-horizon identity work creates short-term urgency.
• Preserve the core outputs named in the curriculum: future-self vision, reverse-engineered plan, identity audit, 10-Year Bet document, and presentation.
• Preserve the website prompt's tone: editorial, precise, future-facing, manifesto-like rather than startup-hype.
• Treat the written 10-Year Bet as the central product and the annual revisit as part of the course's real value.
• Resolve the curriculum metadata inconsistency pragmatically:
• The metadata says "5 sessions x 90 minutes," but the body of the curriculum and the website prompt clearly define 6 live sessions.
• This implementation spec therefore uses 6 live sessions over 3 weeks as the operational model.

2. Design decisions at the top

Instructional design decisions

1. The course teaches thesis-building, not prediction.

Students are not being asked to guess the future correctly. They are being asked to develop a coherent, revisable directional thesis that improves present-day decisions.

1. Specificity is the main pedagogical lever.

The course only works if students are pushed out of vague aspiration and into concrete imagined reality, concrete milestones, concrete tradeoffs, and concrete next moves.

1. Identity is treated as a buildable system.

The curriculum's future-self work is implemented as capability design, relationship design, and artifact design, not just values journaling.

1. Reverse engineering matters more than inspiration.

The 2036 vision is only useful if it produces a 2026 action rule. Every live session must move from vision toward milestones, optionality, and near-term action.

1. The minimum viable path must be constrained.

Students should not create exhaustive life plans. They should identify the few highest-leverage moves that make the desired future more likely while preserving flexibility.

1. Identity investments vs. identity consumption is a recurring decision lens.

This distinction should appear in Session 1, become explicit in Session 4, and shape final document scoring.

1. The document is the product; the presentation is a pressure test.

Students should leave with a written artifact robust enough to revisit in one year without needing facilitator explanation.

1. Assessment rewards coherence, honesty, and actionability over ambition.

A modest but well-reasoned 10-year thesis should outscore an impressive-sounding fantasy with no pathway.

1. The annual ritual is part of implementation, not a sentimental add-on.

The course is incomplete unless the student has set up a real future revisit mechanism and named how the thesis may evolve.

1. AI evaluation should judge visible reasoning, not worldview alignment.

LLMs may evaluate structure, specificity, internal logic, and rubric fit, but should not reward prestige goals or punish unconventional paths if the thesis is coherent.

Facilitation design decisions

1. Facilitators should adopt a calm, exacting tone.

The course should feel consequential and thoughtful, not motivational or therapeutic.

1. Students must be interrupted when they stay abstract.

Useful redirect questions include:

• "What city?"
• "Doing what, exactly?"
• "With whom?"
• "What evidence would exist by then?"
• "What would have to become true in the next two years?"

1. The course must welcome multiple definitions of success.

Students may envision company-building, craft mastery, family leadership, public service, research, artistic work, or portfolio careers. The course grades quality of thesis design, not type of ambition.

1. The facilitator should normalize revision.

Students often freeze because they think the thesis must be perfect. Repeatedly state that the course values directional clarity plus adaptability.

1. Emotional exposure should be managed explicitly.

Students may discover that their current behavior does not match the future they claim to want. Facilitators should name that discomfort as productive rather than shameful.

Experience design decisions

1. Use the website prompt's editorial-manifesto aesthetic as a teaching motif.

Materials should feel like a personal doctrine document: serif headlines, mono years, clean margins, strong use of white space, and precise language.

1. Years should be visually salient throughout the course.

Use 2026, 2028, 2031, 2034, and 2036 as recurring anchors in slides, worksheets, and templates.

1. Exercises should alternate between imagination and reduction.

The course works best when expansive future vision is followed by narrowing, prioritization, and pruning.

3. Delivery model assumptions

Core format

• Duration: 3 weeks
• Live sessions: 6
• Session length: 90 minutes each
• Total live time: 9 hours
• Async work: 4 to 6 hours total
• Recommended cadence: 2 sessions per week

Target learner

• Age range: 15 to 25
• Backgrounds: mixed
• Prerequisites: none
• Prior work experience: not required

Recommended cohort model

• Ideal cohort size: 10 to 24 students
• Minimum viable cohort: 6 students
• Maximum without support facilitator: 28 students
• Recommended staffing for 20+ students: 1 lead facilitator plus 1 support facilitator or TA

Delivery assumptions

• The course may run in person, hybrid, or live online.
• Students should have access to a laptop, notebook, or typed document platform.
• The facilitator should use a visible timer for writing blocks and presentations.
• Students should have access to calendar tools or equivalent reminder systems for the annual revisit ritual.

Required course materials

• Slide deck
• 2036 day-in-the-life worksheet
• identity inventory template
• reverse-engineering timeline template
• identity investment vs. consumption audit worksheet
• 10-Year Bet document template
• presentation rubric
• accountability partner prompt or email template

Async workload assumptions

• Between Sessions 1 and 2: short reflection or notes capture
• Between Sessions 3 and 4: reverse-plan refinement
• Between Sessions 4 and 5: identity audit completion
• Between Sessions 5 and 6: final document revision and presentation prep

Accessibility and inclusion assumptions

• Students may not know what industry or career they want yet; the course should allow bets centered on problems, domains, lifestyles, or identities rather than job titles alone.
• Students may come from unstable or constrained environments; the course should allow multiple path shapes and should not assume family wealth, geographic mobility, or elite credentials.
• Students may write ambitiously; evaluators should score the logic and structure of the path rather than penalizing scale of aspiration.

Safety and support assumptions

• The facilitator should explicitly state that the course is not financial, medical, or legal planning.
• Students may anonymize names, employers, schools, and personal details in submissions.
• If a student's written work reveals severe hopelessness, self-harm ideation, coercion, or crisis, refer for human follow-up rather than relying on rubric scoring alone.

4. Detailed course content broken down by module, session, lesson, activity, timing, facilitator moves, and student outputs

Module 1 / Session 1: The Long View

Session goal: Establish why 10 years is the right unit for identity-level decisions and introduce the difference between a plan and a thesis.

Session outcomes:

• Students can explain why short-horizon thinking produces reactive choices.
• Students can describe the compounding logic behind long-term capability building.
• Students produce an initial 2036 direction statement.

Session 1 agenda

Facilitator notes for Session 1

• Do not let students hide inside generic success language like "happy," "successful," or "wealthy" without defining those terms.
• If students confuse a 10-year bet with an inflexible life script, repeat that the thesis is revisable.
• Use the phrase: "Long horizon does not mean low urgency. It means better urgency."

Student artifacts generated in Session 1

• Opening quickwrite
• Compounding worksheet
• Decision-horizon self-assessment
• 2036 one-sentence thesis seed

Module 2 / Session 2: Designing Your Future, Part I - The Day-in-the-Life

Session goal: Make the future self concrete enough that it can later be reverse-engineered.

Session outcomes:

• Students produce a vivid day-in-the-life narrative for 2036.
• Students move from abstract goals to situational detail.
• Students identify the emotional texture of the future they want.

Session 2 agenda

Facilitator notes for Session 2

• Students often write luxury signals instead of meaningful identity markers. Redirect from consumption symbols toward work, relationships, and contribution.
• If a student's vision is too safe, ask what meaningful stretch it implies.
• If a student's vision is grandiose but empty, ask what they actually do between 8 a.m. and 11 a.m.

Student artifacts generated in Session 2

• Day-in-the-life draft
• Peer question list
• Authenticity note

Module 3 / Session 3: Designing Your Future, Part II - Identity Inventory and Reverse Engineering

Session goal: Convert the future scene into capability, relationship, and milestone logic from 2036 back to 2026.

Session outcomes:

• Students identify what their future self knows, has built, is known for, and refuses.
• Students construct a backwards timeline from 2036 to now.
• Students begin distinguishing foundational moves from later-stage outcomes.

Session 3 agenda

Facilitator notes for Session 3

• Do not allow milestone inflation. If everything happens in years 1 to 2, the reverse plan is under-thought.
• Push students to separate status labels from earned evidence. "Founder" is not a milestone; "shipped three products in this domain" is.
• Treat "what I refuse to do" as a serious identity constraint, not a cosmetic values statement.

Student artifacts generated in Session 3

• Identity inventory
• Reverse plan timeline draft
• Assumptions list
• Starting-now commitment

Module 4 / Session 4: The Minimum Viable Path and the Identity Audit

Session goal: Help students identify the few highest-leverage moves on the path and distinguish identity investments from identity consumption.

Session outcomes:

• Students can define a minimum viable path.
• Students can audit recent behavior as investment vs. consumption.
• Students can reduce an overbuilt plan into a smaller set of compounding moves.

Session 4 agenda

Facilitator notes for Session 4

• Many students will label all learning as investment. Push on whether it changed capability or produced artifacts.
• Many students will over-design the path. Force pruning. The minimum viable path should feel slightly uncomfortable in its simplicity.
• Use repeated contrast pairs:
• investment vs. image
• optionality vs. noise
• compounding vs. stimulation

Student artifacts generated in Session 4

• MVP path draft
• Identity audit worksheet
• Audit synthesis

Module 5 / Session 5: Writing the 10-Year Bet

Session goal: Synthesize prior work into a formal 10-Year Bet document that is complete, readable, and revisitable.

Session outcomes:

• Students draft the full 10-Year Bet document.
• Students integrate future-self vision, reverse plan, assumptions, contingency, and minimum viable path.
• Students prepare a document worthy of annual review.

Session 5 agenda

Facilitator notes for Session 5

• The best documents sound self-authored, not like polished corporate strategy decks.
• If the thesis reads as imitation of someone else's life, ask: "What in this document is unmistakably yours?"
• Require at least one explicit contingency statement. The curriculum makes adaptation part of the thesis.

Student artifacts generated in Session 5

• Full 10-Year Bet draft
• Peer review sheet
• Presentation outline

Module 6 / Session 6: Presentations and Commitment

Session goal: Test clarity, conviction, and present-day translation; close the course with a concrete annual revisit ritual.

Session outcomes:

• Students present their 10-Year Bet clearly and concisely.
• Students receive peer and facilitator challenge.
• Students complete the accountability and revisit setup.

Session 6 agenda

Facilitator notes for Session 6

• Keep presentation feedback centered on clarity, coherence, and actionability.
• Do not reward charisma over substance.
• Make the annual ritual feel serious. The course's long-term effect depends on this step being real.

Student artifacts generated in Session 6

• Final presentation
• Revision reflection
• Revisit reminder confirmation
• Future-self note

Recommended async checkpoints across the course

Between Session 1 and Session 2

• Capture 3 moments when current behavior aligns or conflicts with the 2036 thesis seed.

Between Session 2 and Session 3

• Re-read the day-in-the-life and underline implicit skills, relationships, and artifacts.

Between Session 3 and Session 4

• Refine the reverse plan so each time band has evidence-oriented milestones rather than titles or vibes.

Between Session 4 and Session 5

• Complete the identity audit with honest examples from the last 6 months.

Between Session 5 and Session 6

• Finalize the document and rehearse the 3-minute version aloud at least once.

5. Assignments and artifacts

Assignment 1: Future Self Exercise

Weight: 20%

When: Sessions 2 and 3

Purpose: Create a vivid, specific future identity that can be reverse-engineered.

Submission format: 700 to 1,000 words total across the day-in-the-life and identity inventory

Required artifact:

• 2036 day-in-the-life narrative
• explicit setting and environment
• named work/problem focus
• relationships or key people described concretely
• identity inventory covering knowledge, network, built artifacts, reputation, and refusals
• short authenticity note on what feels real vs. borrowed

Assignment 2: Reverse Plan

Weight: 25%

When: Draft in Session 3, revised by Session 5

Purpose: Translate the future vision into a backwards logic chain from 2036 to now.

Submission format: Timeline, structured table, or memo equivalent

Required artifact:

• milestones for Year 8-10, Year 5-7, Year 3-4, Year 1-2, and now
• capability needs for each stage
• relationship or network needs for each stage
• evidence of progress for each stage
• at least 2 explicit assumptions or "what I'm betting against" statements
• one starting-now move

Assignment 3: Identity Audit

Weight: 15%

When: Session 4

Purpose: Diagnose current behavior as identity investment vs. identity consumption.

Submission format: Worksheet plus 150 to 250 word synthesis

Required artifact:

• audit of recent time, effort, money, or attention across at least 8 examples
• classification of each example as investment, consumption, or mixed
• short rationale for each classification
• one pattern to increase
• one pattern to stop or reduce
• synthesis paragraph on current ratio and implications

Assignment 4: 10-Year Bet Document

Weight: 25%

When: Draft in Session 5, final by Session 6

Purpose: Produce the course's core written thesis document.

Submission format: 1,200 to 2,000 words or equivalent polished memo

Required artifact:

• title and version/date
• the world in 2036
• who I am in 2036
• reverse plan
• what I'm betting against
• minimum viable path
• what I'd do differently if a key assumption is wrong
• revisit date one year out

Assignment 5: Final Presentation

Weight: 15%

When: Session 6

Purpose: Test whether the thesis is clear enough to communicate and robust enough to defend.

Submission format: 3-minute presentation plus 1-minute Q&A

Required artifact:

• who I am in 2036
• minimum viable path
• one crucial assumption
• what starts this week
• proof of annual revisit setup

Ungraded but required implementation artifacts

• one-sentence thesis seed
• starting-now commitment
• peer review sheets
• future-self note to be opened at annual revisit

6. AI/LLM grading and assessment framework

Assessment model

The LLM should function as a structured first-pass evaluator. It should check completion, score observable rubric dimensions, extract evidence, and generate targeted feedback. Human review should remain in the loop for ambiguous, emotionally sensitive, or unusually inconsistent submissions.

Core grading principles

1. Grade internal coherence, not prestige.

A thesis about becoming a respected local craftsperson can score as highly as a thesis about founding a billion-dollar company if the reasoning is stronger.

1. Reward specificity over intensity.

Vivid, grounded detail should score above dramatic but generic ambition.

1. Reward present-day translation.

A strong future vision must generate a credible "start now" implication.

1. Reward explicit assumptions and contingency.

Students should receive credit for naming what must be true and what they would change if wrong.

1. Do not confuse polished writing with strategic quality.

Writing quality matters only insofar as it affects clarity and evaluability.

1. Penalize fantasy without pathway.

If the student names a destination but provides no believable capability, relationship, artifact, or milestone logic, scores should stay low.

1. Do not over-penalize ambition.

The issue is not scale. The issue is unsupported scale.

1. Treat honesty as evidence of maturity.

If a student plainly admits uncertainty, mismatch, or fear, that should improve self-awareness scores when paired with analysis.

What the LLM is well suited to evaluate

• presence or absence of required sections
• specificity of future scene
• completeness of identity inventory
• logical flow from future vision to reverse plan
• presence of assumptions and contingencies
• evidence that the minimum viable path is selective rather than exhaustive
• quality of identity investment vs. consumption distinctions
• clarity and actionability of immediate next steps

What the LLM should not pretend to know

• whether the student's life plan will actually work
• whether a personal aspiration is objectively realistic given facts not present in the submission
• whether the student's preferred industry or geography will evolve exactly as they describe
• whether prestige-oriented markers are inherently more valuable than quieter forms of success

Concrete assessment heuristics for LLM evaluation

Heuristic 1: Future-scene concreteness

• Full credit if the day-in-the-life contains specific environment, work, people, routines, and pride markers.
• Partial credit if the scene is somewhat vivid but still generalized.
• Low credit if the scene reads like a motivational slogan or aesthetic mood board.

Heuristic 2: Identity delta clarity

• Full credit if the student clearly shows what the 2036 self knows, has built, knows whom, and refuses.
• Partial credit if some dimensions are clear but others are thin.
• Low credit if the inventory mostly lists labels or outcomes without underlying capability.

Heuristic 3: Reverse-chain logic

• Full credit if the reverse plan makes each time band meaningfully dependent on the previous one.
• Partial credit if the plan has plausible stages but weak causal connection.
• Low credit if the timeline is just a stack of titles or achievements.

Heuristic 4: Minimum viable path quality

• Full credit if the student identifies a small number of leverage moves that create directional pressure and optionality.
• Partial credit if the moves are sensible but too numerous or unprioritized.
• Low credit if the path is bloated, prestige-driven, or disconnected from the future thesis.

Heuristic 5: Investment vs. consumption judgment

• Full credit if the audit shows honest classification with real examples and nuanced "mixed" cases.
• Partial credit if the audit is mostly correct but under-argued.
• Low credit if all activities are labeled investments or if the student avoids self-critique.

Heuristic 6: Assumption awareness

• Full credit if the student names key assumptions and shows how at least one wrong assumption would change the plan.
• Partial credit if assumptions are named but not operationalized.
• Low credit if the thesis reads as certainty with no exposure to being wrong.

Heuristic 7: Present-day actionability

• Full credit if the thesis names a start-now move that is concrete, near-term, and aligned.
• Partial credit if the next action exists but is generic.
• Low credit if the thesis ends in inspiration without action.

Heuristic 8: Revisit readiness

• Full credit if the document feels usable one year later and includes a clear revisit date plus comparison logic.
• Partial credit if the ritual is mentioned but not integrated.
• Low credit if the annual revisit appears as an afterthought.

Artifact-specific LLM evidence checks

Future Self Exercise

• Check for concrete setting, work, relationships, routines, and identity markers.
• Check whether the future scene implies capabilities and artifacts rather than status alone.
• Flag submissions that are entirely aesthetic, fantasy-driven, or empty of actual work.

Reverse Plan

• Check that every time band is populated.
• Check for cause-and-effect logic across stages.
• Check for at least one starting-now move and at least two assumptions.
• Flag plans where early years are vague but later years are overconfident.

Identity Audit

• Check that at least 8 recent examples are classified.
• Check whether rationales distinguish compounding from comforting activity.
• Flag audits where no consumption is admitted or where examples are too abstract to judge.

10-Year Bet Document

• Check for all required sections.
• Check whether sections reinforce one another rather than contradict one another.
• Check whether the contingency section changes behavior meaningfully if the assumption is wrong.
• Flag documents that are polished but structurally hollow.

Final Presentation

• Check whether the student can state the thesis, MVP path, key assumption, and weekly action clearly.
• Check whether presentation claims match the written document.
• Flag presentations that are emotionally persuasive but evidentially thin.

Suggested scoring mechanics

• Use a 1 to 4 scale for each criterion:
• 4 = strong and specific
• 3 = competent and mostly complete
• 2 = partial, vague, or uneven
• 1 = missing, generic, or misunderstood
• Convert criterion scores to weighted artifact scores.
• Provide evidence-based notes before overall judgment.

Automatic LLM flags requiring human review

• Severe hopelessness, self-harm ideation, or crisis signals in future-planning language
• Submission is so generic that authorship or authenticity is unclear
• Internal contradictions severe enough that the intended thesis cannot be determined
• Repeated claims of certainty with no assumptions or contingency despite rubric prompts
• Strongly personalized content that may require pastoral, counseling, or safeguarding follow-up

Recommended LLM output structure

For each artifact, the LLM should return:

1. completion check
2. rubric scores by criterion
3. 2 to 4 evidence-based observations
4. 2 concrete revision moves
5. overall level: Exemplary / Strong / Competent / Emerging / Incomplete
6. human-review flags, if any

7. Rubrics, scoring criteria, and evaluator prompt guidance

Recommended course-weighted rubric

Aligned to the base curriculum assessment framework:

A. Future Self Exercise Rubric

B. Reverse Plan Rubric

C. Identity Audit Rubric

D. 10-Year Bet Document Rubric

E. Final Presentation Rubric

Performance bands

Evaluator prompt guidance for LLM use

Recommended system prompt

You are evaluating student work for Course 08, "The 10-Year Bet."

Use only the student's submission and the rubric provided.
Do not invent missing evidence.
Do not reward prestige, confidence, or dramatic ambition unless the submission provides specific reasoning and pathway logic.
Reward specificity, internal coherence, explicit assumptions, honest self-awareness, and clear present-day implications.
It is acceptable for a student's future to be uncertain or unconventional if the thesis is well structured.
If the work signals crisis, self-harm, or serious safeguarding concerns, flag it for human review.
```

#### Recommended evaluator user prompt template

Evaluate the following student submission for Course 08, "The 10-Year Bet."

Apply this rubric: [paste relevant rubric section]

Required output format:

1. Completeness check
2. Criterion-by-criterion scores with one-sentence justification each
3. Total score
4. Three strengths
5. Three improvement priorities
6. One paragraph of student-facing feedback
7. Human-review flags, if any

Student submission: [paste submission]

#### Recommended structured output schema

```json
{
  "completeness": {
    "status": "complete | partial | incomplete",
    "missing_items": []
  },
  "scores": {
    "criterion_name": {
      "points_awarded": 0,
      "points_possible": 0,
      "rationale": ""
    }
  },
  "total_score": {
    "points_awarded": 0,
    "points_possible": 100
  },
  "strengths": [],
  "improvement_priorities": [],
  "student_feedback": "",
  "human_review_flags": []
}

Additional evaluator heuristics

• If the day-in-the-life contains almost no concrete verbs, the future-self specificity score should not exceed competent.
• If the reverse plan mostly names titles, credentials, or outcomes without capability milestones, milestone quality should score 2 or below.
• If the identity audit classifies nearly everything as investment, honesty and classification judgment should be capped unless the rationale is unusually strong.
• If the 10-Year Bet document includes no contingency section that changes behavior, contingency thinking should score 2 or below.
• If the final presentation cannot identify a this-week action, present-day actionability should score 1 or 2.

Prompting guardrails for evaluator consistency

• Instruct the LLM to cite or paraphrase the student's actual language when possible.
• Instruct the LLM to distinguish "missing evidence" from "bad logic."
• Instruct the LLM not to punish a student for revising the thesis across drafts if the reflection is explicit.
• Instruct the LLM to keep tone direct, specific, and developmental.
• Instruct the LLM to recommend revisions the student can complete within one revision cycle.

8. Feedback strategy: what strong/average/weak responses look like and how an LLM should respond

Global feedback posture

The LLM should sound precise, unsentimental, and useful. It should not flatter students for having dreams. It should identify exactly what makes the thesis strong or weak, then direct the next revision step.

What strong responses usually look like

• The future scene is vivid without becoming ornamental.
• The student can explain what they do, not just what title they hold.
• The reverse plan has clear stages with believable capability buildup.
• The minimum viable path is selective and high-leverage.
• The identity audit contains honest admissions about current drift or consumption.
• The document names at least one assumption that could fail and shows what would change.
• The thesis creates a real this-week action.

How the LLM should respond to strong work

• Name the strongest strategic move specifically.
• Show the student why the thesis feels credible.
• Push for one layer of sharper assumption testing or milestone evidence rather than broad rewrite.

Example LLM feedback pattern for strong work

• "Your strongest move is that your 2036 identity is defined by work, capability, and relationships rather than status labels."
• "The reverse plan works because each stage depends on skills and evidence built in the previous stage."
• "To strengthen this further, tighten the contingency section by naming what would make you change domains rather than just timing."

What average responses usually look like

• The student has the right sections but not enough depth.
• The future scene contains some detail but still leans on general success language.
• The reverse plan is plausible but has jumps.
• The audit identifies consumption but does not yet expose the real pattern.
• The MVP path is sensible but still too crowded.
• The final action is present but generic.

How the LLM should respond to average work

• Confirm what is already structurally working.
• Identify the single missing layer that would improve the submission most.
• Convert vague ambition into one revision question at a time.

Example LLM feedback pattern for average work

• "You have a workable thesis and a recognizable path, but the middle years need more causal detail."
• "The main thing limiting this document is specificity: several milestones are titles or outcomes rather than evidence of capability."
• "Revise by replacing two abstract milestones with proof-of-work milestones you could actually observe."

What weak responses usually look like

• The future reads like fantasy, aesthetic taste, or social-media aspiration.
• The student describes what they want to have, not what they want to do or become.
• The reverse plan is empty or prestige-driven.
• The identity audit avoids real self-critique.
• The contingency section is absent or meaningless.
• The document ends in inspiration rather than behavior.

How the LLM should respond to weak work

• State clearly that the submission is too vague or unsupported to score highly.
• Point directly to missing required components.
• Give a rescue path in sequence: make the future concrete, rebuild the reverse logic, then name one action.

Example LLM feedback pattern for weak work

• "Right now the bet sounds ambitious, but it is not yet a thesis because the pathway logic is missing."
• "Your next revision should begin by rewriting the 2036 day-in-the-life with real work, real people, and one hard problem you are solving."
• "After that, rebuild the reverse plan around capabilities and evidence rather than titles."

Artifact-specific feedback guidance

Future Self Exercise

Strong: The future is vivid, situated, and reveals who the student has become through capabilities, relationships, and artifacts.

Average: The future is partially concrete but still contains borrowed language or broad status markers.

Weak: The future is mostly aesthetic, emotional, or title-based, with little evidence of lived identity.

LLM response pattern:

• ask what work fills the day
• ask what evidence exists in the world by 2036
• ask what part feels truly theirs versus culturally inherited

Reverse Plan

Strong: Each time band explains what must be in place for the next stage to become possible.

Average: The path is plausible but contains jumps, crowded timelines, or missing assumptions.

Weak: The plan is a wish list, not a sequence.

LLM response pattern:

• challenge the biggest leap
• ask what capability is missing from the current version
• ask what would have to be true by the end of years 1 to 2

Identity Audit

Strong: The student names uncomfortable truths and can distinguish compounding activity from soothing activity.

Average: The student sees the pattern but softens it.

Weak: The student protects self-image and avoids real examples.

LLM response pattern:

• identify one classification that seems honest and useful
• ask for one stronger example of identity consumption
• ask what current behavior most conflicts with the 2036 thesis

10-Year Bet Document

Strong: The document reads like a coherent operating thesis the student can revisit and revise.

Average: The document is complete but unevenly integrated.

Weak: The sections are present but disconnected, generic, or hollow.

LLM response pattern:

• point to the strongest section
• identify the section where the logic breaks
• recommend one revision that increases cross-section coherence

Final Presentation

Strong: The student can summarize the thesis clearly and defend the immediate implications.

Average: The student communicates the gist but struggles under follow-up.

Weak: The presentation is abstract, overly polished, or disconnected from the written thesis.

LLM response pattern:

• ask for the exact weekly action sentence
• test whether the key assumption truly matters
• check whether the spoken version matches the written document

Recommended LLM response style by performance level

Sentence starters the LLM can use productively

• "The most credible part of your thesis is..."
• "The biggest leap in your current logic is..."
• "This becomes a stronger 10-year bet when you..."
• "The most important missing evidence about your future self is..."
• "Your reverse plan gets more believable if..."
• "The current audit suggests you are still optimizing for..."
• "The next move that best matches your thesis is..."

Final evaluator instruction for feedback generation

When generating feedback, the LLM should:

1. State the current performance level plainly.
2. Cite one concrete strength with evidence.
3. Cite one concrete weakness with evidence.
4. Give the highest-leverage revision step.
5. End with one forward action the student can take now.

Implementation notes for operators

• Preserve the seriousness of the written artifact. Students should leave with a document they would still understand and care about on April 20, 2027.
• If the course is run online, require a shared template so students do not lose structure.
• If AI grading is used, keep humans in the loop for presentation scoring, emotional-risk flags, and unusually ambitious submissions that may be structurally sound but easy to misread.
• The course succeeds when students make better present decisions because they have a thesis, not when they produce the most impressive-sounding future.