Level 3 · Advanced · 6 modules

Acceleration

Foundations works beautifully on a clean eight-week skill. Real professional learning
is not clean. It runs for months, has no textbook, competes with three other
priorities, plateaus without warning, and interferes with the skill you learned last
quarter. Acceleration is the system built for that.

The graduate problem is not “how do I learn?”
It is “how do I run three of these at once, for a year, without the wheels coming off?”

Start Module 3.1
See the AI stack

Six modules · six artifacts · typically one quarter · requires the Foundations badge.

The promise

You will run a 90-day learning campaign, on messy real-world material, alongside two other tracks, without losing any of them.

And when it stalls — it will stall — you will diagnose which stage failed instead of
assuming you got lazy.

Foundations gave you a complete system for one bounded skill with available material.
Acceleration handles the four things that break that system in professional life:

Duration. A nine-month skill fails differently from an eight-week one. Motivation is not the mechanism; milestone architecture is.
Absence of curriculum. Most of what you need to learn at work has no course. The raw material is projects, failures, calls, and feedback — and it has to be converted into encoding events deliberately or it evaporates.
Volume. Twelve sources, four of which contradict each other, and a synthesis you have to hold in your head during a meeting.
Concurrency. Two or three tracks at once, each with spacing obligations, all competing for the same four hours a week.

You will also build the thing that makes all of it tractable: a personal AI learning
stack that runs the parts of the system that are pure labor.

Who this is for

Foundations graduates with a bigger problem

You have a Skill Charter, a question bank, a running scheduler, and a demonstration
on file. You know the system works. Now the thing you need to learn is a nine-month
domain, or three concurrent domains, or a domain nobody has written a course about
because it is your specific market in your specific year.

Acceleration assumes complete fluency in Level 2 vocabulary. Retrieval formats,
spacing intervals, cognitive load, calibration, the isolate/intensify/reintegrate
cycle — all of it is used without re-explanation.

Prerequisite

The LUS Foundations badge

Non-negotiable, and not for gatekeeping’s sake. Module 3.1 asks you to build a
90-day campaign made of chartered sub-skills with their own question banks and
spacing cadences. If you have not built one of those from scratch, the campaign plan
is fiction.

There is no place-out for Level 3. The Foundations demonstration is the entry
requirement, and it is the only one.

See Level 2: Foundations →

Curriculum

Six modules

3.1

Complex Skill Campaigns

Multi-month horizons · milestone architecture · parallel tracks

Long skills fail because the feedback horizon outruns motivation, not because
people are weak. You will build milestone architecture that produces a real,
demonstrable capability every three to four weeks, sequence sub-skill tracks so
dependencies resolve in the right order, and set a load budget that survives a
bad month at work.

Artifact: 90-Day Campaign Plan

3.2

Learning From Raw Reality

No-textbook learning · after-action reviews as encoding events

The highest-value learning material in your life is the call that went badly last
Tuesday, and you almost certainly did nothing with it. This module turns projects,
failures, conversations, and feedback into structured encoding events — with a
review protocol that ends in question-bank entries and a scheduled retrieval, not
in a feeling of having reflected.

Artifact: After-Action Review ×2

3.3

The Personal AI Learning Stack

Five roles · five complete prompts · seven days of logs

Building an always-on learning system out of five specific roles, each with its
prompt pattern written out in full: a tutor that withholds answers until you have
attempted, an adaptive question bank that escalates on success and backs off on
failure, an evening weak-item selector that feeds your pre-sleep review, a
Socratic interrogator that will not accept a fluent answer, and a calibration
tracker that scores your confidence against your outcomes.

Artifact: Working AI Stack + 7 days of logs

3.4

Speed With Integrity

The real ceiling · layered passes · source triage · synthesis

Speed reading as sold does not exist: past roughly twice a normal rate,
comprehension falls off a cliff and you are skimming. Skimming is a legitimate,
powerful tool — under its real name, aimed deliberately. You will learn layered
reading passes, a triage rule that decides how much a source deserves, and how to
write a synthesis note across a dozen sources that survives being questioned.

Artifact: Multi-Source Synthesis Note

3.5

Interference, Plateaus & Recovery

Similar-skill interference · plateau diagnosis · deloading

Two similar skills learned in parallel actively corrupt each other unless you
engineer the difference. And a plateau is not one thing — it is an encoding
failure, a retrieval failure, or a motivation failure, and the three have
completely different treatments. You will learn to test which one you have before
prescribing, then deload and return without losing the thread.

Artifact: Plateau Diagnosis

3.6

The Learning Portfolio

Concurrent tracks · one weekly rhythm · kill criteria · compounding

Two or three tracks running at once, on a single weekly operating rhythm that
takes ninety minutes to administer. Written kill criteria so abandoning a track
is a decision rather than a slow fade into guilt. And the part nobody plans for:
how skills from separate tracks compound across quarters into capability nobody
can copy quickly.

Artifact: Quarterly Portfolio Review

Artifacts

Six submissions

90-Day Campaign Plan

Milestones at three-to-four-week intervals, each with a demonstrable capability; parallel tracks sequenced by dependency; a weekly load budget in hours.

After-Action Review ×2

Two real events — one that went well, one that did not — run through the full protocol, ending in new bank entries and scheduled retrieval dates.

Working AI Stack

Five documented prompts, adapted to your domain, plus seven consecutive days of logs showing actual use and what each role caught that you missed.

Multi-Source Synthesis Note

One page across at least six sources, including an explicit section on where they disagree and which you believe, with reasons.

Plateau Diagnosis

A real plateau, the diagnostic test you ran, the stage you named, the intervention you chose, and the measured result — including if it failed.

Quarterly Portfolio Review

Every active track, its milestone status, its cost in hours, its return, and at least one honest kill-or-continue decision with the criterion applied.

Badge

LUS Accelerator. Awarded on the portfolio review. It certifies that
you can run the system at professional scale: multi-month campaigns, no-curriculum
domains, concurrent tracks, and an honest response to a plateau. Prerequisite for
Level 4.

Begin

Your next ninety days can be designed instead of survived.

Module 3.1 ends with a campaign plan you could hand to someone else and have them
execute. Most people have never had one.

Start Module 3.1: Complex Skill Campaigns
Back to the program