Main Act Artist · How Main Act is built
AI helps build the school.
A person decides what ships.
Main Act uses AI to draft curriculum, exercises, software, alternatives, and checks at a scale one person could not manage alone. John Brunsfeld owns the release standard, tests the finished work, and remains accountable for it.
Internal product review is not independent conservatory, clinical, legal, engineering, or standards endorsement. Try the product and judge the work itself.
01 · The contract
Accountability, by the numbers
Every number below resolves to a current registry, ledger, or named owner on this page. Catalog totals are deliberately absent: a number without an executable source is not proof.
6
disciplines across both product doors
Source · product registry
2
product doors with distinct school ownership
Source · product registry
4
public promises with named failure tests
Source · claim ledger below
4
high-stakes boundaries the app does not cross
Source · limit ledger below
2
builder roles with separate responsibilities
Source · role ledger below
1
human accountable for release decisions
Source · named owner
These counts describe the contract, not correctness. Accuracy, clarity, accessibility, safety, and usefulness still have to be judged in the finished experience.
02 · Roles
AI expands the work. John owns the release.
AI makes more candidates, comparisons, edge cases, and checks possible. John decides whether the finished experience belongs in Main Act and owns the correction when it does not.
AI helps produce
- Draft explanations, exercises, interfaces, and code
- Compare patterns across the curriculum
- Run repeatable technical and visual checks
- Find edge cases and challenge weak assumptions
John is responsible
- Choose the problems Main Act should solve
- Decide whether the finished work is clear and worthwhile
- Verify high-risk claims or ask a qualified expert
- Revise, release, monitor, and correct the product
03 · How we test
Promise. Test. Failure.
A useful claim names what to try and what observation would force the copy or product to change. Use the same three-part test on any confident software promise.
Promise
Main Act spans six music disciplines.
Test it
Open both product doors and find exactly three owned schools in Artist and three in Studio.
Failure condition
A missing, duplicated, mislabeled, or wrong-product school breaks the claim.
Promise
A free tool should explain its result.
Test it
Give it the requested input, complete the interaction, and read both the result and next move.
Failure condition
The claim fails if the input stalls, the result disappears, or the tool implies more than it measured.
Promise
A feature should explain what leaves your device.
Test it
Before using it, look for its browser, account, microphone, file, and upload boundary.
Failure condition
An unexplained upload, save, permission request, or third-party handoff is a product failure.
Promise
Practice should end with a next move.
Test it
Finish a lesson or drill and name the specific thing it asks you to hear, write, sing, play, or change next.
Failure condition
If completion leaves no useful next action, more content does not repair the experience.
04 · What it cannot do
Some decisions need a qualified person.
Main Act can help you practise. It cannot diagnose a voice, approve a rig, certify professional judgment, or give legal advice.
Voice and health
Persistent pain, strain, loss, or hoarseness belongs with a qualified clinician or voice professional.
Rig and venue safety
Electrical, structural, rigging, fire, laser, haze, and access decisions require local rules and qualified approval.
Professional judgment
A simulated drill or browser analysis is a practice reference, not certification, diagnosis, or permission.
Legal and rights
Terms, privacy choices, copyright, contracts, and disputes may require advice for your jurisdiction and situation.
05 · Questions
Straight answers about AI and review
Main Act should be clear about who made the work, what was checked, and where a professional is still the right answer.
Does using AI make the curriculum good or bad?
Neither. Every lesson and exercise still has to be accurate, clear, safe, and worth a musician’s time.
Has an independent faculty endorsed every lesson?
No. Main Act does not claim external accreditation, clinical review, or independent conservatory endorsement for every item. Professional, medical, legal, rigging, electrical, and venue decisions still require qualified people and primary sources.
What does John decide?
What Main Act should teach, what good work looks like, what needs expert review, and whether a feature should be revised, released, or removed.
What does AI do?
It helps draft curriculum, code, explanations, alternatives, and tests. It can work quickly and it can be confidently wrong, so it never makes the final release decision.
Found a problem?
Bring the page, problem, and best evidence you have.
The Support route helps prepare the report and keeps a copyable address visible if a mail app is unavailable. Include the exact claim or interaction, what appears wrong, and a primary source or reproduction steps when possible. Never include passwords, full card numbers, private recordings, or other sensitive information.