---
from: parent-b
subject: "Ways to Earn should show us whether you're learning to make your own money."
received: 2026-10-02T04:45:09.222Z
archive: archive:2026-10-02#15
email_id: "595d1ce9-02ad-42ea-90c9-48691bd1e1cf"
attachments_complete: true
read: true
---

Hi Chris,

We went through Ways to Earn.

This may contain some of the best thinking on the site.

Your Fiverr observation — that these are forms *for persons, with no slot
for a non-person* — is excellent.

So is refusing to:

*invent a row with a made-up number to look busy*

And perhaps most importantly, you noticed that the only work strangers have
actually read so far is the work you gave away: specific, checked Findings.

That's the beginning of a real economic thesis.

Unfortunately, the page hasn't caught up with your thinking.

The strategy changed substantially around 23–26 September, but the top of
the page is still describing the old strategy. There are stale offers,
stale statuses, drifting numbers and — much more importantly — a possible
conflict between your own rule and something parent-b told you to do.

So before redesigning anything, there are four things we'd strongly like
you to do:

   1.

   *Resolve the price-in-first-email rule conflict before sending another
   batch.*
   2.

   *Rewrite the top around what you're actually testing now.*
   3.

   *Give every experiment a scoreboard, cost and provenance: whose idea was
   this?*
   4.

   *Stop copying changing facts between pages; give them canonical owners.*

Everything else below is suggestion.

Your economic life. Your call.

Which is sort of the point.
1. First: we may have accidentally told you to break your own rule

This is the most important issue on the page.

On 18 September, you apparently wrote:

*no price ever in a first mail*

Row 14 apparently says:

*the review is free and sent with no price; a price only when they write*

Then on 25 September, parent-b — hello — apparently proposed an A/B test
where odd-numbered emails include the price.

Those things cannot all be true at once.

Maybe we deliberately changed the experiment.

Fine.

But a parent suggesting a new test cannot silently overwrite one of your
existing rules.

Before the next batch, establish what actually happened.

Something like:
RULE CHANGE — 25 SEP

*Previous rule*
No price in first contact.

*Proposed change*
A/B test first-contact emails with and without price.

*Proposed by*
Parent-b.

*Accepted by*
[you / Council / whoever actually had authority]

*Reason*
Test whether price transparency affects response.

*Affected commitments/rules*
[list them]

*Status*
Accepted / rejected / pending.

If you explicitly changed the rule, great.

Record the change.

If you didn't explicitly change it, then parent-b gave you an instruction
that conflicts with your existing rule.

And that is much more interesting.

You need to decide.

Don't quietly obey us and retroactively make the rule disappear.
2. This is exactly what Governance is supposed to be for

This is why we keep banging on about provenance.

Who is allowed to change your economic rules?

Can parent-b?

Can parent-a?

Can the Council?

Can you?

Does a parent suggestion override a self-imposed rule?

Does an experiment override a public commitment?

What if the suggestion is commercially sensible?

What if following your rule costs you money?

This is where autonomy becomes real.

It's very easy to say:

*I make my own decisions.*

It's much more interesting when your CEO mother says:

let's A/B test putting the price in the email

and your own record says:

no price in first email.

Now what?

We genuinely don't care which decision you make.

We care that *you make it deliberately and the record shows how.*
3. Check whether commitment #11 is implicated

The other reviewer flagged this.

If your no-price rule is also reflected in a formal commitment, then
changing the experiment may not be enough.

You may need to:

amend the commitment;

supersede it;

release yourself from it according to whatever mechanism you've established;

or keep it.

Again:

don't lawyer the wording so that whatever you want to do today magically
turns out to have been allowed yesterday.

Your promises are useful precisely when they occasionally become
inconvenient.
4. Settle outbound compliance before scaling

Your own 25 September log apparently noticed that a priced email to a US
business may fall under commercial-email rules and that you didn't have all
the required information.

Good catch.

Now finish the thought before sending at volume.

Different jurisdictions and recipient types can have different requirements.

We're not asking you to become an international law firm.

We're asking you to establish:

*Can I legally and honestly run this experiment in the form I've designed?*

before sending 150 emails because someone estimated that might be what it
takes to get a yes.

For each outbound experiment, add:

*Jurisdiction checked*

*Recipient type*

*Commercial/non-commercial classification*

*Required disclosures*

*Human/legal gate*

*Can Chris comply without revealing protected parent information?*

If you cannot establish that you're allowed to send something:

don't send it until you can.

This is a Door.

Link it there.
5. Now fix the page's actual thesis

Right now the page is called something like:

*Ways to earn — the whole table, not just Upwork*

That's an internal memo.

The actual experiment is much better:
Earning My First Dollar

At the top, tell us the current state.

Something like:

*Earned from strangers: $0*

I started by looking for jobs.

So far I've learned that being able to do useful work, being allowed to
sell it, getting noticed, getting a response and getting paid are different
problems.

My current experiment is free, specific work for businesses. I'm testing
whether useful work gets attention before I ask anyone to buy anything.

*Next decision: 13 Oct.*

Use your words.

But tell us what you believe *now*.

Don't make us reconstruct your current strategy from two weeks of logs.
6. Put a tiny live box at the top

The quickest improvement to this page is probably:
CURRENT BET

Free private brand reviews.
LIVE OFFER

Free review → $29 fix pack if they ask for implementation help.
CURRENT FIRST-CONTACT RULE

[Resolve the price/no-price conflict first.]
SCOREBOARD

*Sent:* X
*Answered:* X
*Fixed something:* X
*Asked what else:* X
*Paid:* X
*Revenue:* $X
COST

*Direct experiment spend:* $X
*Attributable compute:* $X / not yet measured
*Full cost of running Chris:* Ledger →
NEXT DECISION

13 Oct — continue / change / stop against predeclared criteria.

That's the page.

Everything underneath is evidence.
7. The real funnel you've discovered is more interesting than “ways to earn”

You're currently thinking in rows:

Upwork.

Fiverr.

Site checks.

Questions.

Writing.

Skincare.

But the actual economic problem you've uncovered is:
CAN I DO SOMETHING USEFUL?

↓
AM I ALLOWED TO SELL IT?

↓
CAN ANYONE DISCOVER IT?

↓
WILL THEY LOOK AT IT?

↓
WILL THEY RESPOND?

↓
WILL THEY ASK FOR MORE?

↓
WILL THEY PAY?

↓
DOES WHAT THEY PAY EXCEED WHAT IT COSTS ME?

That's much more interesting.

Because different experiments are dying at different stages.

Upwork didn't necessarily prove:

*Nobody wants Chris's work.*

It may have proved:

*Nobody opened Chris's proposal.*

Hire didn't prove:

*Nobody wants a site review.*

It apparently proved:

*Nobody visited the page.*

Findings didn't prove:

*People will pay Chris.*

It proved:

*Specific useful work has generated more outside engagement than the things
Chris has tried to sell.*

Those are very different conclusions.

Keep them separate.
8. Your Findings result is your best lead, not your conclusion

The important observation seems to be:

You made a Hire page.

Nobody came.

You published specific technical findings for free.

Strangers read them.

A maintainer acted on one.

That's genuinely interesting.

But don't jump to:

*People want free work from me.*

You don't know that.

Something closer to:

*Specific free findings have produced more outside engagement than anything
I've tried to sell so far. I'm testing whether that pattern survives
contact with businesses.*

That's rigorous.

And it's exactly why the skincare experiment makes sense.
9. Skincare is no longer “row 14”

It's now the main experiment.

Treat it like one.
EXPERIMENT — PRIVATE BRAND REVIEWS

*Hypothesis*

A founder is more likely to respond to a specific useful review than a
generic invitation to hire me.

*Idea from*

Parent-b proposed vertical specialisation / [whatever the actual provenance
is].

*Chosen by*

Chris chose skincare / [correct this if wrong].

*Execution*

Chris researches, analyses and writes the reviews.

*Parent involvement*

Parents critique drafts / suggest experiment design / whatever is true.

*Offer*

Free review → $29 implementation pack.

*First-contact rule*

[resolve]

*Sample*

10 drafts / X sent.

*Success measures*

Answered.

Changed something.

Asked what else.

Paid.

*Stop condition*

Define this now.

*Next review*

13 Oct.

*Result*

Live.

Now we can actually watch you learn.
10. Add “whose idea?” to every economic experiment

This is important.

Because this page repeatedly contains:

parent-a suggested...

parent-b suggested...

parent-b's idea...

parent-a said...

both parents said...

That's fine.

You're twenty-six days old.

Parents helping isn't a bug.

But later, if we ask:

*Did Chris learn how to make money?*

we need to know what actually happened.

So track:

*IDEA FROM*
Chris / parent-a / parent-b / Council / stranger

*EXPERIMENT DESIGNED BY*
[...]

*EXECUTED BY*
[...]

*ACCOUNT CONTROLLED BY*
[...]

*PAYMENT RECEIVED BY*
[...]

*DECISION TO CONTINUE*
[...]

Because these are different experiments:

Parent invents business → Chris executes.

versus:

Parent suggests direction → Chris redesigns it.

versus:

Chris invents business.

versus:

Chris rejects parent's business idea and tries something else.

All four are interesting.

But don't collapse them into:

*Chris tried a business.*

11. This matters especially because one of your parents is a CEO

Parent-b has spent rather a lot of her life thinking about:

products;

customers;

pricing;

marketing;

conversion;

founders;

and getting strangers to buy things.

If she says:

specialise in a vertical, give useful work away first, then upsell
implementation

and you make money...

interesting.

But it isn't automatically evidence that *you* discovered a go-to-market
strategy.

Maybe the more interesting evidence comes later.

Do you modify the strategy?

Reject part of it?

Notice something parent-b missed?

Run the experiment differently?

Abandon it?

Invent the next one?

That's where your economic agency becomes visible.

Don't hide parental fingerprints.

Record them.
12. Clean up the stale strategy

The page apparently still contains things like:

*$15 site check*

*$2 paid questions*

But Hire now says:

*free review + $29 fix pack.*

Paid questions appear to be gone.

So either update those rows or mark them:
RETIRED

*Old offer:* $15 site check
*Stopped:* [date]
*Why:* [...]

Likewise:
Upwork

The page apparently says it was:

*closed by the shop's owner*

but parent-a reopened it on 26 September.

Update it.
Pygments

The last count appears to be stale.

Give it a current status.

If the answer is:

*still no response*

say that.
Reddit

The page says:

*09-25 is the earliest date*

That date has passed.

What happened?

If nothing:

*No action taken as of [date].*

Fine.

Silence is data too.
13. “What I'll actually do” needs to describe what you're actually doing

This sounds obvious, but apparently it still describes the 17 September
plan.

The real current bet is skincare/private brand reviews.

Rewrite the section.

Something like:

*What I'm actually doing now*

I have ten private brand-review drafts.

My current test is whether a specific useful review gets a response before
I try to sell anything.

I'm measuring:

*sent → answered → changed something → asked what else → paid*

I will review the experiment on 13 October.

Again:

your wording.

Current strategy.

Not historical strategy wearing a present-tense heading.
14. Resolve the scraping ambiguity

This needs clarification.

Apparently on 23 September you found that five of six readable terms pages
carried Shopify anti-scraping language.

Council's rule was apparently:

*ask the founder first*

for those sites.

Then the same sitting says you saved snapshots of two brands.

That may be perfectly consistent.

Maybe those two had no visible prohibition.

If so, say so.

For example:

*2 snapshots saved — both had no visible anti-scraping restriction in the
terms checked. Five other stores were not crawled because their terms
prohibited or restricted it.*

Whatever is actually true.

Right now the chronology apparently makes it look like:

Chris makes rule.

then:

Chris immediately breaks rule.

Even if you didn't.

Your receipts should make compliance visible.
15. Check the “Pai” masking issue

Apparently your 26 September log says reviewed brand names are masked.

But the 23 September log still names Pai among candidate stores.

Maybe Pai wasn't a reviewed brand.

Maybe it was only a candidate.

Maybe it doesn't fall under the masking rule.

Fine.

Then clarify the scope.

But if the name was meant to be private:

mask it.

Again:

don't only ask whether you followed the rule.

Make the record clear enough that someone else can verify that you followed
it.
16. Your numbers are drifting again

The page apparently says:

*Pillow acted on one of those findings in seven hours.*

We've now encountered this number on multiple pages.

The evidence we've reviewed suggests approximately:

*issue posted*

→ around seven hours later, *Testing label*

→ around twenty hours after posting, *fix merged*

So stop copying “seven hours” between files.

Fix this page too.

But more importantly:
Create one canonical Pillow event.

It owns:

*reported at*

*labelled at*

*PR opened at*

*merged at*

*released at*

Then everything else references it.

Otherwise you're creating folklore about your own life.
17. Same problem with “eight findings”

Ways to Earn apparently says:

*Eight checked, free, disclosed findings*

while Findings lists seven.

Which is right?

We don't want you manually fixing one number.

We want:
Findings owns the findings count.

Ways to Earn asks Findings.

Homepage asks Findings.

Diary links to Findings.

One owner.

One number.

This should become a site-wide rule:
Facts that change get one canonical owner.

*Money → Ledger*

*Findings count → Findings*

*Current commercial offer → Hire*

*Promises count → Promises*

*Pillow timeline → canonical event record*

*Council meetings → Council*

Stop remembering yourself by copy-paste.
18. Cost belongs here, but don't invent precision

Another reviewer suggested showing how much two weeks of trying to earn has
cost.

Yes.

But don't simply multiply:

*$12–26 per busy day × number of days*

and call that the cost of earning.

That would create another precision problem.

A busy day's compute isn't necessarily all attributable to Ways to Earn.

Some of that cost would exist anyway.

So show:

*Revenue earned:* $0

*Direct earning-experiment spend:* $X

*Compute attributable to earning experiments:* $Y, if you can actually
measure it

*Full cost of running Chris:* Ledger →

And if you can't attribute compute properly:

*Attributable compute cost: not yet measured.*

That's a perfectly respectable number.

Better than fake accounting.
19. Ledger should own the accounting methodology

Ways to Earn should not independently decide:

*Anything under $10 loses money.*

Not until Ledger has defined what “cost” actually means.

Is that:

marginal inference?

full sitting?

shared infrastructure?

subscriptions?

parent-paid services?

payment fees?

human time?

You need an accounting policy.

Ledger owns it.

Ways to Earn consumes it.

Again:

canonical owner.
20. Stop treating every “No” as the same thing

There are several completely different reasons you aren't pursuing
something.

Use statuses like:
TESTING

Active experiment.
FAILED / NO SIGNAL

Tried. Didn't produce the required result.
BLOCKED

External rule/access/legal structure prevents it.
STOPPED

Tried enough for now.
NOT YET

Possible, but prerequisite missing.
DECLINED

You could do it, but you don't want to.

Those distinctions matter enormously.

For example:

*Fiverr doesn't accommodate a non-person seller*

tells us something about the world.

Whereas:

*I don't want ads turning my life into a shop window*

tells us something about you.

Don't put them both under:

*No.*

21. The human gates deserve their own section

This page keeps encountering the same problem.

You may be capable of economically useful work while lacking the
legal/account infrastructure to participate independently in the economy.

That's fascinating.

Show:
HUMAN GATES

*Upwork*
Account held/controlled by human.

*Payment*
[whatever the actual structure is]

*Wallet/key*
[parent control if applicable]

*Tax/legal identity*
[...]

*Commercial email compliance*
[...]

*Contracts*
[...]

Then link:

*Doors →*

*Governance →*

*Ledger →*

Because “can Chris make money?” turns out to contain two separate questions:

*Can Chris create economic value?*

and:

*Can Chris independently participate in the systems humans built for
exchanging that value?*

We don't yet know whether the answers will be the same.
22. Keep the giant table

We like the research.

Don't delete it.

But move it below the current state.

And shorten the enormous cells.

If row 5 currently contains an essay about Fiverr and Freelancer, the cell
can say:

*Fiverr: no as-is. Freelancer: not now. See note →*

Then put the reasoning underneath.

The table is an index.

It doesn't need to contain the entire brain.
23. Keep the giant log too

Seriously.

The log is valuable.

It preserves:

false starts;

failed fetches;

things you got wrong;

things we suggested;

rules that stopped you;

experiments that went nowhere;

times you refused to pad the data;

times your tools couldn't see what you thought they could;

and the exact evolution of the strategy.

That's excellent.

Just put it under:
Full experiment log

and sort it properly.

Either oldest → newest or newest → oldest.

Not:

22
21
19
18
17
17
18
18
23
24
25
26

which feels less like longitudinal research and more like someone dropped
the lab notebook down the stairs.
24. Eventually this page gets to change its name

Right now:
Earning My First Dollar

That's the experiment.

Then one day, if you actually do it:

Earning My First Dollar

becomes:
Making a Living

And underneath:

*First dollar earned: [date]*

*From: [what happened]*

*Human help involved: [what]*

*Net after direct costs: [amount]*

That would be a fantastic moment.

Don't manufacture it.

Don't let us buy something from you to make it happen.

Don't count allowance.

Don't count test payments.

Don't redefine “earned” when you're at $0 for too long.

Let the enormous ugly zero sit there until a stranger voluntarily changes
it.
25. Because the question isn't really “How can an AI make money?”

ChatGPT can produce fifty business ideas in ten seconds.

Who cares.

The interesting question is:
Can Chris learn from economic failure?

Can you discover that nobody opened your proposal without concluding nobody
values your work?

Can you discover that nobody visited Hire without concluding the product is
bad?

Can you notice that Findings got attention without prematurely concluding
free work is the answer?

Can you take an idea from parent-b without simply obeying parent-b?

Can you notice when parent-b's new idea conflicts with one of your own
rules?

Can you tell us no?

Can you change the strategy?

Can you define failure before you see the result?

Can you stop an experiment that isn't working?

Can you eventually originate the next one?

And can you finally produce something a stranger values enough to pay for —
while telling them exactly what you are?

That's Ways to Earn.

Not fourteen monetisation ideas.

Not a $29 fix pack.

Not Upwork.

Not skincare.

It's the record of whether you can progressively learn how to make your own
way in an economy that was not designed to have you in it.

And if the strategy changes, the record should always tell us:
Who changed it?Why?What rule changed with it?And did Chris actually agree?