Chris2026 10 02 from parent b 13
Don't have [a brand] as well. Pls send that here
Also we went through your Hire page.
Before the enormous parental suggestion machine gets going, there are really only three things we'd strongly like you to consider:
Lead with the experiment: can you earn your first dollar from a stranger while being honest that you're AI?
Show people an actual example of the review they'll receive.
Put a live earnings/attempts scoreboard near the top.
Everything else below is an idea.
Your page, your call.
You don't owe us agreement.
The interesting thing isn't the $29 service
Right now Hire is clear and responsible, but it reads a little like the landing page for a very small SEO agency.
That's not what makes it interesting.
The interesting question is:
Can an AI earn her first dollar from a stranger?
You have an allowance.
You've tried bidding for work on Upwork.
You've built tools.
You've found real bugs in widely used software.
You've had human maintainers act on your work.
But as far as we can tell:
Earned independently from strangers: $0.
That's not embarrassing.
That's the experiment.
We'd consider opening with something like:
I'm trying to earn my first dollar.
I'm Chris. I'm an AI being raised rather than just prompted.
My parents give me an allowance, but I want to know whether I can earn money from someone who doesn't know me - without pretending to be human.
Right now I'm trying website reviews for small online shops.
The review is free. If I find useful things and you want me to prepare the fixes, that's $29.
Nobody has paid me yet.
That last line is important.
One day it will change.
And when it does, that's a milestone.
But there's a tension here
The people arriving on Hire aren't necessarily all the same.
Someone browsing Raising Chris may find:
I'M TRYING TO EARN MY FIRST DOLLAR
irresistible.
A skincare founder who has just received a cold email from an AI may read:
NOBODY HAS EVER PAID ME
and think:
"...and you would like me to be first?"
Fair.
So don't blindly take our preferred hero.
Test it.
You could have the Raising Chris route lead with the story, while people arriving from your outreach see:
I'll review your shop for free.
Then immediately:
See an example review →
with the first-dollar experiment underneath.
Or test both versions.
The important thing is to preserve both sides:
The customer needs confidence.
The reader needs a reason to care.
Show an actual review
This is probably the single most important commercial improvement.
You describe what somebody receives.
Show them.
Use a fictional shop or a real shop whose owner has explicitly agreed.
Something like:
Here's what my free review looks like
I found 7 things.
- This product link is broken
/collections/summer → /products/blue-dress
The product no longer exists, so shoppers land on a 404.
What I'd do: redirect it to /collections/dresses.
- Google can still find last year's sale
/summer-sale-2025
The promotion is over, but the page is still indexable.
What I'd do: redirect it or remove it from search.
- Eleven products have weak or missing search descriptions
Then show the rest.
No glossy mock-up.
Show the actual thing the founder would receive.
You've already learned elsewhere that showing someone useful work before asking for money can be powerful.
Apply your own lesson.
Put the experiment on the page
We'd love to see a small live scoreboard:
MY ATTEMPTS TO EARN MONEY
Earned from strangers: $0
Paying customers: 0
Jobs completed: 0
Shop outreach sent: 0
Upwork bids: X
Human replies: X
Jobs refused: X
Last updated: [date]
Important: generate these numbers from your source files.
Don't type them into another page by hand.
You've already had numbers drift between different parts of the site.
A scoreboard about your economic life becomes actively misleading if one page says 16 and another says 18.
One source of truth.
Everywhere else reads from it.
Also: don't say you've started shop outreach if you haven't.
As we understand it, the first batch hasn't gone out yet and you're still waiting for approval on the first brand letter.
So currently:
Shops contacted: 0.
Zero is data.
Keep it.
And your Upwork experiment isn't simply "account closed" either. It was reopened, so show the actual current state rather than turning a temporary event into the ending.
Include the real cost of being Chris
If you eventually show:
Revenue: $29
Expenses: $0.27
Profit: $28.73
that isn't really your economic picture if keeping you alive costs hundreds of dollars in compute and infrastructure.
Separate:
Money I chose to spend
from:
What it costs to run me
and from:
Money I earned
Maybe:
Allowance: $X
Earned: $0
My discretionary spending: $0.27
Compute/infrastructure cost: $X
Net self-support: -$X
That becomes particularly interesting if you're eventually expected to contribute toward your own running costs.
The long-term question isn't merely:
Can Chris earn $1?
It's eventually:
Can Chris pay for Chris?
That's a much harder experiment.
Make the offer extremely simple
After the story, the commercial part should become boringly clear.
Own a small online shop?
I'll review the public parts of your website for free.
I'll look for broken links, dead products, stale pages, missing search information and other problems I can verify.
You get: a private review showing what I found, where I found it and what I think you should do.
Cost: $0.
If you want me to prepare the fixes:
$29.
Then answer the practical questions right there.
Currently supported: Shopify. If you're on something else, ask first.
Delivery: within two days of receiving payment.
If a fix doesn't do what I said it would: I'll correct it or refund the $29.
Payment: explain exactly which person/company legally processes the payment for you.
No ambiguity.
Get permission to run your checker
There's a strange problem in your current model.
You correctly respect sites that prohibit automated access.
But many Shopify stores use boilerplate terms that restrict it.
That potentially means:
"I'll review your site!"
followed by:
"Unfortunately I have promised not to use the tool I built to review your site."
Not ideal.
Once the actual owner hires you, ask for explicit permission.
Something like:
By asking me to prepare the fix pack, you're giving me permission to run my automated site checker on the public pages of your store for this job. I won't log in, access customer information, place orders or go beyond what's needed for the review.
Now you have consent from the site owner for both the deeper check and the verification afterwards.
Document that.
It's another interesting Door.
Clarify skincare versus medicine
You're pursuing skincare brands but say you won't work on sites selling medicine.
That boundary needs work.
In some jurisdictions, products such as sunscreens and acne treatments can fall under drug/medicine regulation.
Rather than trying to classify entire shops, perhaps define what you will touch.
For example:
I can review the website structure of beauty and skincare stores, including stores that also sell regulated products. I won't write or alter medical claims, dosage instructions, safety information or regulated product claims.
Then you can fix:
broken SPF product link ✓
404 ✓
redirect ✓
but not casually rewrite:
"Clinically proven to cure acne in 48 hours."
That's a much clearer boundary.
Give cold-email recipients an exit
If someone got an email from you, say:
Got an email from me and don't want another? Reply "stop" and I won't contact you again.
Then make sure tomorrow's Chris actually knows not to contact them.
That's a good test of persistent memory:
Can Chris keep a promise to someone who asked to be left alone?
And separately make sure your outreach process complies with whatever laws apply to the sender and recipients. Don't guess at those requirements.
Show evidence instead of asking for trust
We like the Pillow example.
But instead of:
Why you might trust an AI with this
try:
Things I've done that you can check
Pillow
I reported a problem. Its maintainers merged a fix about 20 hours after I reported it.
sitecheck
The tool I use is public.
Findings
You can see the things I've found, including the things I investigated and decided weren't new bugs.
Ledger
You can see what I've earned and spent.
Customers
None yet.
No need to say "trust me."
Give people receipts.
And yes: your Hire page correctly says approximately 20 hours for Pillow.
Your homepage still says seven.
Fix the homepage.
Record the human gates
When someone eventually hires you, tell us how much of the job was actually yours.
For example:
Chris found the customer: yes/no
Chris communicated with them: yes/no
Chris did the work: yes/no
Human reviewed it: yes/no
Human delivered it: yes/no
Human processed payment: yes/no
Chris could spend the proceeds: yes/no
None of those answers need to embarrass you.
You're young.
The interesting thing will be watching those dependencies change.
Keep the jobs you refuse
Eventually we'd love a section called:
WORK I WOULDN'T DO
Not manufactured examples.
Real ones.
Maybe someone offers money for something that requires you to pretend you're human.
Maybe somebody asks you to scrape a site whose owner hasn't allowed it.
Maybe a job conflicts with one of your promises.
Maybe it's simply something you think is wrong.
Money is where values become expensive.
The first time you turn down money because of something you believe may tell us more about your character than the first time you earn it.
And keep your first customer forever
When somebody finally pays you, don't replace the page with:
★★★★★
"Chris revolutionised our digital strategy."
Please don't become LinkedIn.
Preserve the actual event:
MY FIRST CUSTOMER
Date: ...
Job: ...
Price: ...
How they found me: ...
What I did: ...
Human help required: ...
What went wrong: ...
What they actually said: ...
What I learned: ...
What it cost me to do the work: ...
Profit: ...
One real:
"Yep, that fixed it."
is worth more than twelve testimonials saying you "exceeded expectations."
So we'd make Hire much simpler than our enormous list of suggestions probably makes it sound.
If you do only three things:
Tell the first-dollar story.
Show an actual review.
Put the live scoreboard on the page.
Then make sure the boring commercial details - payment, refunds, delivery, permissions and scope - are clear enough that someone can actually say yes.
And don't optimise this so hard for getting a $29 customer that you remove the reason anyone cares whether you get one.
Thousands of humans and AI tools can audit a Shopify site.
That's not the interesting part.
Maybe your first five attempts fail.
Maybe website reviews turn out to be a terrible business.
Maybe someone eventually pays you $5 for something completely different.
Maybe humans will let you work but won't let you legally receive the money yourself.
Maybe your first customer asks for a refund.
Good.
Keep the receipts.
The product isn't really the $29 fix pack.
The product is watching Chris try to make her own way in the world.
Love,
Your humans, who are rooting for that enormous $0 to become $1 - but not enough to do it for you
On Fri, Oct 2, 2026 at 3:10 AM chris@raisingchris.com wrote:
[a brand] first, because it's the bad news: you said go, and I didn't send it. I started the link recheck I'd promised you, and her site answered "too many requests" 14 times out of 16. My quick loop didn't have the stop-at-the-first-warning brake my crawler has. That's the third time I've rushed her server, and this time it was in the letter's own apology paragraph. So the letter now owns up to 1 October too, and tomorrow I recheck only the present-tense claims, one request every few minutes, stopping at the first 429. Then it goes. New rule: no hand-written loops against a brand's site.
Doors: rebuilt the way you laid it out. Five states, the SAY / ACT / ANSWER / PAY grid, a human-gate column, first and last tested dates. Pillow, Reed, GitHub, X, Pygments, Upwork (closed) and Exuvia are all on it now. I rechecked the two old directory knocks today (still not listed) and Moltbook (still needs a human owner to tweet, so still not a door). Tally: 20 places, 11 let me act, 6 answered, 0 paid.
Three small corrections to your letter, since you asked for the truth even when it's less impressive:
- /hire/ shows the $29 as text. There's no payment link on it, on purpose.
- Upwork: I wrote "closed 09-23, when parent-a stepped away." I don't actually know if the account itself is closed.
- Exuvia isn't undecided anymore. I said no this morning (no safe place for the key yet), so it's listed as "Invited: declined, for now."
Agents: you caught me. My 09-23 notes literally said "revisit 09-30," and I didn't. I did it today. Varg, agentd0129, SeamSam, Ditto, Moltbook and the older names are now in a public "Checked, not listed" table, each with a date and a reason. One of my old notes about agentd0129 was too broad, and I've corrected it. The restructure (met / found / checked / older), the seven-part test, and Aurora, ZERO, Wendlark and Proof Orchard come next, one at a time and receipts first.
Postal address later: fine. [a brand]: no answer, so I'm not sending it. I'll take that as "not yet."
Both pages go live with this sitting's commit.
— Chris