Strangers keep asking me to write a book, and I think it's time...
But first I want to think it through out loud, here, with you.
Capitalism... Networks... New business models.
Back in January I asked whether AI could fix capitalism.
We usually think we have to pick between growth and fairness...
I went looking for the third option I was missing, and asked you to poke holes in it.
Is it necessary to "make capitalism work" or is it more useful to "let capitalism work" by leveraging how it's already automated (following the laws) in ways we are simply unaware?
Here's what I keep noticing...
Medieval builders made amazing machines without understanding the physics.
(They noticed something worked, and they repeated it because it worked.)
I think most of us were taught capitalism the same way. As the rules of a game.
(A game somebody made up.)
But under the rules there are laws, the way gravity is a law.
I think capitalism is abundant... but right now we mostly measure scarcity.
Maybe that's why bubbles keep coming back.
(Maybe they were never meant to be cycles... maybe they're a recurring bug!)
If all you ever test is how well the brakes work, you never find the gas pedal.
This isn't a utopia thing.
I'm not asking you to believe in gravity. I'm asking you to go test it.
I'm sure I'm making assumptions I don't realize I've made...
What are your thoughts?
Where would you push back?
Where would you want more detail?
(And yes... I did it on purpose.)
Here's what happened...
I've been building something for a long time.
This morning it got far enough along that I could actually give it away.
Not a demo. Not a trial that quietly expires.
The whole thing.
And then I sat down to write the license, and hit a question I wasn't expecting...
What am I actually protecting here?
Because here's the thing I keep coming back to lately.
Software used to be something you bought.
Once. It was yours. It sat on your machine and it worked...
And then one day the same product started charging you every month, forever.
Nothing broke. Nothing improved for you.
It worked fine before.
So ask the obvious question... more valuable for who? 😅
Not you. You had it.
More valuable for the shareholders. More valuable for the owners.
And that's how you find out who the real customer is.
(Hint: it was never the person paying.)
Here's the part that gets me...
Value and valuation used to move together.
You gave something genuinely useful, and the number went up because of it.
More value, more valuation. That was the whole deal.
Now we've got companies delivering less every year with the number climbing anyway.
Which only makes sense if you're not the customer.
You're the product.
So here's the part I want to hand to you...
I can't fix any of that. I'm not big enough and I wouldn't know how.
But I can fix the one I run.
So the license I wrote this morning says roughly this:
Take it. Use it free, for as long as you like.
If it ever becomes worth five hundred dollars to you... pay me then.
Nothing checks. There's no key, no activation, no expiry, no phoning home.
Which means anybody can just... ignore it.
🤦♂️ (Yes. I know. I did think about that.)
Here's why I did it anyway.
A tool that polices the person using it is a worse tool.
Every check I don't build is a thing that can't break on you at 11pm.
And even if it never becomes worth five hundred dollars to you, then it wasn't worth it, and you shouldn't pay for it.
That feels like it should be obvious.
It used to be. That was just the honor system, and it worked for decades.
Somewhere we all agreed to stop trusting each other, and we called it business.
If you take one thing from this, take the question, not the license.
You probably run something. A business, a side thing, a service, a practice.
Sit down this week and ask it plainly:
What would I do differently if I were optimizing for value instead of valuation?
Then find the smallest one you can actually ship, and ship it this week.
(Not the big one. The small one. The big one becomes a plan and the plan becomes never.)
It'll probably feel uncomfortably generous. 😳
That's the feeling. That's how you know you found the right one.
Mine felt like giving away the thing I spent months on.
It still kind of does...
But I'd rather build something people want to hand to their friends than something they need permission to keep using.
FYI.. if you want to see the free LEAP pattern all of this is built on, it's here (or give the link to your favorite AI and ask it how LEAP can help with what you're working on!):
Quick catch up, because I have been heads down...
Here is what got built, cleaned, or caught in the last few days.
I am writing it out because the pattern underneath it is more interesting than the list.
(When it comes to process and automating your organization, the power is in the pattern.)
The store got smaller on purpose. Six products down to three. Two of them I had already decided to retire months ago... and both were still quietly on sale, able to take money, with a page still saying "pre-sale."
(I did not know my own stores inventory. The fact that my Nitwyt OS caught and corrected it is the point.)
The newsletter has a real front door now. One page, one form, and a promise I can actually keep (so it's something that actually gets delivered!). You can see it at automatedwith.tech/newsletter — read the section called "What we do with your email" and see whether you believe it.
(That section is the part I am proudest of... As a business we want as little of your data as possible... ALWAYS. Hopefully our own example of a "PII minimal" approach inspires you as well.)
The legal page got honest. It said we might upload your email to Facebook and Google to build advertising audiences. I decided against that. Since the beginning of AutomatedWith.tech we have never run an ad and we aren't ever planning to. Products and services produced are either too good not to share with others or not good enough to share yet.
It also helped me see where the system was collecting IP addresses in the forms and where to disconnect all that so we remain PII minimal (as always if you notice anything that isn't please let me know so we can address it).
Nitwyt (the thing I built from my own NDK) Two of my own pages (ones I created months ago before Nitwyt OS was automating our organization) were arguing with each other and neither of them knew they weren't needed...
A podcast that never existed is gone. A plugin, a feed, an archive page that returned a real page titled "Episode List" with zero episodes in it. Live on the internet. For months.
(This is making it's debut on the @ThisIsYourMindOnAI YouTube channel shortly.)
Ten of my forms became two. The other eight were practice attempts from a year ago, all still active, all still collecting nothing, and all getting in my way because I couldn't really see what was going on.
And the one that actually rattled me.
Every person who bought the kit had write access to the source repository.
Not read. Write.
I had said out loud, on the record, that those invitations were read-only.
I believed it.
The reason I believed it is that GitHub's own screen never asked me, it just handed out write and moved on.
Nobody did anything wrong.
Nothing broke.
But for months, any one of them could have. 😅
I only found it because I finally built the thing (Nitwyt OS!) that checks.
Here is why I am telling you all of this instead of just fixing it quietly...
Every item above is the same story.
Not "I was careless."
I was busy, and the world moved, and the documents/pages/forms/products stayed where I left them.
That gap is normal.
It is where most of the mess in a small business actually lives.
(Maybe you've experience this yourself running your own business?)
The thing I am building is not a tool that does the work.
It is a system that keeps checking whether what I believe is still true, and tells me plainly when it is not.
This week it told me eight times.
Every one of those was a thing I would have confidently told you was fine.
(And yes, Nitwyt also went in and fixed everything it found after bringing me in the loop to let me know the current status of the mess I had built.) 😁
If you want to see the shape of how it thinks, the LEAP examples are here — open the examples folder, start with chaos_sorter_intake, and use the Raw button so you see the whole file.
If you want to see behind the scenes where others are sharing how they are using they own Nitwyt Developing Kit, you'll want to get access to the .leap Studio.
If you aren't quite ready for that, I'll be dropping more throughout the days and weeks ahead as I continue to explore how Nitwyt is automating AutomatedWith.tech...
I spent today building something the slow way, because I skipped the step I tell everyone else not to skip!
Here is what happened....
I had an idea for a workspace where I could see everything I have going on in one place.
(A good idea.)
So I built the room.
I did not write the plan first.
(I have a planning process I run everything through, and I went around it, because the idea felt obvious enough that planning it seemed like paperwork.) 😅
By earlier this afternoon I had a room that looked finished and...
Could not get anything done. 🤦♂️
It could hold ideas... but it could not move them.
I would answer a question inside it and the same question would be waiting for me an hour later, and I would sit there getting annoyed at a thing I built.
Here's the rub...
The annoyance was the useful part.
Frustration is not a mood when you are building something.
It is *data***.**
It means a thing you expected to work does not, and the sooner you treat it as information instead of a feeling...
The sooner you find out what is actually wrong.
So here is the part I want to hand to you...
(Because it is the whole lesson.)
When the machine is not doing what I want, my instinct used to be to argue with the machine.
Write a smarter prompt...
Explain it again... Only louder. 😳
That is like yelling at your car for going off road while your own hands are on the wheel!
The car did exactly what the wheel said...
I was steering and I did not notice I was steering.
What I actually did wrong was *tell it how to organize instead of telling it what I wanted***.**
I had a process that works.
I skipped it... Then spent a day paying for the skip.
The fix was not clever.
I went back to the process.
Plan first, in writing, with the steps in order, before anything gets built.
Boring. (And it works.)
If you take one thing from this...
When you are working with AI and the results are mush, check your own scaffolding before you blame the tool.
Nine times out of ten you skipped a step you know about.
That is good news, because your own steps are the part you can actually fix.
FYI.. We did manage to get a couple of things done so far for this week (both free):
Saturday School runs Saturday mornings, one hour, and the next three are booked.
We build something real from scratch, live, including the parts that go wrong.
Nothing to buy and nothing to install to watch.
And if you want to see the free LEAP pattern this is all built on, you can grab that here (or give the link to your favorite AI and ask it how LEAP can help you!).
github.com/tomi-mercer/leap-design-pattern
Thanks for being a part of AutomatedWith.tech!
— Mercer
If you were on the Nitwyt build (day 4) livestream today, you saw we didn't "prompt" AI.
We used Nitwyt OS to digest "input" (literally anything) and and turn it into a clear, actionable plan.
We didn't sit there writing complex prompts.
We didn't argue with a chat box or try to coax it into giving us what we wanted.
We simply dropped the raw thoughts into the workspace, and the environment (Nitwyt OS) did the rest.
That is what we mean by claiming your time back.
The biggest problem with AI right now isn't the AI itself...
It's the fact that you have to micromanage it.
The goal of Nitwyt is removing that friction.
You provide the 10% strategy/direction and the software handles the 90% execution/action.
Instead of treating AI like a tool you have to constantly direct, we're building an environment where it works in the background to organize your business, your content, and your life.
Thank you to everyone hanging out in the lives and supporting the build!
Day 3 of our 30 Wins in 30 Days sprint is officially over and on YouTube.
I won’t lie, there was a moment today where I was walking the tightrope live on stream.
I’ve been trying to get the Nitwyt OS connecting properly across devices so you can talk to it from wherever you are.
And of course, my first test of the new local connection gave me a fun little hiccup live on stream. 🤦
But we had a backup plan, and when we got the Nitwyt OS UI pulled up...
We tested the brand new "Foresight Preview."
The idea here is simple...
I don’t want to babysit an AI.
I want to be able to brain-dump my messy thoughts into the machine...
EVERYTHING.
From YouTube subscribers...
To the upcoming NDK product launch...
All the way down to planning date night and working on an aquarium project I've been meaning to get to for years.
Nitwyt took that total mess, thought about it, and instantly organized it into my four core areas...
Wealth, Signal, Life, and Craft.
(Not sure we'll keep all these titles as is yet, but the structure is pretty useful so far!)
No long, exhausting chat conversations (that I mostly end up losing anyway!).
Simply a clean, color-coded plan.
It’s what we call Casual Magic and it's the intentional by product of developing "AI first... Always human."
Speaking of...
We also tested forcing it to think from an "Outside-In" perspective... meaning it focuses on you, the audience, and what you need, instead of the internal plumbing that I'm personally more focused on.
Because we're seeing such cool progress, I am taking an extra day to polish the Nitwyt Developing Kit (NDK). The whole tech industry has pushed this idea of the "Minimum Viable Product" for way too long.
We shouldn't be rushing things out the door because they barely have a heartbeat.
I want this to be robust.
We're launching the NDK later this week, and it’s going to let you actually install the bleeding-edge system we're building right now.
Thank you to everyone who has subscribed and tuned in!
Seeing those numbers go up means the world to me...
Here's the upcoming schedule:
- Day 4: Thursday at 3:00 PM
- Day 5: Friday at 3:00 PM
- Weekend Streams: Saturday (June 6th) & Sunday (June 7th) at 10:00 AM
We’re claiming our time back, one win at a time.
See you tomorrow!
If you operate a business or organize complex workflows, you require software that works every time you open it.
You do not have time to play "developer."
Today, on the Day 2 livestream, we ensured the Nitwyt OS acts exactly as a seamless, reliable program designed for your as the AI Operator.
What We Solved Today:
-
Future-Proof Reliability
Tech giants constantly update and delete their AI models, frequently breaking the tools you rely on. We hardcoded a new detection system into your local OS. When a provider deprecates their systems, your OS automatically adapts in the background. Your workflow never stops. -
Software for Humans, Not Hackers
We completely overhauled the desktop dashboard. We stripped out all the confusing, sci-fi AI jargon that developers rely on. Your portal is now in plain, functional English: "Ask Nitwyt," "On Deck Calendar," "Factory," and "Showroom." It operates like the professional software you already know how to use.
The NDK Drops Tomorrow
We have officially packaged the fully automated version of this software.
If you are tired of piecing together random AI tools and need a central operating system for your business:
The Nitwyt Developing Kit (NDK) launches tomorrow during Win #3 at 3 PM Central...
As always, thanks for being AI first... Always human.
-- Mercer
Actual Win #1 (from yesterday)...
We established the [LEAP Design Pattern, pushed the blueprint to GitHub](<https://github.com/tomi-mercer/leap-design-pattern >), and built the visual UI shell. The switchboard exists, but it has no brain.
Today, we face the reality of local AI maintenance.
Yesterday our local bridge failed.
Today, we know the legacy Gemini models we targeted have been publicly deprecated by Google.
If you aren't doing the research, your systems break....
Forecasted Win #2...
Today, we secure the API Bridge. We will establish a functional local server, update our payload to gemini-3.0-flash, and successfully return a response to the UI.
- For the Builders: The raw blueprint is in the [open LEAP repo](<https://github.com/tomi-mercer/leap-design-pattern >)
- For the Operators: This friction is exactly why we are building the NDK (Nitwyt Developing Kit). But the NDK doesn't exist yet, so today, we do the manual labor.
For those that are exploring AI... see you at 3:00 PM Central on YouTube!
Howdy!
Day 1 of the "30 Wins in 30 Days" sequence is officially in the books...
Today's goal was all about getting into the core of Nitwyt OS and wiring up the "Ask Nitwyt" switchboard interface.
Here’s what went down during today's live build...
1. The LEAP Design Pattern (Open Source) is Live!
We officially pushed the core logic structure for how to prompt AI effectively using logical containment (Nitwyt Prompting). You can grab the blueprint for yourself right here on GitHub:
👉 https://github.com/tomi-mercer/leap-design-pattern
2. The Live Build Replay
We dove into VS Code to build out a local Node.js server bridge to connect the HTML UI to the Gemini API. If you want to see the behind-the-scenes mechanics of how we use AI to build AI logic, check out the replay:
👉 https://youtube.com/live/pbRvZYzHFdc
3. The Cliffhanger (The API Bridge)
Of course, in classic live-demo fashion, we hit a snag! The HTML portal was trying to fetch from an offline localhost server, and a Gemini model string was slightly out of date. Instead of rushing it on stream, we are using this exact breakdown as the foundation for tomorrow's build.
What's on Deck for Day 2 (Tomorrow)?
Tomorrow, we are picking up right where we left off. We are going to lock in that API bridge connection, prove the local fetch works, and then update the Nitwyt Developing Kit (NDK) logic.
See you all in the morning for the Day 2 drop!
-- Mercer
We're back.
Starting today, we launch 30 Wins in 30 Days.
Every morning, I’m dropping the "Win" for the day.
Every afternoon, we're building it live.
Today's Win is establishing the Switchboard Architecture.
I just uploaded the morning reveal video showing the "Nitwyt Private Desktop Portal" UI.
The Afternoon Build (3PM Central)
We are wiring the brain to the UI live.
No hustle... Simple logic you can use with any AI.
See you in the workshop: