Transcript: AI-written software, patents, and the rise of care work
An AI-generated audio overview (Google Gemini), transcribed by machine on Blaine's Mac (Whisper). The audio was checked against primary sources; what it gets wrong is listed at https://blainomd.com/learn#audio-headless-autonomous-workflows

Imagine a world where, you know, the brilliant software engineers who built your favorite app are suddenly just out of a job.
Right. Yeah.
But at the same time, the person making your grandmother's bed is suddenly the most sought after asset in the global economy.
Welcome to the deep dive.
It is quite the shift. Yeah.
Today, we have a mission for you.
We are unpacking this dense but honestly fascinating strategic report about how AI is destroying traditional software ownership.
And completely flipping the global economy upside down.
Exactly. And surprisingly, it's actually saving the human care industry.
So to frame this shift, think of it like a restaurant where suddenly the secret recipes are totally free for everyone.
Yeah. Anyone can just make the food.
Right. So the only thing of actual value left is that invisible manager, the one who flawlessly orchestrates the kitchen and the servers.
The report calls this a shift to headless autonomous workflows.
Right. Because for decades, those recipes, I mean, the source code, they were written by humans.
And that meant you got automatic copyright protection.
Which is a huge deal for software companies.
Oh, absolutely. But now we are seeing this thing called vibe coding.
Vibe coding, right?
Yeah. Where an AI just generates the entire code base from like a natural language prompt.
And without a human author, that automatic copyright just vanishes completely.
So I know companies tried pivoting to trade secrets, right?
Like to protect the harness, the instructions managing the AI.
They did. But trade secrets are just incredibly fragile.
You just have to look at that massive anthropic leak back in March 2026.
Oh, wow. Yeah.
Yeah.
That was a huge deal.
Yeah. Human error accidentally exposed like 500,000 lines of their clawed code harness.
And once it's out in the wild, the legal protection is permanently gone.
Wait. OK. I'm genuinely confused about something here, though.
The report mentions dark code and this particularity paradox.
Right. The particularity paradox.
Yeah. So if my company's AI invents this brilliant new routing system, right, and it's operating securely on my servers, why can't I just sue someone if they steal it?
Well, why does it matter if my own engineers can't explain exactly how the AI came up with it?
Because the courts require you to describe a stolen trade secret with, quote, sufficient particularity.
Oh, I see.
Right. So if your engineers cannot explain the AI's emergent black box reasoning, like how the code actually works, you legally just cannot enforce the secret.
That is wild.
It is. And it's exactly why the smartest companies are pivoting to utility patents now, specifically under the Bascom legal standard.
OK. Let's unpack that. What does the Bascom standard actually do for them?
So it allows you to patent a specific, unconventional arrangement of steps, right, even if the individual steps themselves aren't patentable.
OK. So you aren't patenting the AI's dark code then.
Exactly. You are patenting the deterministic orchestration backbone.
Going back to your restaurant analogy, you basically patent the manager's specific system for running the restaurant.
Not the uncopyrightable recipes.
Exactly. Not the recipes.
OK. So if the real value is just in the orchestration, the manager, and the code itself is basically free, the whole way we buy software has to change, right?
It's like the death of the seat license.
Totally. We are rapidly transitioning from software as a service to service as a software or sasses.
Service as a software, right?
Right. So we are moving from buying hammers for our workers to just, like, buying the finished house.
That is a perfect way to put it. Vendors no longer sell you a tool to do the work. They sell the completed, verified outcome.
So no more buying subscriptions to a customer support platform.
Right. You just pay a fraction of a cent per perfectly resolved customer ticket.
But wait. If buyers are only paying for outcomes, and AI achieves those outcomes for literal pennies, what happens to the overall size of the market?
The money-changing hands has to plummet.
It does. We get what's called a total addressable market, or TAM, deflation paradox.
AI reduces the cost of delivering cognitive work by about 90 percent.
Ninety percent.
Yeah. And competitive pricing forces everyone to pass those savings on.
So a 12 trillion euro global services market could shrink to just 2 trillion euros.
Just because cognitive work is now incredibly cheap to execute.
That much money evaporating sounds like a recipe for massive job panic, honestly.
It does. But when you look at the 2025 and 2026 payroll data in our sources, the fallout isn't hitting everyone equally.
Right.
Early career workers in highly exposed cognitive roles, you know, like software engineering and customer support, they saw a 19 percent drop in employment.
But the countertrend is stunning.
Because employment for young care workers like home health, nursing, psychiatric aides, it's booming.
It is absolutely booming.
And the mechanism behind this is just fascinating.
It's because AI is finally automating the bloated administrative shell of health care.
Yes, exactly.
We're talking about natural language AI agents automatically transcribing patient visits, instantly generating those complex billing codes, and negotiating directly with insurance APIs.
Which completely eliminates the need for massive back office billing departments.
And when you demolish that incredible administrative overhead using headless AI, you free up massive amounts of capital.
Right.
Which brings us to those new models mentioned in the report, like Vimte and Coop.care.
Yeah, because they have virtually zero back office costs.
These local co-ops can charge a family just $35 an hour for elder care, but they pay the W-2 caregiver a guaranteed $26 an hour.
Which is unheard of.
It's literally having industry turnover.
By stripping away all that digital busy work, they elevate embodied human presence as the ultimate scarce asset.
Exactly.
We are officially in a post-digital economy.
Routine cognitive execution is simply too cheap to meter.
So, as AI drives the cost of digital tasks to practically zero, the future doesn't belong to the machines.
It belongs to accountable, physical human presence.
It really does.
Which brings us right back to our restaurant.
When the recipes are free and the kitchen runs itself, you will still pay for the human warmth of a great server, right?
Or the shared experience of a meal with friends.
Absolutely.
So, here's a thought to leave you with today as you go about your week.
In a post-digital economy where routine brain work is virtually free, what uniquely human, uncopyable value are you bringing to your life and work that a machine can never replicate?