A couple of days ago, I went to the milk tea shop downstairs and noticed an interesting detail. Many shops no longer actively greet customers.
You scan, place an order, pay, earn points, and use coupons. The whole process is automated by the system.
I suddenly realized: the internet has been doing the same thing for years — turning 'human behavior' into computable data.
What are you ordering?
What time are you coming?
What flavor do you like?
How long will you stay?
How often do you reorder?
These assets will eventually transform into platform assets.
But the problem is: after the platform makes money, the people who actually provide this data get almost nothing.
The more I dive into AI, the more I realize this issue is going to get worse. Because in the AI era, what's truly valuable is no longer just the code. It's 'human behavior itself.'
Your one sentence.
One click.
One edit.
One judgment.
All could become model training material.
This is also why I've recently started paying attention to @OpenLedger again. Because at least they're seriously tackling a tough question that many projects dare not touch: should data contributions be priced?
In the past, many AI projects loved to hype their model parameters.
How many B.
How many Tokens.
How fast is the inference.
But OpenLedger is currently heading in a direction that feels more like researching 'labor relations in the AI economy.' Especially with their current focus on Proof of Attribution. Many just see it as 'data attribution.'
But the more I look at it, the more I think: they're actually trying to establish a new form of 'digital piecework pay.'
Who contributes.
How much was contributed.
Which data was called.
Which contents really influence model outputs.
Complete re-accounting. This is actually a particularly sensitive matter.
Because the default logic of the past internet has always been: the platform owns the data, and users are merely 'consumers.'
But OpenLedger now wants to shift to a model where users are both consumers and producers. And theoretically, they can also share in the profits. If this logic actually works, it could be truly disruptive.
Because the biggest battle in AI's future might not be a model war, but a 'high-quality data supply war.'
Many still haven't realized: models are getting cheaper. Open-source models are proliferating. The cost of computing power is also gradually decreasing.
But the truly scarce resource is starting to become: long-term, real, continuously updated human behavior data. Especially in niche areas.
For example: healthcare, law, finance, trading behavior, and specialized knowledge. These things can't just be solved by scraping a few web pages.
What OpenLedger is doing with Datanet is essentially laying the groundwork for this. They want to untangle the data contribution networks from different fields. Then use PoA to track value flows. Honestly, I actually agree with this direction.
Because AI will inevitably face a question: what if no one is willing to contribute high-quality data? In the past, internet platforms could rely on monopolizing traffic to get by.
But if the value of data keeps rising, contributors will start negotiating terms in reverse. At that point, 'who allocates more fairly' will become a core competitive advantage.
Of course, I'm not just mindlessly hyping it up. The issues OpenLedger faces are pretty apparent.
First is cold start. Without enough calling volume, data contribution earnings simply won't take off.
Secondly, there's data pollution. Wherever there are incentives, there will always be people churning out junk content.
Thirdly, governance. Many on-chain governance models eventually drift towards 'big players deciding everything.' These issues haven't really been solved yet.
At least the issues they're discussing now are ones the AI industry will inevitably face.
Instead of just rolling out a few Agent Demos every day and shouting revolution.
I'm increasingly convinced that the most valuable resource in the future AI industry won't be GPUs.
It's about those willing to contribute real behavior over the long term. Whoever can rework this value distribution is the one who might truly reap the big rewards of the AI era.
At least so far, @OpenLedger has started to move down this path.