There's a conclusion I've come to recently.
When it comes to talking about AI, the conversations always seem to find the same conclusion. They converge on a discussion about more powerful models, more efficient structures, larger and better-distributed datasets, models with the highest system intelligence, etc.
These are plausible.
These are valid.
But the more time I spend trying to figure out what is happening around OpenLedger, the more I am convinced we are looking at the “front” of the machines while neglecting everything going on behind them.
I actually might be wrong.
This is where my head has been lately.
When humans interact with an AI system, what they see are only “results.”
This appears on a screen.
This is what we judge the system on.
It can be a good/bad answer.
It can be a correct/incorrect answer.
It can be a useful answer.
It can be a simple answer.
But where did the answer come from?
This is everything that exists before the machine learning model.
Everybody who contributed the information.
Everybody who curated it.
Everybody who validated it.
Everybody who contributed information that built on top of existing information.
They are all invisible.
I am actually guilty of this as well, and I never really thought about this before.
I used to think this was just how AI systems worked.
Not so sure I can say that now.
The best analogy I have is this.
Getting an answer from a machine learning system is like going to the supermarket and picking an item off a shelf.
To you, all you see is the product.
Everything that came before it being put on the shelf was the factory and the delivery trucks.
You can't see the warehouses. You can't see the people who brought it to you.
You can see the product.
You can't see the process.
Al is becoming something similar. OpenLedger catches my attention for different reasons. Not because I'm sure it's a perfect solution. Every project mentions solving attribution. Reality is messier than that. What I find intriguing is where the conversation is heading. It's forcing people to grapple with a question Al has not done a good job with:
Who has really created value? Not in terms of the platform owner. Not the one who created the last step. Who did the value adding? Who did the value creating? It is quite possible that the same people who created the most value along the process were the most invisible throughout the process.
Some things become highly visible because their contribution is easier to see, and that is not a critique. It is simply how systems work. When systems try to track inputs, they can only record what is visible. Once I started considering this, I started recognizing the same pattern absolutely everywhere.
Social media. Ranking systems. Search algorithms. Everything works this way. The things that are visible tend to become more visible, and the things that are not visible start to fade from the narrative. This is not because there is no significance attached to the things that are invisible; it is simply because they are difficult to track.
Al believes that attribution will become a greater challenge for everyone in the industry. As the industry evolves, the need for verification will increase. So will the need for trust. As such, the issue is no longer just whether information is available.
Now we must consider whether the system will even know information is available. That's a much different issue, and maybe that's why this topic constantly draws me back. We're about to cross into the reality where information is nearly infinite. The challenge will be verifying the sources of information, contributors, and the acknowledgments, as well as omissions. I really don't believe this is being discussed enough. Unfortunately, the discussion is mostly about constructing more sophisticated models.
Meanwhile there is a whole layer, beneath that is quietly becoming just as important.
Maybe more important.
Because intelligence alone is not enough.
People want to know where value comes from and who helps create it.
They want to be able to trust.
They want transparency.
I think that's what OpenLedger keeps making me think about.
Not how smart AI can become,
Something much simpler.
When value is created, who gets. Who gets forgotten?
Honestly I think that question might matter more than people realize today.