Recently, Vitalikhas once again targeted AIIn the era, Ethereum has made a series of remarks on the role it should play. Among them, the core views that I care about are:

- AIshould not speed up blindly

- The value of Ethereum lies in providing AIProvide verifiable, decentralized direction choices

- Make better markets and governance truly feasible

This article generally inherits VitalikConsistent thoughts and logic, unswervingly insisting that Ethereum should be decentralized based on technology in the AIera. text="">AIThe ideal that puts "shackles" on to prevent loss of control.

I very much recognize and agree with these views.

But during this period, due to a series of new events and situations, I have higher expectations and greater expectations for the future development of Ethereum.

Last two weeks, due to the development of AI, Wall Street conducted an almost irrational "bloodbath" on the stocks of software companies. The common view on the Internet is that:

In theAIera, software will become obsolete and software companies will be eliminated,AIwill reshape everything about software.

This view seems a bit exaggerated, but this time, I think Wall Street's view has some truth and is worthy of serious consideration.

All the software we use in modern society is created for humans---User experience, interface design, usage habits, calling methods, etc. are all ultimately faced by humans.

But now, we have seen a new trend: After the emergence of Clawdbot/Molt, do humans still need to deal directly with software? It seems that it is still needed at present, but this demand will definitely decrease exponentially in the next period, and eventually in the future society, the vast majority of "software" will be used by AI Agents instead of humans.

text="">AI AgentsThe software they call and the way they use the software can be completely different from humans. They can be more efficient and subversive.

This means that most of the software we see today is very likely to be refactored, subverted, and eliminated in the future.

This is what is and will happen in the software field.

Such things are also fissioning and catalyzing in other fields.

Comparing the past two daysVirtualandMoltIn the article, I vaguely feltVirtualThere are similar problems, according to the original words of the article:

"We look at the world from God's perspective and know that countries will definitely appear in this world in the future, so we first build the country's structure and mechanism according to our own ideas, and then immigrate to this country, or attract people to this country"

The biggest potential problem with this approach is: Do the structures and mechanisms we build in this country really meetAI Agentsdemands? Are they really the most adaptable to the growth of AI Agents?

It is precisely because of the continuous accumulation of these ideas that I thought of leaf="">ERC-8004, thought ofX402And a series of verification mechanisms and payment mechanisms designed by us humans forAI Agents......

Then a question appeared:

These protocols and mechanisms are all imagined by us humans according to our own ideas, but are they really suitableAI Agents? Are they really the most suitable form for AI Agents?

Will some inconspicuous protocols and mechanisms appearing in a corner, in a community that we completely ignore, end up being our favorite, most native, and most suitable?

I don’t have an answer right now.

But I always feel that this exploration process may not be too smooth.

Duan Yongping emphasized two sentences in his investment Q&A:

Do the right thing, do it right.

According to this logic, VitalikI firmly believe that Ethereum and the Ethereum he leads are doing the right thing, but are they doing the right thing now?

This is something I am increasingly uncertain about now, becauseAIThe changes that the development of AI has brought to us are too fast, too disruptive, and too unpredictable.

However, the correct exploration and practice in any field is this process: make sure to do the right thing first---Use encryption technology to add boundaries and constraints to the development of AI, and then explore how to do these things right in practice.

If Ethereum is to be realized in theAIeraVitalik's ideals and goals may require more frequent iterations and closer attention and tracking of AIin the future.

This is a more ambitious and long-term goal and task.