When AI becomes the new consensus, what is left of Web3?

A late-night discussion about industry migration, talent re-evaluation, and long-term thinking.

On the evening of March 26, a livestream that lasted for more than an hour finally put into clear words what many people had been feeling but couldn’t quite explain.

Over the past year, almost everyone has felt the same thing: attention is migrating from Web3 to AI.

Last year, there were still huge numbers of people discussing switching to Web3, talking about public chains, narratives, and project opportunities; but this year, industry buzzwords, event hubs, hiring requirements, and even entrepreneurs’ self-introductions are increasingly being redefined by AI. Web3 hasn’t disappeared, but it’s clearly no longer the place that single-handedly hogs the spotlight.

The most important value in the livestream isn’t a simple sigh of “the trend has changed,” but that it points to a deeper reality: Web3 is shrinking back from a high-hype narrative industry into a tougher, more bottom-layer, more compliant, more infrastructure-driven industry.

That means the real barriers for the industry have changed, too.

1. What disappears first isn’t the industry—it’s the illusion

In the discussion, there’s a very real judgment: last year, many people wanted to enter Web3—even people from traditional industries were rushing in; but this year, the market is like a sieve, filtering out followers. What remains may not be the people who are best at telling stories, but the ones with clear goals, who are willing to invest long-term, and who can continuously adapt to changes in the environment.

The livestream mentioned that people who are truly willing to pay for the transition, who are willing to learn systematically, and who are willing to rebuild their capabilities come more often from backgrounds like funds, banks, and big tech.

Behind this, it’s not simply a “bull-bear cycle switch,” but a more typical signal of industry maturity.

In the early stage, the industry is easiest to attract people with strong expectations for financial leapfrogging; in the mature stage, the industry needs people who can deal with complex reality. The former pursues opportunity density; the latter handles systemic friction. Once an industry moves from “telling new stories” to “delivering real work,” its requirements for people rise immediately: no longer only looking at enthusiasm, no longer only looking at insider identity, no longer only looking at whether you can memorize a few sets of terms—but whether you can keep working, keep learning, and keep producing results in an uncertain environment.

In this sense, what AI is taking isn’t Web3’s future, but the overly bubble-like attention Web3 had in the past.

2. Web3 is becoming an “infrastructure industry” again

A trend mentioned in the livestream is especially key: the strategic narratives of several major public chains have already clearly shifted. For example, Solana’s direction is more focused on payments and enterprise-level financial rules; BNB places more emphasis on AI, payments, privacy, and RWA, and is trying to move closer to Web2 user experience; TON is putting attention on a payments ecosystem and AI agent ecosystems; and some platforms that have obtained compliance licenses are starting to move toward institutional-grade tokenization.

In other words, the industry is shifting from the old narrative of “anonymity, tax avoidance, wealth myths” toward these heavier themes: “payment infrastructure, enterprise services, institutional compliance, and assets on-chain.”

This is a signal that’s well worth paying attention to.

Because it means that the place where Web3 will truly crystallize value in the future won’t be the hottest emotional area—it will be the most boring one. Not the place where people shout slogans the loudest, but the place where compliance, payments, data flows, asset mapping, enterprise coordination, and global implementation can be handled.

Many people will feel that this makes the industry “less sexy.” But the opposite is true: this is what it looks like when the industry starts getting truly close to mainstream economic structures. Only when an industry moves from peripheral narratives into the institutional interfaces does it have a chance to obtain longer-cycle capital, customers, and talent.

The livestream also mentioned that capital is flowing to more compliant regions, and practitioners are flowing with it—some people shifting from Dubai to places like Hong Kong and the U.S. This kind of migration itself shows that the industry’s focus is no longer only on where things are freer, but on where things are more sustainable.

So when looking at Web3 today, you can’t only view it from the perspective of “projects inside the circle.” You have to start looking at it through a perspective closer to infrastructure research: Who is handling real-world enterprise needs? Who is handling payments and clearing/settlement? Who is handling compliance and issuance? Who is handling real adoption at the institutional end?

That’s the next stage’s more valuable capability layer.

3. AI isn’t “another track”—it’s the new bottom-layer capability for people working in Web3

One of the most piercing lines from the livestream is: when companies hire people now, it’s not Web3 or AI—it’s Web3 + AI.

This is a very important structural shift.

In the past, Web3 entry barriers mainly came from technical understanding, industry jargon, resource networks, and information asymmetry. Today those still matter, but they’re no longer enough. Companies now require candidates not only to understand Web3, but also to be able to use AI for marketing, research, data analysis, content, product demos—sometimes even to complete a task on-site directly. In other words, AI is no longer a “nice-to-have”; in many roles it has become a new baseline literacy.

This will directly rewrite the industry’s talent structure.

The first roles to be hit are those with clear processes, delivery that’s templated, and value that’s not close enough to business outcomes. For example, basic content roles, parts of entry-level operations roles, and pure text-based customer acquisition roles will all be greatly compressed by AI. The livestream explicitly mentioned that some basic copywriting output jobs may be replaced by digital employees.

But some other roles become more valuable precisely because of AI. Especially roles involving complex judgment, deep trust, cross-boundary coordination, and closed-loop outcomes. For example, in the livestream’s discussion of BD: if it’s only sending messages, coordinating with lists, and doing mechanical outreach, that kind of BD is easy to replace; but if it’s about building deep trust with customers, understanding the customer’s real problems, and推进 complex collaborative conversion, then that kind of BD is harder to replace.

The logic behind it is simple: AI will compress the value of “information transportation,” but it will amplify the value of “relationship judgment” and “complex delivery.”

So in the AI era, what makes a person truly valuable isn’t how much information you know—it’s whether you can integrate information, tools, relationships, and judgment into results.

4. What companies truly want isn’t “people who know how,” but “people who can grow bigger”

The livestream had an observation that was especially accurate: companies now hire people increasingly not just by looking at resumes, but by valuing curiosity, learning ability, and the ability to think on the spot. They even test a person’s real capabilities through fast Q&A, giving questions on the spot, requiring them to produce plans quickly, and so on.

This actually reveals a harsh reality: in an industry changing too fast, companies no longer trust static proof.

Education, work history, and past titles still matter, of course, but their reference value is declining. Because tools are changing, the market is changing, narratives are changing, and even the roles themselves are changing. Today a person’s core competitiveness isn’t “what I already know how to do,” but “whether I can quickly learn what I will need to know next.”

From a sociological perspective, this means the logic of how the labor market evaluates people is shifting toward “dynamic adaptability.” Companies aren’t looking for the most standard person for some fixed job function; they’re looking for people who can understand the business across modules, who can enhance themselves with AI, and who can still maintain momentum and action in instability.

The livestream even mentioned that rotation systems, training for comprehensive capabilities, and a “deal-maker/operator mindset” are becoming part of talent development.

Behind this is actually a change in organizational form: companies don’t need large numbers of people who only complete single-point actions; they need people who can complete a small closed loop together with AI.

That’s also why “long-term thinking” becomes important again today. Not because those three words sound correct, but because only long-term thinkers are willing to keep learning, iterating, and getting stronger in an environment that keeps changing.

5. Do newcomers still have opportunities? Yes—but not by relying on fantasy

In the livestream, the answer to whether “newbies still have a chance to enter Web3” was relatively clear: there is an opportunity, but it requires systematic learning, understanding the industry map first, then migrating your existing professional capabilities, and then entering the stages of practice and interviews. It’s not “zero-fund beginners can get rich”; it’s “zero-fund beginners can switch, but they must seriously rebuild their capabilities.”

This is an important reality check.

Today’s Web3 no longer allows many people to just rush in based on enthusiasm and impulse the way it used to. It’s more like a cross-disciplinary industry: you need to understand technology, and you need to understand the market; you need to understand products, and you need to understand narratives; you need to be able to use tools, and you need to know where real business is happening.

For newcomers, the most effective path isn’t to ask right away “which project will go up,” but to ask first: What problems can I solve for this industry? What existing skills can I migrate into it? Can I turn my learning process into visible assets?

The livestream mentioned that building social accounts, continuously outputting insights, analyzing the industry, and showcasing real experience with agents or projects you’ve actually participated in can all increase your chances of entering the industry.

This actually aligns with today’s job-hunting logic: a resume is the past, work is the present, and consistent expression is the future.

6. The asset most underrated in the AI era isn’t prompts—it’s training data (corpus)

One viewpoint from the livestream that I think has the most staying power is about “accumulating a corpus.”

The discussion mentioned that one very worthwhile thing to do in the future is to record your everyday conversations, your work process, the content you learn in classes, and your real thinking—and continuously feed it to AI so that AI can form contextual connections and gradually grow something like a “soul.”

That sounds a bit emotional, but underneath it’s very real: in the AI era, what’s most scarce isn’t the model, but high-quality, continuous, true personal and organizational context.

Many people are competing over tools, models, and who can connect to new products first. But what may have long-term value are those least noticeable accumulations: how you judge problems every day, how you speak, how you make decisions, in what situations you change your views, and how you turn scattered information into your own methodology.

Once these things are continuously recorded, they won’t just be “materials”—they’ll gradually become a new kind of means of production. The livestream said that even if these real corpora can’t be seen as commercial value right now, they may爆发 commercial value in the future. I strongly agree with that judgment.

Because future competition may very likely not be about who gets AI first—it will be about who has a self that AI can understand deeply enough.

7. Who is truly worth betting on today?

If I compress the whole livestream into one answer, I’d say:

In the AI wave, Web3 isn’t over. It’s just changed from “hot narratives anyone can talk about” into a “hard industry where real capabilities are required to enter.”

And within this industry, the people truly worth betting on probably have a few characteristics:

• First, willing to keep learning instead of consuming anxiety.

• Second, able to treat AI as a workflow capability, not a shiny new toy.

• Third, understand that Web3 is moving toward payments, compliance, RWA, and enterprise-level infrastructure—not still stuck in old-cycle illusion.

• Fourth, have real delivery capability, able to independently complete a business closed loop.

• Fifth, can build trust, move forward complex relationships, and understand customers’ real needs.

• Sixth, willing to do things long-term rather than always searching for the fastest shortcut.

In the end, what the era truly rewards is never the person who chases hotspots the hardest, but the person who keeps growing through shifts in what’s hot.

AI will rewrite many roles. Web3 will eliminate many illusions. But after these two waves stack together, it will actually make some people stand out even more—those who understand both technology and people; those who can call on tools and build trust; those who can see structural changes and are willing to deeply build a niche for the long haul.

They may not be the loudest. But they very likely will be the core assets of the next stage.