In the AI Era, the Only Asset You Take With You Is Reputation

Some recent thoughts, shared here for discussion.

1. When you switch jobs, what can you actually take with you?

I’ve had a few thoughts about the AI era lately, and I want to gather and record them here.

Start with a scene most of us have lived through: changing jobs. At your old company you might have done really well — you had a title, resources, a network, status. But once you leave for the next job, how much of that actually comes with you?

Think it through, and almost none of it does. The one thing that truly follows you is this: what other people think of you. Your reputation.

That was the first thing that struck me.

The second was watching what the latest AI can do with code. Someone used a state-of-the-art model to rebuild the classic game Red Alert in a single night, after hundreds of rounds of testing and debugging. AI’s ability to get things done — to write code — has already come a very long way, and it will only get stronger.

So in the AI era, what is genuinely valuable about us as individuals? My answer is that one word: reputation.

2. Why reputation?

To answer that, look back at what the internet era actually changed, and what was most valuable in it.

The most valuable companies of the internet era basically did three things:

Connected people with goods — Amazon, Taobao: wildly successful e-commerce companies.
Connected people with information — Google and Baidu.
Connected people with people — Tencent and Facebook (Meta).

These three kinds of companies became the most valuable of the internet era.

What about the AI era? AI will eventually replace these platforms themselves. In the future, people will connect to goods, to other people, and to information through AI. When we tell an AI “I’m looking for this kind of thing, this kind of person, this kind of information,” what it hands back to us depends on your reputation inside the AI’s world.

When an employer needs someone to do a job, whether that person has a reputation decides whether the AI will recommend them.

In the future, skills themselves will mean less and less, because AI can handle most knowledge-type skills, and the value a person adds there is very limited. At that point, a person’s value lies in their reputation.

I’ve also noticed that a lot of very smart, very forward-looking people — people who were already extremely successful in their careers — are now writing furiously on X, blogging, making videos. Many of them clearly don’t need to, and yet they’re out there doing it.

3. “Playing to the crowd” isn’t a dirty word — it’s how the world works

There’s a belief I used to take for granted: that “playing to the crowd,” seeking the spotlight, making yourself seen, is somehow undignified. I’m increasingly skeptical of that. The truth may be this: if you have no reputation, you are essentially invisible in this world.

In the AI era, if the AI doesn’t know you — if you’ve left no trace in the AI’s world — then it will never recommend you.

Picture a scene from the internet era we’ve all experienced: you want to buy something, but it can’t be found on any website anywhere online. Would you trust it? Hardly.

That was the lesson of the internet era. For every product, every company, even for ourselves as individuals, we went and built a profile on LinkedIn. In the AI era, this becomes compulsory — you have no choice but to do it.

How do you build it? Leave your track on the internet. Leave traces. Whatever you’re doing, if you think it has value, find a way to make it stick. Writing on X, making videos — all of it helps.

To be more concrete, I think reputation breaks into three layers:

Identity — who you are, recognizable and traceable by AI;
Record — what you’ve done, with dates and details;
Verification — third-party proof: customer reviews, real transactions, peer endorsements.

Of these three, the third is worth the most. Because in the AI era, anyone can mass-produce the first two — AI drives the cost of content production toward zero. When the internet is flooded with AI-generated “traces,” AI will lean even harder on the signals that are hard to fake. So leaving traces isn’t about quantity; it’s about leaving verifiable traces: real work, real customers, a consistent record across a long span of time.

Sure, you could choose to be a handyman, a plumber — the kind of work that genuinely requires a human body. But how long does that hold? As AI keeps expanding its reach into knowledge, it will move into the physical world too — maybe in ten or twenty years, maybe sooner. By then, how much those jobs are worth is anyone’s guess. But whatever you do, people in the future will use AI to find everything they need. At that point, all of our value shows up as reputation. Without reputation, none of it means anything.

4. Education in the AI era: what should we actually accumulate?

Now to education — or really, self-education: what should we actually learn?

I saw a lot of discussion recently arguing that university education has lost its meaning in the AI era. Looking back, Europe’s earliest universities were founded to educate monks, in an age when knowledge was monopolized. By the industrial era the nature of universities had changed — they were mostly training technical workers, blue-collar or white-collar. Today’s universities mainly train white-collar skills.

We all know AI can replace white-collar work to a huge and growing degree. Which means university education itself starts to lose its point.

The moment anyone can use AI, it’s as if they’re working with a team of world-class advisors at their side. At that point, spending many years learning one knowledge-type skill is an investment with a very, very low return.

So what should we really accumulate in the AI era? I still think it’s reputation.

But reputation isn’t built by doing something today and having it appear tomorrow — it takes time to accumulate. AI is smart: if a person (or a website) has had no presence online for the past ten years and only just popped up this year, then in the AI’s eyes they’re a newcomer with no record, and it’s hard to prove they’re trustworthy. But if they have a record of work every year, every month, with reasonably detailed descriptions that prove a good track record in some field or skill — then they have a good reputation.

Back to that idea of “playing to the crowd”: when you’ve left no trace in this world, you’re a nobody. Even leaving some bad reputation is still reputation — it’s still your name out there, and a slightly bad reputation beats no reputation at all.

So my conclusion: the most valuable thing in the AI era is reputation, and we should start accumulating it right now, immediately.

5. A startup direction: a profile for AI to read

Finally, the startup direction I think is most valuable in the AI era.

The internet era had something very important called SEO — search engine optimization — the study of how to get search engines to surface your company, your product, you.

The most valuable thing now is the equivalent: making AI able to find you — you the person, your company, your business.

At its core this is still reputation: record what we do as completely as possible, and collect as much feedback as possible. Then, when an AI is looking for relevant information, products, or companies, it can do a good job of picking us. The more people use and see what we make, the better our reputation, and the more valuable we become.

How? Back to the opening point: accumulate our own reputation, and then help our customers accumulate theirs — help individuals and companies build an AI-era profile, automatically turning it into a “Facebook for AI to read,” a “search engine for AI to read,” a “shopping site for AI to read.”

There’s one key distinction here. If all you do is mass-produce content for people and ghost-write their posts, the barrier is low and it turns into a red ocean fast. The real moat is that third layer — the verification layer: structuring a person’s or company’s real work record, customer feedback, and transaction history into a format that AI can read and trust. By analogy, the thing to build isn’t “advertising for AI to read,” it’s a “credit report for AI to read.” Anyone can buy ads; a credit record can’t be faked — and that’s what’s genuinely scarce in the AI era.

This is what helps people connect, in the AI era, to the goods, information, and other people I mentioned earlier. I think it’s the startup direction with the most potential and the most room in the years ahead.

Will the incumbent internet giants get replaced? Hard to say — I think it’s possible. Even if they survive into the AI era, they’ll look quite different. And there’s a very real chance an entirely new company emerges to replace the ones that were so successful in the internet era.

A note at the end

Walking the talk: this article is itself a step on my own road to accumulating reputation.

From having these thoughts, to using AI to help me organize and publish them — the process itself is me practicing what I’m describing: turning valuable thinking into a dated, traceable record left on the internet.

These are my personal thoughts — discussion and pushback welcome.