Claude Code 之后 Anthropic ตัวต่อไปจะกลายเป็นผลิตภัณฑ์爆款อะไร?

Video Title: Anthropic's Hunt to Find the Next Claude Code

Video Creator: ACCESS Podcast

Translation: Peggy, BlockBeats

Editor's Note: As large model capabilities continue to advance and AI coding tools rapidly proliferate, the industry discussion is shifting from "whether models can complete tasks" to "how model capabilities can be organized into products, workflows, and business systems."

Over the past year, products like Claude Code, Codex, and Co-work have entered the workflows of developers and knowledge workers. AI is no longer just a chatbot for answering questions; it is becoming a production-level interface capable of calling tools, executing tasks, and verifying results. But while the consensus that "agents will become the next software paradigm" is solidifying, a more critical question is emerging: Who will be the first to transform model capabilities into reusable, distributable, and scalable work systems?

This article is compiled from an interview with Mike Krieger on the ACCESS Podcast. Mike Krieger is the co-founder of Instagram and currently serves as the Chief Product Officer at Anthropic, leading Anthropic Labs. His goal is to guide the team in exploring Anthropic's next frontier product directions after Claude Code.

Alex Heath (left) and Mike Krieger (right)

In this conversation, Mike doesn't just speculate on Anthropic's next product. Instead, he deconstructs the AI product competition into a set of more fundamental structural issues: how model capabilities enter real workflows, how AI companies organize innovation internally, how platform companies manage boundaries with ecosystem customers, and where human judgment will be repositioned in the production chain as AI execution power grows stronger.

First, the product paradigm is shifting from "chat" to "tasks." Previously, large models primarily existed as dialog boxes—users input prompts, models generate responses. Now, Claude Code, Co-work, and Claude Design represent a different product logic: enabling AI to work persistently toward a goal, calling tools, generating results, and performing verification along the way. This means the key to AI products is no longer just the quality of answers, but the ability to decompose tasks, maintain context continuity, call tools, and verify results. Whoever can package these capabilities into a seamless workflow will be closer to the next generation productivity gateway.

Second, the organizational approach is shifting from "big team planning" to "small team experimentation." Anthropic Labs operates more like an entrepreneurial unit embedded within a large company: starting with two or three people, reviewing progress every two weeks, and using high-frequency feedback to decide whether to continue a project. Historically, innovation labs in large companies have suffered from long cycles, ambiguous responsibilities, and projects being dragged out on "good enough" ideas. Now, models have lowered the cost of building, making judgment, taste, and decision-making speed the truly scarce resources. This means organizational efficiency in the AI era depends not just on the number of engineers, but on the ability to use smaller teams to validate directions faster.

Third, the boundary between platforms and applications is being redrawn. The success of Claude Code means Anthropic is no longer just a model provider; it has begun defining application forms itself. The controversy surrounding Claude Design and Figma shows that when model companies build their own applications, they inevitably touch the interests of customers and ecosystem partners. Previously, foundational model companies primarily provided underlying capabilities, with vertical applications like Cursor and Figma handling the user interface and scenario packaging. Now, model companies also need to demonstrate the agent-first future paradigm through their own products. This means AI platform competition is not just about APIs, but also about product paradigm competition.

Fourth, the stronger AI gets, the scarcer human judgment becomes. Mike repeatedly emphasizes that Claude can write code faster, generate prototypes, and execute tasks, but it cannot replace the most difficult part of the journey from 0 to 1: asking the right questions, understanding real users, defining the product's north star, and judging what is truly "right." In the past, execution capability was the main bottleneck in knowledge work. Now, execution is being accelerated by models, and human value is more concentrated in upfront judgment, creativity, relationship networks, and organizational ability. AI will not automatically eliminate difficult decisions; instead, it will amplify wrong directions faster.

If this conversation were compressed into one judgment, it would be: After Claude Code, Anthropic isn't looking for a single blockbuster product, but rather a methodology for transforming model capabilities into production systems. In this sense, the subject of this article is no longer just Anthropic's next product roadmap, but the structural shift of the entire AI industry from a "model competition" to a "system competition."

The following is the original content (edited for readability):

TL; DR

·AI product competition has shifted from "stronger models" to "how to land capabilities," essentially meaning large model companies are now vying for workflow entry points.

·The significance of Claude Code goes beyond writing code; it proves that agents can persistently execute tasks towards clear goals, pushing AI from a chat tool to a production system.

·Anthropic Labs' core value isn't in how many products it launches, but in using small teams to quickly validate what capabilities the model should acquire next.

·Co-work represents Anthropic's attempt to extend the Claude Code methodology to non-programmers, essentially abstracting "programming ability" into workflow automation capabilities for ordinary people.

·OpenAI's Codex catching up means Claude's advantage is no longer just technological leadership; it depends on whether Anthropic can integrate Claude Code, Co-work, and Claude.ai into a unified experience.

·Model companies building their own applications will intensify boundary conflicts with customers, but this is also the inevitable path for them to define the next generation of AI product paradigms.

·The faster AI can execute, the more human value is concentrated on upfront judgment, product taste, and problem definition, because wrong directions will also be amplified faster by AI.

·The impact of AI on employment is not a problem a single company can solve; it will fundamentally force society to re-discuss skill retraining, distribution mechanisms, and uniquely human irreplaceable capabilities.

Original Content

Alex Heath (Host): After Claude Code, what is Anthropic's next big product going to be? On this week's show, we have Mike Krieger. He's the co-founder of Instagram and now leads the team at Anthropic working on 'moonshot projects.'

Mike Krieger (Chief Product Officer, Anthropic):

One of my darkest days at Anthropic was naming it 3.5 Sonnet v2. I can explain why we ended up with that name.

Alex Heath: Mike and I recorded this conversation in person in San Francisco during Anthropic's recent Claude Code meetup. At that event, Anthropic announced a new major compute partnership with Elon Musk. So, are you guys going to space with Elon now?

Mike Krieger: Absolutely. Yes, we are looking for new and sometimes unexpected sources of compute.

Alex Heath: We talked about what Mike is working on now, the fierce competition between Anthropic and OpenAI, and what parts of human work Mike thinks will remain important even as AI gets more powerful.

This is Access.

Mike, great to see you here in San Francisco at the Claude Code event. I was just thinking back to our last conversation. You had just taken over Labs not long ago, but it's been a few months now, right?

Mike Krieger: Yes, almost four months.

How Labs Operates: Bi-weekly Triage, Validating Big Products with Small Teams

Alex Heath: Almost four months. I want to start here for people who don't know about Labs, because it's a pretty unique organizational structure. We talked about this when I visited your office a few months ago. What exactly is Labs? What is its mission within Anthropic?

Mike Krieger: Simply put, my understanding of Labs—the current version, which I'd call Labs v2, we can talk about what Labs v1 did and what v2 wants to do later—is that Labs does two main things.

First, it's about closing the gap between what Claude is theoretically capable of and what an average person experiences day-to-day. Claude can do a lot of things in theory, but how do those capabilities actually enter people's daily work and life? What products, prototypes, or projects do we need to build to demonstrate how to unlock more of this potential, to shrink that gap as much as possible?

Second, we act more like a 'frontier scout' team, trying to figure out where the model needs to evolve next to meet the needs of different users.

So, a successful Labs project isn't necessarily one that gets released as a product. It could be a prototype. We build it and find out: the model isn't good enough yet; it can't do this task right now. So we put it aside, re-evaluate when the next generation model comes out, or turn it into an evaluation metric for future model development, and keep iterating.

Therefore, unlike a product lab in a pure product company, where success might be measured by "did you ship a product," at Anthropic, Labs' value can also be seen in other ways: it can influence Anthropic's future direction.

Alex Heath: Labs has made some hits, right? Claude Code is one. MCP is another. What else?

Mike Krieger: Agent Skills is another important thing Labs built. I can also talk about a project we didn't release but was very helpful for research: computer use—getting Claude to use a computer.

I joined Anthropic in May 2024. Next week will be my two-year anniversary, what we call an 'antiversary' internally.

Alex Heath: You mean anniversary?

Mike Krieger: It's antiversary. Everything at Anthropic has to relate to ants. I was resistant at first. We don't say dogfood; we say antfood.

After I joined, we started building Labs. One of the earliest projects proposed was: why not try to get Claude to use a computer?

Alex Heath: That's computer use.

Mike Krieger: Exactly.

Alex Heath: What model era was that?

Mike Krieger: That was Claude Sonnet 3.5. It was also the first model I helped launch. I was in my third week and started working on this launch. We often joke that Anthropic doesn't have an 'onboarding program'; they just throw a really hard project at you. And for me, it was jumping into a launch in my third week.

Sonnet 3.5 was an interesting model because it was one of the first to truly unlock some coding scenarios. Not full agentic coding, but you could see the beginnings.

So, we put Sonnet 3.5 in and built a computer use product around it. But it had a lot of problems. It was too slow, not accurate enough, vision wasn't good enough. It would see the screen and say "I need to click that button," but then click somewhere else.

But building that 'not-fully-functional' testing framework was incredibly helpful. Because later, when we got to Sonnet 3.5 v2—the naming we can discuss later, truly one of my darkest days at Anthropic—we could put the new model into that framework and test it.

Then we tried 3.6. Still not good enough, but showing some signs of life. Then came 3.7. I remember that day vividly. I was on a business trip in New York, meeting the NY team. Suddenly, someone messaged me saying: We think that thing Labs built, that computer use project which had been sitting for nine months, is actually showing real signs of life on Sonnet 3.7. We think it's time to open this up as a capability and talk about it publicly.

That was about a nine-month journey. Every few months, we'd test the new model in the same framework. Even though Labs had shelved the project, it remained extremely useful as a test suite for tracking the evolution of computer use capabilities in our models.

Alex Heath: When you first joined Anthropic, it was as Chief Product Officer. I remember thinking at the time: Mike Krieger, the Instagram co-founder, in my mind, a very consumer-product founder, why would he join an enterprise AI company?

Mike Krieger: Right.

Alex Heath: We probably talked about this back then. I thought it was a really interesting choice. In hindsight, it was a good one. And the timing was excellent.

I'm curious how you went from being CPO, overseeing the whole product line—and 'AI product' is a vague, fast-changing concept—to moving to Labs about four or five months ago? From what I understand, you're more of an IC now, an individual contributor? Do you still manage people?

Mike Krieger: I don't manage anyone right now. We're just entering the performance review cycle.

Alex Heath: So this is what you wanted, right? To escape writing performance reviews?

Mike Krieger: Exactly. I open the system, see who I need to review, and found out: you just write your self-review and your manager's review.

Alex Heath: That's it?

Mike Krieger: That's it.

Alex Heath: Now Claude writes the performance reviews.

Mike Krieger: Claude does help write some reviews, which is quite useful. It doesn't write them entirely, but helps you remember: What did I actually do in the last six months?

I think companies go through different stages, and the alignment between those stages and what I'm truly passionate about changes.

When I joined, the entire product and engineering team was maybe 30 people, perhaps half and half. Sure, we had engineering teams working on research infrastructure, scalability, etc., but if you looked at the people purely building the product, it was mainly Claude.ai and what we then called the API—we didn't even call it the Claude Platform yet—maybe 30 to 35 people total. Very, very small.

It still felt like an early-stage startup. Many things were being defined. 'What is this product?' wasn't settled at all. Claude.ai back then had no Projects, no Artifacts. It was basically a list of your conversations with Claude, with almost no extra features.

So joining Anthropic then felt like joining a startup finding its product-market fit. Sure, it had a tailwind.

Alex Heath: When you joined, the Claude 3 family was already out, right? Opus, Sonnet, Haiku.

Mike Krieger: Yes. That was the first time Anthropic had a series of models that were at least near-frontier. There was just so much product work to do: what is this product going to become?

Even though my background is more consumer-focused, I was excited because, in the period between Instagram and Anthropic, Kevin and I did a lot of investing. We had a whole set of investment themes, one of which was 'the future of work.' How will work get done in the future?

And Anthropic seemed very likely to unlock that theme: what happens when you have a really smart assistant to help you work? I didn't even foresee at the time how disruptive this would become.

Alex Heath: You probably thought: it's an interesting little AI company; maybe it can help me understand some investment themes.

Mike Krieger: Yeah, maybe it would help us understand some themes we were thinking about. But actually, it changed a lot more than I imagined.

So that was Phase 1: small team, projects you could count on one hand. Then fast forward to the end of last year. The product team was hundreds of people. We had a portfolio of work. A lot of work became about deployment, understanding customer needs, being customer-facing, management layers, and all the things that come with scaling.

I realized that some people really love that kind of work and are very good at it. I respect them immensely. For me, I had a great coach who called this state the 'zone of competence'—things you're good at, you do well, you can handle, but it's not ultimately what truly ignites or drives you.

It's actually a dangerous position. You can stay there for a long time and seem to be performing well, but it's not where you have the most fire and drive.

So, in Q4 of last year, I started talking to Daniela about this. I said, the company has grown. We've run the standard five-year growth curve in a compressed timeframe. It was really only about two years.

Alex Heath: Yeah, I think you guys grew pretty well.

Mike Krieger: Yeah, growth was okay. Team size and product portfolio expanded rapidly. So I said, I think I want to go build a new company.

Daniela asked: Is this because you want to leave Anthropic, or because you want to change what you do here? I said, I really like this company. The people are great, I love the technology and the mission.

Around that time, we were also restarting Labs. Because Labs v1 was so successful, all its projects had 'graduated,' and no one was left. Labs was essentially put on the shelf.

So we decided to restart Labs, and I went back to being a builder. Everyone I meet inside and outside work says to me: 'Mike, you look so happy.'

Alex Heath: Some of your colleagues told me earlier today. They said, Mike is in such a great place, he's having so much fun.

Mike Krieger: Yes. Of course, I'm still my own harshest critic. So every day I think: How can I be better? What can we do? What can we build? What are we actually validating?

So it's not an easy job. But it aligns much more with what truly drives me.

Alex Heath: I don't