{"id":2228,"date":"2026-04-29T20:54:40","date_gmt":"2026-04-30T03:54:40","guid":{"rendered":"https:\/\/dmkthinks.org\/blog\/?p=2228"},"modified":"2026-04-29T22:40:38","modified_gmt":"2026-04-30T05:40:38","slug":"observations-on-anthropic","status":"publish","type":"post","link":"https:\/\/dmkthinks.org\/blog\/observations-on-anthropic\/","title":{"rendered":"Observations on Anthropic"},"content":{"rendered":"\n

Take a look at this chart:<\/p>\n\n\n\n

\"Anthropic<\/figure>\n\n\n\n

And this LinkedIn post:<\/p>\n\n\n\n

\"A<\/figure>\n\n\n\n

These numbers are nuts.<\/p>\n\n\n\n

Last year, the conversations online were about how amazing Lovable was for being the fastest-growing company in history from zero to $100M ARR (8 months).<\/p>\n\n\n\n

But Anthropic increased their run rate from $9 billion to $30 billion in the span of 4 MONTHS.<\/strong> These numbers are on another level. Especially when the CEO and one of the founders of the company, Dario Amodei, has said the founding team didn’t necessarily want to create their own company.<\/p>\n\n\n\n

Starting as a spinoff from OpenAI in 2021<\/a>, Anthropic started as a research lab<\/strong> focusing on safety and niche products for a niche demographic.<\/p>\n\n\n\n

When OpenAI launched their first public version of ChatGPT on November 20th, 2022, it took Anthropic almost 4 months to catch up and launch their first version of Claude (in March 2023).<\/p>\n\n\n\n

Anthropic was late to the show, but they’ve made up incredible ground against OpenAI in both usage and revenue.<\/p>\n\n\n\n

Today, you can’t turn anywhere without someone on LinkedIn sharing Claude Code advice. Even Mark Zuckerberg is using the product: Sources: Mark Zuckerberg is back to writing code after a two-decade hiatus, submitting three diffs to Meta’s monorepo, and is a heavy user of Claude Code CLI<\/em><\/a>.<\/p>\n\n\n\n

And big AI conferences, like HumanX, are buzzing about Anthropic’s products. From a recent CNBC article<\/a>:<\/p>\n\n\n\n

\n

\u201cIt has become a religion, that\u2019s the level of that mania,\u201d Jain said in an interview. \u201cEverybody, if you go and ask them today, \u2018Hey, if I gave you one AI tool, what tool would you want?\u2019 The answer would be Claude.\u201d<\/p>\n<\/blockquote>\n\n\n\n

And from a blog post by legendary tech blogger Steve Yegge<\/a>:<\/p>\n\n\n\n

\n

As you\u2019ve probably noticed, something is happening over at Anthropic. They are a spaceship that is beginning to take off.<\/p>\n<\/blockquote>\n\n\n\n

With Claude Code the top choice of technical users, Claude.ai exploding in popularity, and the launch and immediate hockey stick growth of Cowork…<\/strong><\/p>\n\n\n\n

Anthropic is no longer the hipster, underground alternative. They’re leading frontier AI.<\/strong><\/p>\n\n\n\n

From a recent article in TIME<\/a>:<\/p>\n\n\n\n

\n

Already, its $380 billion valuation eclipses those of Goldman Sachs, McDonalds, and Coca-Cola. Its revenues are a rocket ship. Claude is considered a world-class model, with products like Code and Cowork upending what it means to be a programmer. Its tools are so good that each new release causes stock-market shocks, as investors grasp the likelihood the advances will upend entire categories, from law to software development. Over the past few months, it emerged as the company most poised to disrupt the future of work.<\/p>\n<\/blockquote>\n\n\n\n

And from the article Everyone Is Talking About AI. The People Actually Winning Are Using Claude<\/em><\/a>:<\/p>\n\n\n\n

\n

From $1B in annualized revenue at the end of 2024 to $14B by February 2026. That\u2019s a 14x jump in 14 months.<\/p>\n<\/blockquote>\n\n\n\n

These aren’t normal numbers, and this isn’t a normal company.<\/p>\n\n\n\n

Since I started using Claude Code heavily in July 2025, I’ve followed Anthropic’s journey closely.<\/strong><\/p>\n\n\n\n

I’ve spend thousands of hours using their products both professionally and personally, reading their white papers, optimizing my setup, browsing niche subreddits, reading tell-all articles, and listening to interviews with team members.<\/p>\n\n\n\n

And today, I’m sharing what’s sticking out to me \u2014 about where they’re going, how they think about business, and my observations on why they’ve grown.<\/strong><\/p>\n\n\n\n\n\n\n\n

One of my favorite ways to explain Anthropic comes from this chart:<\/p>\n\n\n\n

\"\"<\/figure>\n\n\n\n

And this quote:<\/p>\n\n\n\n

\n

Growth skyrocketed. Annualized revenue from the coding agent alone topped $1 billion by the end of 2025. By February, it had more than doubled to $2.5 billion.<\/p>\n<\/blockquote>\n\n\n\n

Most companies would kill to be this big.<\/p>\n\n\n\n

But, it’s not a company.<\/p>\n\n\n\n

It’s a single product from Anthropic called Claude Code.<\/p>\n\n\n\n

To understand Anthropic, you need to first understand Claude Code<\/strong>. This product was the tipping point of public attention.<\/p>\n\n\n\n

Increasingly, the Claude Code team has been sharing their learnings publicly. This starts with Cat Wu, the head of product at Claude Code.<\/strong><\/p>\n\n\n\n

In an interview on AI product leader Peter Yang’s podcast<\/a>, Cat shares behind-the-scenes on how the Claude Code team thinks about building:<\/p>\n\n\n\n

\n

“Our design philosophy is to make sure the CLI is very, very simple to onboard to. We have this philosophy that for new features there should be no onboarding UX. It should be intuitive via the feature name, via a one-line description of what it does. And you should be able to just get started.”<\/p>\n<\/blockquote>\n\n\n\n

As companies get bigger, it’s easier to add more complexity and features. A $30B company still focused on simplicity is a strong sign of customer alignment (and, in my experience, continued success).<\/p>\n\n\n\n

Deeper in the interview, Peter follows up with a question to learn more about how the Claude Code team functions:<\/p>\n\n\n\n

\n

Peter: “There are new features coming out all the time, and I’m just wondering how the Claude Code team actually work in a given week.”<\/p>\n\n\n\n

Cat: “Our team is very strong. A lot of strong product engineers who love to have end-to-end ownership over features. The way things normally work is there’s some higher level design principles, but within that there’s a lot that you can build and a lot of our best features was like an engineer prototyping an idea, shipping it to dogfooding [internally], and we just hear what the feedback is. For the features that people love, that’s a strong signal that we should fast track it to sharing it to the public. A lot the features that you’ve seen that we have shipped have gone through two to three iterations internally before we made it public, and there’s a lot of features that we’ve played around internally that we’ve decided not to ship.”<\/p>\n<\/blockquote>\n\n\n\n

As a Product leader, I also really liked a comment Cat made about the role of a PM in a deeply technical product that has strong developer ownership:<\/p>\n\n\n\n

\n

“I see the role of a PM as, one setting the broader direction of what is the minimum or aspirational bar that we want our products features to fall between and shepherding it through the process. And also, in the age of AI tools a lot of the PM role is around pricing and packaging as well, so developers can focus on the best coding experience possible and then the PM’s role is to make sure it’s accessible for the world.”<\/p>\n<\/blockquote>\n\n\n\n

Cat has a lot of interesting product philosophies, and she also wrote a blog post on Anthropic’s official blog about the changing role of product management in AI<\/a>.<\/p>\n\n\n\n

Her opinions on product leadership during the time of AI is similar to what fast-moving companies, like ours with AppSumo Originals<\/a>, have been doing for years. Now with AI increasing velocity even more.<\/p>\n\n\n\n

Going back to to the interview with Peter, Cat gives more detail on what shorter timeframe planning means:<\/p>\n\n\n\n

\n

Peter: “What do you think Claude will look a year or two from now?”<\/p>\n\n\n\n

Cat: “A year or two is a really long time, I can talk about the next few months.”<\/p>\n<\/blockquote>\n\n\n\n

With models rapidly improving their capabilities, the team at Anthropic is way more nimble than I’ve ever seen for a ~3,000 person company. Usually there’s a lot more structure, layering, planning, and organizational red tape. But it seems like Anthropic is highly aware that that’s a waste of time when models step change in their capabilities every 4-8 months.<\/p>\n\n\n\n

Cat’s colleague, and another person on the Claude Code team is Boris Cherny \u2014 who created the initial Claude Code.<\/strong><\/p>\n\n\n\n

In her post on Anthropic’s blog, Cat talked about encouraging team members to go on “side quests”. This really fits Boris’ behavior. He came up with the prototype for Claude Code in his first month at Anthropic<\/strong>.<\/p>\n\n\n\n

From a recent TIME article<\/a>:<\/p>\n\n\n\n

\n

Boris Cherny, the creator of Claude Code, had a simple question for his new tool: \u201cWhat music am I listening to?\u201d It was September 2024, the Ukrainian-born engineer\u2019s first month at Anthropic. Cherny shared his prototype internally. Claude Code spread so quickly that in Cherny\u2019s first performance review, Dario Amodei asked if he was forcing colleagues to use it. When a research preview of the tool was publicly released in February 2025, programmers outside Anthropic flocked to it too. Then in November, Anthropic released a new version of Claude that, when strapped into Claude Code, was good enough at spotting its own mistakes to be trusted to complete tasks on its own. Cherny stopped writing his own code entirely.<\/p>\n<\/blockquote>\n\n\n\n

Interestingly, Claude Code didn’t even release to the public till five months after the prototype launched. And then it was another 10 months until it really got popular. This wasn’t an overnight success.<\/strong><\/p>\n\n\n\n

Back to Boris, his journey to Anthropic \u2014 and back again \u2014 is a fascinating path. He joined Anthropic from Meta, where he was one of the team’s top performers, and he left his first stint at Anthropic for a bit to join Cursor, before returning to Anthropic.<\/p>\n\n\n\n

In an interview with Wired<\/a>, Boris goes into his views on simplicity:<\/p>\n\n\n\n

\n

\u201cWe built the simplest possible thing,\u201d said Cherny. \u201cThe craziest thing was learning three months ago that half of the sales team at Anthropic uses Claude Code every week.\u201d<\/p>\n<\/blockquote>\n\n\n\n

A core belief I’ve had in the businesses I’ve helped over the years is to get the simplest idea to market based on a loose idea of what a customer wants, and then iterate. Progress over perfection.<\/p>\n\n\n\n

Launching fast is a tradeoff, though. A lot of founders and executives want to launch fast… but then get frustrated when they see bugs or issues. Launching fast means that mistakes will be made, and we have to live with those mistakes.<\/p>\n\n\n\n

In light of those tradeoffs, I like another part of the Wired article where Boris shows strong understanding of the risks in launching early, while also having foresight to the larger picture of iteration and potential:<\/p>\n\n\n\n

\n

Cherny acknowledges that early versions of Claude Code often stumbled, making errors or getting stuck in costly loops. Cherny says Anthropic built Claude Code for where AI capabilities were headed, rather than where they were at launch.<\/p>\n<\/blockquote>\n\n\n\n

In another interview, Boris went on the Lenny’s Newsletter<\/a> podcast where he shared more about what happens at Anthropic, his beliefs on AI, and how Claude code was built.<\/p>\n\n\n\n