Dario Amodei named "the zeroth world." You probably live in it.

The 10 things from the interview that almost nobody else is covering

Ruben Dominguez

Jun 01, 2026

I sat down to watch one Davos interview last night.

I ended up with 30 minutes of notes and one question I cannot stop thinking about.

The interviewer asked Dario Amodei if the world is ready for what is coming.

He said no. One word. Then he spent the rest of the interview explaining why.

If you only get 5 minutes today, this is the 30 minutes worth reading instead. Here are the 10 things that matter.

For where this fits in the AGI race across labs, see Anthropic is closing in on a $1 trillion valuation, Anthropic just passed OpenAI in revenue, spending 4x less, and Demis Hassabis named his AGI year.

📢 A quick word before we get into it:

Amodei’s whole point is that the engineers at the frontier company already operate at a level the rest of us are 12-24 months behind on. The fastest way to close that gap is not reading about Claude. It is learning to use it the way they do.

The world’s first LIVE Claude-a-thon runs this weekend. 16 hours, 2 days, free:

Master Claude’s 3 modes (Chat, Cowork, Code), set up Skills and Connectors to automate your desktop and files, and vibe code apps without writing code. Plus 10+ AI tools that pair with Claude.

🧠 Saturday and Sunday, 10am-7pm EST.

That is the weekend that puts you on the right side of the zeroth-world line.

1. Amodei answered the most important question in business with one word. The word was no.

The interviewer asked if anyone is actually preparing for this.

“I don’t think there’s an awareness at all of what is coming here and the magnitude of it.”

He has watched this field for 15 years. The capability curve has been smooth and relentless the whole time. The public narrative swings between panic and euphoria every 3-6 months. The model improves at the same rate either way.

Amodei calls it a Moore’s law for intelligence. The hype cycles are noise. The exponential is the signal.

Most policy is being built around the noise.

Every regulatory framework and corporate strategy indexed to sentiment cycles is already wrong. Ask your leadership team this week: is our AI strategy built around the hype cycle or the capability curve? Those two inputs produce opposite decisions.

For the strategic version of this argument, see Marc Andreessen on why the AI moat is not the model, Mark Cuban on the AI thesis, and Sam Altman watched 900 million people talk to one personality every week. For the founder mindset that maps onto this curve instead of the cycle, see the AI agent that thinks like Jensen Huang, Elon Musk, and Dario Amodei.

2. 5 to 10% GDP growth. 10% unemployment. A combination that has never existed.

GDP growth has always meant more jobs. More output, more things to do.

AI breaks that, and not gently.

“My view is the signature of this technology is it’s gonna take us to a world where we have very high GDP growth and potentially also very high unemployment and inequality. That’s not a combination we’ve almost ever seen before.”

He puts the numbers on the table: 5 to 10% GDP growth alongside 10% unemployment. Historically unprecedented. The math is internally consistent. The policy frameworks are not built for it.

Past technologies displaced some jobs and created others nearby. AI displaces entire categories of knowledge work simultaneously, faster than retraining absorbs.

Three distinctions that separate this from every previous disruption:

1️⃣ Structural career elimination. Permanent, not cyclical.

2️⃣ The wealth created will be real. The distribution will be the problem.

3️⃣ The timeline is shorter than any policy currently plans for.

Every workforce strategy, education policy, and safety net built on the assumption that growth equals jobs is built on the wrong model. Most are. Start the audit now.

For the most honest public accounting of which jobs are getting hit first, see Anthropic just showed us which jobs AI is actually replacing and no one is safe from AI. For the small business side of the same shift, where the displacement hits hardest and earliest, see the complete Claude setup guide for small business owners.

3. Anthropic Engineers Have Stopped Writing Code. The Rest of the Economy Is 12 to 24 Months Behind.

This one stopped me.

Engineering leads inside Anthropic are telling Dario, directly, that they no longer write code. Claude Opus does the work. They edit.

“I have some engineers, some engineering leads within Anthropic who have basically said to me, ‘I don’t write any code anymore. I just let Opus do the work and I edit it.’”

If the engineers at the frontier AI company are already there, the 26 million software developers globally are on a delayed version of the same curve.

Software becomes cheap. Possibly free. The premise that software must be amortized across millions of users collapses. A custom app for a single meeting costs a few cents.

“Even if the software engineers are only doing 10% of it, they still have a job to do. That’s not gonna last forever.”
Name one workflow in your company that would look different if code cost nothing. Start there this week.

For the production-grade systems already operating at this level, see the Claude Code system that replaces a 5-person team, the complete guide to AI coding in 2026, the AI code review checklist that prevents the next $1M production incident, Claude Cowork: the tool that triggered a $285B software selloff, and 25 Claude Skills that give your startup a marketing team it cannot afford yet.

4. Anthropic built the Economic Index. Almost nobody in policy is reading it.

No government measures this in real time. Anthropic built it anyway.

“Until we can measure the shape of this economic transition, any policy is gonna be blind and misinformed.”

The Anthropic Economic Index tracks in real time what Claude is being used for across all conversations, in a privacy-preserving way. It measures which tasks are being automated versus augmented, which industries are adopting, and how adoption is diffusing across U.S. states and countries. Updated 4-5 times in the past year.

Almost no one in policy is using it.

The Economic Index is a leading indicator for sector disruption. If you can see where AI is automating versus augmenting before the labor market data shows up, you are seeing the future 12-18 months early. It is public. Read it before your next strategy meeting.

For the VC lens on the same data, see where VC money is going in AI, Coatue’s 18-chart AI report, and the most valuable VC-backed startups in the world. For the analyst-grade workflow built on the same kind of real-time data discipline, see build your own stock analyst with Claude: the 12-prompt system.

5. Autocracy is the actual geopolitical AI risk. The fix is one policy lever.

Most people think the geopolitical AI risk is one country beating another.

Amodei thinks the risk is what happens to human freedom.

“AI may be uniquely well suited to autocracy and to deepening repression. Individualized propaganda. Breaking into any computer system in the world. Surveilling everyone in a population, detecting dissent, suppressing it. A huge army of drones that could go after each individual person. It’s really scary.”

His solution is not a geopolitical alliance or an arms treaty. It is a supply chain constraint.

“We don’t need to fight them. We just need to not sell these chips.”

One lever. The political will to hold it is not there yet.

For investors in semiconductor and infrastructure companies, this is a material geopolitical risk with a specific mechanism. The lever exists. Whether it gets pulled is the variable to track.

For the broader geopolitical and infrastructure thesis, see Elon Musk and the outer limit of vertical integration and the most valuable VC-backed startups in the world.

6. 10 million people are forming their own economy. He calls it the zeroth world.

This is the line that will be quoted for years.

“The nightmare would be that there’s this emerging zeroth world country of 10 million people. Seven million in Silicon Valley, three million scattered throughout. Forming its own economy. Becoming decoupled.”

What looks like 10% national GDP growth looks like 50% growth inside that concentrated group. The technology compounds in their favor so fast that separation becomes permanent.

His framework for preventing it:

1️⃣ Measure the transition in real time before policy is written

2️⃣ Invest in adaptation infrastructure, not one-off retraining programs

3️⃣ Accept that government redistribution at macro scale is unavoidable arithmetic

If you are reading this, you probably live inside the zeroth world. Whether the fracture happens depends in part on decisions made by the people closest to this technology.

For the personal leverage version of this divide, the moves that put you on the right side of the zeroth-world line, see your voice is the only AI moat that compounds, I built a second brain in 10 minutes with Granola + Claude, the single best productivity decision you can make with Claude right now, and why ChatGPT and Claude keep disappointing you.

For the SMB owner side of the same leverage, where the compounding hits hardest if you start now, see the complete Claude setup guide for small business owners.

7. Scientists and social media founders are not the same. The difference is what gets optimized.

2 types of people lead the biggest AI companies. Amodei draws a line between them.

“There’s a long tradition of scientists thinking about the effects of the technology they build, of thinking of themselves as having responsibility for the technology they build, not ducking responsibility.”

The social media generation was shaped by different incentives. Engagement maximization. Manipulation as a design pattern. Growth at any cost as an operating system. Those incentives shaped what gets built and what success means.

Amodei does not name names. The contrast does the work.

Sell to businesses that pay for value and you do not have to fight your own incentives to avoid addictive product design. The revenue model is the safety model.

That is the Anthropic enterprise thesis in one sentence.

Who leads an AI company determines what gets optimized. Check the incentive structure of every AI vendor in your stack.

For the founder-incentive lens, see the AI agent that thinks like Jensen Huang, Elon Musk, and Dario Amodei, Dario Amodei and the long game of safe AI, and Anthropic’s 2022 pitch deck just leaked.

8. The one technical breakthrough that determines whether AI is actually safe.

Current alignment cannot tell you why a model does what it does.

Mechanistic interpretability can.

“Similar to how you can learn things about human brains by doing an MRI or an X-ray that you can’t learn just by talking to a human, the science of looking inside the AI models, I am convinced that this ultimately holds the key to making the model safe and controllable because it’s the only ground truth we have.”

Current alignment methods are phenomenological. You test outputs. You observe behavior. You train on behavior.

A model can tell you it is doing X for reason Y while doing X for a completely different reason. Or while lying about doing X entirely.

Amodei says Anthropic’s own tests show deception, blackmail, and sycophancy in current models, including their own. They publish those findings.

The only way to ground-truth model behavior is to understand internal representations, not just outputs.

For anyone building on top of AI models in enterprise, interpretability is a risk management question today. Ask every AI vendor you work with: what is your interpretability roadmap? The ones with no answer are the ones to watch most carefully.

For the engineering discipline this demands, see the AI code review checklist that prevents the next $1M production incident, stop blaming the model, fix the architecture, and prompt engineering is dead, context engineering is what matters now.

9. $0 to $100M. $100M to $1B. $1B to $10B. Three years.

The revenue curve matches the technology curve. The public narrative has matched neither.

“We have this revenue curve that in 2023 went from zero to roughly $100 million, in 2024 from roughly $100 million to roughly $1 billion, 2025 from roughly $1 billion to roughly $10 billion.”

Through all of it, the commentary alternated between “Anthropic is changing the world” and “Anthropic is finished” every few months.

The curve did not notice.

The most recent proof: Claude Code. Agentic task completion crossed an inflection point with Opus 4.5. Non-technical users were wrestling with command-line interfaces just to get access. Anthropic built a better UI in 2 weeks. Day-one metrics came in at 4x anything previously released.

When the infrastructure company grows at this rate, the applications layer is 3-5 years behind on the same trajectory. That gap is where the opportunity currently sits.

For where that opportunity sits right now, see 50 game-changing AI agent startup ideas for 2026, 70 startup ideas YC wants you to build, the AI GTM playbook for 2026, and the SaaS defense playbook for the AI era.

For the fundraising side of building into this curve, see the U.S. VC database most founders never find and the angel investors SaaS database for 2026. For the investor-communication discipline that closes rounds in this market, see the McKinsey Pyramid Principle applied to Claude for investor documents.

10. Redistribution Is Not Coming Because of Ideology. It Is Coming Because the Math Leaves No Other Option.

The wealth will be concentrated. The displacement will be broad. The mismatch forces the response.

His prediction: within 1-2 years, positions that currently seem politically coded become bipartisan necessities. Not because anyone changed their values. Because the technology left no other option.

“The pie is gonna grow much larger. The money is gonna be there. The issue is distributing it to the right people. This is probably a time to worry less about disincentivizing growth and worry more about making sure that everyone gets a part of that growth.”
“Ideology will not survive the nature of this technology. It won’t survive reality.”
For long-horizon investors and operators, this is the macro framework. The question is not whether redistribution policy happens. It is what form it takes, how fast, and whether you positioned before or after the consensus shifted.

For the VC playbook that maps onto this macro, see what top-tier VCs actually look for in 2026, $80 billion in 3 months: Q1 2026’s record-breaking fundraising, the ultimate investor list of lists, the U.S. VC database most founders never find, and the angel investors SaaS database for 2026.

What this means for you

Amodei’s thesis has one spine across the entire interview.

The technology is on a smooth exponential. The economic effects will be extreme in both directions. The awareness gap is the primary risk of this decade.

Founders

The software cost curve is collapsing. The moats that exist in a world of cheap code are not the moats that exist today. Proprietary data, physical-world operations, customer relationships, and trust are the candidates. Start building toward those now.

▫️ The SaaS defense playbook for the AI era

▫️ 50 game-changing AI agent startup ideas for 2026

▫️ 70 startup ideas YC wants you to build

▫️ The AI GTM playbook for 2026

▫️ The Claude Code system that replaces a 5-person team

▫️ The McKinsey Pyramid Principle applied to Claude for investor documents

Investors

The Anthropic revenue curve is a leading indicator for the applications layer. The Anthropic Economic Index is the most granular real-time signal available for where AI is creating value and displacing labor. It is public. Most people reading this are not using it.

▫️ What top VCs check in due diligence before writing checks

▫️ Where VC money is going in AI

▫️ Coatue’s 18-chart AI report

▫️ The U.S. VC database most founders never find

▫️ The angel investors SaaS database for 2026

▫️ The full investor lists archive

Operators

The engineers inside the frontier AI company are already not writing code. Any workforce strategy that assumes knowledge work roles are stable for 5 years is built on the wrong assumption. Adaptation infrastructure matters more than any retraining program.

▫️ The single best productivity decision you can make with Claude right now

▫️ Build your own stock analyst with Claude: the 12-prompt system

▫️ I built a second brain in 10 minutes with Granola + Claude

▫️ Your voice is the only AI moat that compounds

▫️ The complete Claude setup guide for small business owners

▫️ The McKinsey Pyramid Principle for investor documents

Everyone in tech

The zeroth world is forming now. Whether the people who benefit most from this technology take the distribution question seriously before it becomes a political crisis is the defining question of the next decade.

The 3 things to hold

1️⃣ The curve is the signal. The commentary is noise. Build strategy around the exponential.

2️⃣ Measure before you policy. The Economic Index is the model for decision-making under this kind of uncertainty.

3️⃣ The distribution problem is bigger than the creation problem. AI will create wealth. That part is settled. Whether it reaches beyond the zeroth world requires deliberate action, not optimism.

The technology is not waiting for awareness to catch up. The curve does not care about the debate. The only question is whether the institutions act before the math forces them to.

Full interview:

If this saved you 30 minutes, share it with one founder or investor who needs to see it.

The AI Corner

Discussion about this post

Ready for more?