The AI-native startup: 5 products, 7-figure revenue, 100% AI-written code | Dan Shipper (co-founder/CEO of Every)

Jul 17, 2025 Episode Page ↗
Overview

Dan Shipper, co-founder and CEO of Avery, shares how his 15-person team operates at the bleeding edge of AI, building multiple products and a daily newsletter with zero manual coding. He provides actionable insights on leveraging AI for productivity, team structure, and future-proofing skills in the evolving 'allocation economy.'

At a Glance
17 Insights
1h 34m Duration
14 Topics
5 Concepts

Deep Dive Analysis

AI's Potential for Reshoring American Jobs

Underestimated Power of Command Line AI for Non-Coders

Defining AGI: The Length of the AI's Leash

AI's Impact on Learning and Skill Development

Every's AI-First Operating Model and AI Operations Lead

Dan Shipper's Personal and Team AI Tool Stack

The Concept of Compounding Engineering

Accelerating Career Growth with AI Tools

The Enduring Value of Coding Knowledge

Every's Product Incubation Strategy for AI Apps

Innovative Fundraising: The "Sip Seed" Round

Consulting and AI Adoption in Companies: Best Practices

The Allocation Economy: Future Skills and AI

Why Generalists Will Thrive in the AI Age

Compounding Engineering

This concept suggests that for every unit of work performed, the next unit should become easier. It involves investing time to create prompts and automations that streamline repetitive tasks, thereby increasing leverage and efficiency in engineering teams.

Allocation Economy

This describes a future economic model where management skills become broadly distributed and more valuable. As AI makes management cheaper, more people will need to learn how to manage AI agents and allocate tasks, shifting focus from specialized execution to oversight and direction.

Sip Seed Round

An innovative fundraising strategy where investors commit a specific amount of capital (e.g., $2 million) that the company can draw down whenever needed, typically on a SAFE at a set cap. This provides financial flexibility and optionality without the psychological pressure of a large, immediately available bank balance.

Context Engineering

This refers to the crucial process of providing the right context to an AI model at the precise moment it's needed. It is considered a significant factor in determining the performance and effectiveness of AI applications.

AI Operations Lead

A dedicated role within a company focused on continuously building prompts and workflows to automate tasks for all team members. This individual identifies repetitive processes and implements AI solutions to enhance overall organizational efficiency.

?
What are some common misconceptions about AI's impact on jobs and human skills?

Many headlines incorrectly suggest AI will eliminate entry-level jobs or turn off human brains. However, AI can accelerate learning and skill development, and while some skills may be given up, the gains in other areas can be much higher, similar to how writing impacted memory.

?
How can non-coders leverage powerful AI tools like Claude Code?

Non-coders can use command-line interfaces like Claude Code to process large amounts of text data from local files, such as meeting notes or books, to perform complex analysis or generate insights without manual intervention.

?
What is the most significant predictor of a company's successful AI adoption?

The number one predictor is whether the CEO actively uses AI tools like ChatGPT. A CEO who leads by example, expresses excitement, and sets realistic expectations drives company-wide adoption and momentum.

?
How does Every operate as an "AI-first" company with a small team?

Every employs an "AI operations lead" to automate repetitive tasks, uses "compounding engineering" to make future work easier, and leverages multiple AI agents (like Claude Code, Friday, Charlie) for development, allowing a small team of generalists to build multiple products with minimal manual coding.

?
What is "compounding engineering" and how does it increase leverage?

Compounding engineering means making each successive unit of work easier by investing time upfront to create prompts or automations. For example, building a prompt to turn rambling thoughts into a structured PRD reduces future PRD writing effort.

?
Will knowing how to code become obsolete in the AI era?

While AI agents can write code, knowing how to code remains valuable for understanding underlying systems and debugging. It allows engineers to go "down a layer in the stack" when needed, a skill that will diminish in importance over decades, not years.

?
What kind of products are best suited for development in the current AI landscape?

The best products to build are those that historically were very expensive services (e.g., chief of staff, ghostwriter, lawyer) that AI can now make orders of magnitude cheaper and accessible to everyone.

?
How does Every approach fundraising to maintain its creative spirit?

Every uses a "sip seed" round, where investors commit capital that can be drawn down as needed, rather than receiving a large lump sum. This provides a cash cushion and optionality without the psychological pressure of a massive bank account.

?
What skills will be most valuable in the "allocation economy" driven by AI?

In the allocation economy, skills traditionally associated with management will become crucial, such as evaluating AI agent "talent," setting a clear vision, developing taste, and knowing when to dive into details versus delegating tasks to AI.

?
Why will generalists thrive in an AI-first world?

AI acts like having "10,000 PhDs in your pocket," handling specialized tasks and providing vast knowledge. This empowers generalists to dabble across different domains, build diverse skills, and achieve more without deep, decade-long specialization in every area.

1. CEO Must Actively Use AI

For successful AI adoption within an organization, the CEO must be an active, daily user of AI tools like ChatGPT. Their personal engagement drives excitement, sets realistic expectations, and fosters widespread adoption across the company.

2. Hire AI Operations Lead

Appoint a dedicated AI Operations Lead whose sole job is to constantly build prompts and workflows to automate repetitive tasks for the entire team. This role helps overcome the inertia of individuals having to build new AI processes themselves, significantly increasing efficiency.

3. Codify Feedback into AI Prompts

Codify common feedback and style guides (e.g., for writing, headlines) into AI prompts. This allows team members to get ‘simulated’ feedback from AI before human review, pushing quality standards to the edge and significantly reducing repetitive managerial tasks.

4. Practice Compounding Engineering

Adopt a principle where for every unit of work, you invest a little time to make the next unit easier to do. This involves creating and refining prompts and automations (e.g., for PRDs) that generate leverage and accelerate future tasks, even with a small team.

5. Democratize Expensive Services with AI

Identify historically expensive services (e.g., chief of staff, ghostwriter, lawyer, personal organizer) that AI can make orders of magnitude cheaper. Build products or internal tools around these to meet previously unfulfilled demand, creating new market opportunities.

6. Create AI Learning Forums

Establish regular forums, such as weekly meetings or internal emails with usage statistics, where employees can share AI prompts, use cases, and learnings. This highlights early adopters, fosters momentum, and facilitates knowledge transfer across the organization.

7. Utilize Cloud Code for Non-Coders

Encourage non-programmers to use Cloud Code (command-line interface agents) to process large amounts of text, like meeting notes or books, for autonomous task completion and deeper analysis. This tool can work for long periods without intervention, making it incredibly powerful for text-heavy tasks.

8. Leverage Multiple AI Agents

Utilize different AI agents (e.g., Claude, Friday, Charlie) for tasks, recognizing that they have distinct ‘personalities’ and styles. This approach provides diverse perspectives and capabilities, similar to assembling a team of human specialists.

9. Develop AI Management Skills

Focus on developing management skills such as problem communication, information gathering, task division, feedback, and vision-setting. These skills become crucial for effectively ‘managing’ AI agents and leveraging their capabilities in the evolving ‘allocation economy’.

10. Cultivate Generalist Skills

Embrace and cultivate generalist skills, allowing individuals to dabble in diverse domains like coding, video creation, image generation, and writing. AI acts as ‘10,000 PhDs in your pocket,’ handling specialized tasks and empowering generalists to achieve more across various fields.

11. Automate Repetitive Communication

Set a goal to avoid repeating yourself in meetings by codifying common feedback and information into AI prompts and workflows. This pushes your ’taste’ and knowledge to the edge of the organization, freeing up time for higher-value activities.

12. Utilize AI as a Quality Judge

Employ advanced AI models like Claude Opus 4 that can ‘judge’ the quality of creative output (e.g., writing). This allows AI agents to self-improve before presenting work to humans, significantly streamlining creative processes and raising quality bars.

13. Understand Code (Even with AI)

While AI can write code, it remains valuable for individuals to understand how to code and go ‘down a layer in the stack.’ This foundational knowledge accelerates problem-solving and understanding, especially during technological transitions, even if you’re not manually writing lines of code.

14. Utilize Speech-to-Text Interfaces

Integrate speech-to-text tools (e.g., Monologue, Super Whisper, Whisper Flow, Notion’s meeting recording) into daily workflows. These interfaces represent the future of human-computer interaction and can significantly improve efficiency for tasks like note-taking and content creation.

15. AI for Self-Reflection

Leverage AI tools with memory (like ChatGPT-3) for self-reflection and personal growth. Feed them meeting transcripts or personal data to gain insights into your behaviors, identify patterns, and track progress on personal development goals.

16. Adopt ‘Sip Seed’ Funding

Consider a ‘sip seed’ funding model where committed capital can be drawn down whenever needed, rather than receiving a large lump sum. This provides psychological safety for risk-taking without the pressure of a large bank account balance that might encourage excessive burn.

17. Align Business with Personal Joy

Structure your business around what genuinely brings you joy and aligns with your core identity (e.g., writing), even if it deviates from typical industry models. This approach leads to greater personal fulfillment and often better, more unique business outcomes.

I think that AI may be one of the biggest force for reshoring American jobs.

Dan Shipper

No one is manually coding anymore.

Dan Shipper

A good definition of AGI is when does it become economically profitable for people to run agents indefinitely?

Dan Shipper

Whenever I see a kid with ChatGPT, I'm like, holy shit, they're going to go so much faster than any other person that I've worked with.

Dan Shipper

Using AI is a skill.

Dan Shipper

The entire game board has been like totally reset in terms of things you can build.

Dan Shipper

GPT wrappers are amazing and they've been much maligned for absolutely no reason.

Dan Shipper

If the CEO is in it all the time being like, this is the coolest thing, everybody else is going to start doing it.

Dan Shipper

Do things worth writing about and write things worth reading.

Dan Shipper (attributing Pliny the Younger)

AI-First Company Adoption Playbook

Dan Shipper (describing practices of companies like Walleye Capital)
  1. CEO sends an "AI-first company" memo, leading by example (e.g., stating the memo was written with ChatGPT).
  2. Host weekly meetings where employees share AI prompts and use cases.
  3. Send a weekly company-wide email sharing AI usage stats and highlighting employees who contributed new prompts or use cases.

Internal AI Operations Workflow

Dan Shipper
  1. Designate an "AI operations lead" whose job is to automate tasks.
  2. Identify repetitive tasks performed by team members (e.g., CEO, Editor-in-Chief).
  3. The AI operations lead builds prompts and workflows (e.g., using Claude Opus 4 and a style guide for copy editing).
  4. Integrate these prompts into engineers' workflows (e.g., a Claude Code command that checks codebase for copy edits and creates a pull request).
  5. Encourage behavioral change to ensure adoption (e.g., asking if a prompt has been used).

Compounding Engineering Workflow

Dan Shipper (describing Every's Quora team practices)
  1. For every unit of work (e.g., writing a PRD), identify opportunities to make the next unit easier.
  2. Spend time creating prompts or automations (e.g., a prompt to turn rambling thoughts into a structured PRD).
  3. Store and share these prompts/automations (e.g., in a GitHub repository with slash commands for Claude Code).
  4. Utilize multiple AI agents (e.g., Claude, Friday, Charlie) with different "personalities" or strengths for various tasks.

AI Product Incubation Model (Every's approach)

Dan Shipper
  1. Identify historically expensive services (e.g., chief of staff, ghostwriter) that AI can make orders of magnitude cheaper.
  2. Experiment with general-purpose AI tools (e.g., ChatGPT, Claude) to see if they can fulfill these needs.
  3. If a general-purpose tool proves useful, unbundle it into its own dedicated app.
  4. Use internal team members (who are AI-first generalists) as the first users and measure success by internal adoption ("is it a banger inside of Every?").
  5. Leverage the company's audience (who share a similar "vibe") as the next set of users.
15 people
Every's team size Builds and ships four products, publishes a daily newsletter, and operates a consulting arm.
~100,000
Every's newsletter subscribers Core of the business, read by people at top AI labs.
$700,000
Every's pre-seed fundraising round Raised during the height of the creator economy, with a modified SAFE to allow optionality.
$2 million
Every's "sip seed" fundraising commitment Committed by Reid Hoffman and Starting Line VC, drawn down as needed on a SAFE at a set cap.
~$300,000
Estimated cost to build Quora (Every's AI chief of staff for email) Built by two engineers (plus AI agents) and has 2500 active users.
2,500
Quora's active users Every's AI chief of staff for email, recently launched publicly.
~$1 million
Every's consulting arm revenue (last year) Expected to double this year, helps big companies adopt AI best practices.
8%
Percentage of the workforce that are managers (pre-AI) Management skills are not broadly distributed but will become cheaper and more common with AI.