Building an
AI-Native Company
"There's definitely a huge 10x difference between an org where 90% of the engineers are using AI versus an org where 100% of the engineers are using AI. It's totally different."
Dan Shipper
CEO, Every • 00:02:15
Code Written by AI
People, 4 Products
Engineer per App
Faster with 100% AI
The 100% Threshold
Why partial AI adoption holds you back
"If even 10% of your company is using traditional methods, you have to lean all the way back into that world."
The last 10% of non-AI adoption creates massive drag. Even a small group using traditional workflows forces the entire organization to accommodate old patterns, preventing the leap to entirely new ways of working.
Watch explanation (00:02:29)What 100% AI Adoption Looks Like
"99% of our code is written by AI agents. No one is handwriting code. No one is writing code at all."
— Dan Shipper (00:03:48)
"Each one of our apps is built by a single developer, which is crazy. And these are not like little apps."
— Dan Shipper (00:04:10)
Every's Metrics (AI-Native Company)
People
Software Products
Paying Subscribers
Total Raised
Compounding Engineering
The revolutionary framework for AI-native development
"In traditional engineering, each feature makes the next feature harder to build. In compounding engineering, your goal is to make sure that each feature makes the next feature easier to build."
— Dan Shipper (00:09:00)
Watch definitionThe Compounding Engineering Loop
Plan
Create detailed plans before delegating to agents
Delegate
Tell agents what to do
Assess
Tests, trying, agent review, code review
Codify
"The money step"
Compound learnings into prompts
"The last step... the most interesting one is codify. This is kind of like the money step which is where you compound everything that you've learned... back into prompts."
— Dan Shipper (00:09:45)
What Codification Looks Like
- Turn tacit knowledge into explicit prompts
- Build organizational prompt libraries
- Spread learnings across entire company
- Each feature makes the next feature easier
Non-Obvious Benefits
The surprising second-order effects of 100% AI adoption
Parallel Work Execution
Work on multiple features and bugs simultaneously
"The reason we can go much faster is we can work on multiple multiple features and bugs in parallel." — (00:06:11)
Demo Over Memo Culture
Prototype in hours, not days. Show, don't tell.
"You can vibe code something in a couple hours that shows the thing that you want to make." — (00:07:26)
Tacit Knowledge Becomes Explicit
Point AI at teammate's repo, learn their process
"Point your CloudCode instance at the repo from the developer sitting next to you and learn the process." — (00:11:04)
New Hires Productive on Day 1
All company knowledge in their prompt files
"On the first day they have all that set up in their CloudMD files." — (00:12:15)
Managers Commit Production Code
Even the CEO contributes between meetings
"Managers can commit code. Even the CEO can commit production code. I have committed production code over the last couple months." — (00:15:08)
No Stack Standardization Needed
AI translates between languages and frameworks
"We let everyone who's building different products pick the thing that they like best." — (00:14:36)
Moving Up the Stack
The fundamental shift in how we write software
"We're moving from Python and JavaScript and scripting languages up into English."
— Dan Shipper (00:08:40)
Watch paradigm shift explanationThe AI-Native Vision
"A single engineer should be able to build and maintain a complex production product."
— Dan Shipper (00:16:31)
Watch vision statementActionable Takeaways
How to apply these insights today
For Engineering Leaders
- Aim for 100% AI adoption — 90% creates drag
- Implement the 4-step loop — Plan → Delegate → Assess → Codify
- Build organizational prompt libraries — Codify all learnings
- Enable demo culture — Prototype in hours, not days
For Founders
- Target single-developer products — 1 engineer per complex app
- Embrace fractured workflows — 3-4 hour focus blocks obsolete
- Drop in expert freelancers — DJ model for specialists
- Remove stack standardization — Let AI translate
For Engineers
- Work in parallel — Run 4+ agent panes simultaneously
- Obsessively codify learnings — Build prompt libraries
- Learn multiple stacks — AI will translate between them
- Contribute across products — Submit PRs to any app
For HR & Operations
- Redesign onboarding — Day 1 productivity target
- Hire for AI aptitude — Technical skills matter less
- Enable cross-team contribution — Anyone can work on anything
- Share all knowledge openly — AI reads any repo instantly
"We're moving from Python and JavaScript up into English."
— Dan Shipper, CEO of Every (00:08:40)
Watch Full Talk (17 min)Video Reference
Dispatch from the Future: Building an AI-Native Company
Dan Shipper, CEO of Every
Duration: ~17 min
Event: AI Engineer Summit 2024
Video ID: MGzymaYBiss
Speaker: Dan Shipper