For twenty years I've dreamed of building software where users truly control their data and how it shows up in their lives. I never could — the scope was too vast. Two more public efforts are Mitch Kapor's Chandler, Google Wave - ambitious projects with real funding that couldn't pull it off.
This is the year AI changes that. I'm building a Collaborative Conducting Environment — a new kind of tool where you and AI work as partners, complementing each other's strengths and covering for each other's weaknesses. It's not about sending the AI off to do your work. It's about working together.
I'm already using my CCE for 100% of my AI software development, not Cursor, not Claude Code, my tool - because for my workflow it's already radically more productive. I could explain what it *is*, but those would be wasted words, as it changes every day, and I can't explain what it will become, because it's a new kind of thing. Every explanation is based on expectations - and that makes them fall short. You have to use it to understand it.
What I can explain is what I believe and why I'm building it.
Memory Safe Software. Now.
Every week, the software you rely on — your browser, your PDF reader, your phone — gets hacked because it's built with tools that let attackers reach into your computer's memory and take control. These aren't rare events. Chrome alone patches hundreds of these vulnerabilities a year. Your bank, your photos, your passwords — can all be stolen because we keep building software the same way we did in 1972. The answer isn't patching faster. Garbage Collected Memory Safety removes entire classes of vulnerabilites that are present in C, C++, and yes, Rust.
In 2024, our highest office issued a call to action: The White House: Future Software Should Be Memory Safe. I'm answering that call. I am already working in my Collaborative Conducting Environment, well on my way to deliver memory-safe versions of the critical software we all use — a web browser, PDF viewer, image editor, and AI collaboration studio, shipping on all platforms. And I'm doing it by myself. One programmer. Incredible AI leverage. Because I want people to see what AI is truly capable of when you work with it as a partner, instead of sending it off to work alone.
White House: Path Towards Secure and Measurable Software: Final-ONCD-Technical-Report.pdf
AI-Sovereign Software. Yours.
In 1985, Richard Stallman published a radical vision: users should control their software, not the other way around. The GNU GPL and the open source movement that followed reshaped the industry — but forty years later, the software most people actually use is still proprietary. Windows, macOS, iOS, iMessage, Snapchat, Instagram. Open source won the server room and lost the living room.
In the near future, every piece of software and data becomes more powerful when deeply integrated with AI. But that creates a tension. The history of software is locking you in for revenue. As users, none of us want software that captures. We want software that does what we want.
I'm building a configurable AI-embedded platform where you own your data and control your UI — email, messaging, writing, artistic work, coding, everything. Not designed around revenue capture. Designed around you.
The last piece of software you need, because it becomes whatever you need it to be.
http://artificialnecessity.com
My browser engine is already rendering complex HTML5 with Chrome-reference conformance across dozens of layout cases involving the all major layout features.
Here are some recent HTML5 conformance of Chrome vs my FluidDOM, 100% memory safe down to the GPU pixel drawing.
..and here is a report pull of my git-hub repos..
No comments:
Post a Comment