I've tried every productivity system out there. Notion, Obsidian,豆腐, supernotes, AI assistants. You name it.
The promise is always the same: capture your knowledge, retrieve it when needed, think better.
The reality: mostly just another place to store things I never look at again.
So when I heard about Floatboat's approach to AI-assisted work, I was skeptical. Another "learns how you work" claim. Been there.
But the architecture behind Floatboat is actually different from what I've seen before.
What makes Floatboat different
Most AI assistants are stateless. You prompt, they respond, conversation ends. Next day, you're starting over.
Floatboat's Tacit Engine™ maintains a persistent working model of how you work — your patterns, your preferences, your decision criteria. This isn't a knowledge base. It's more like muscle memory for your cognitive process.
Combo Skills: turning process into reusable automation
Here's the part I found most useful: you can capture complex, multi-step workflows as "Combo Skills" and apply them across different contexts. Instead of doing the same 7-step process manually every time a new client comes in, you build the skill once, and Floatboat handles the mechanical parts.
Selfware: context-aware output generation
Instead of generating text from scratch, Selfware generates in your style — based on what it has learned about how you write, decide, present. Voice notes to decks. Research to articles. Rough thoughts to platform-ready posts.
The funding
Sequoia China and Weiguang Venture Capital just led the seed round. That's worth noting if you track who's funding the next generation of AI-native tooling.
The full article is on my Medium. Link in comments.
Top comments (0)