What’s New?
A fascinating project called CleanSlate has popped up on Hacker News — it offers a collaborative workspace for AI with a shared procedural memory. In plain English, this isn’t just another AI tool, but an attempt to build a universal “brain” for AI assistants, where skills can be stored, refined, and reused.
The project’s source code is already on GitHub (link at the end), and this approach could seriously change how we automate, train, and collaborate with AI.
Let’s break down why this matters, how it benefits AI creators, marketers, and entrepreneurs, and how you can put these ideas to work.
How “Shared Procedural Memory” Works
What Is It?
Most AI tools today are task-based: generate some text here, create an image there, analyze some data somewhere else. But each tool lives in its own bubble — it doesn’t “remember” what you did before, and it can’t learn from past actions within your team or project.
CleanSlate changes that by creating a workspace where AI assistants can:
- Remember sequences of actions (procedures)
- Store these procedures as reusable skills
- Make those skills available for future use
- Share skills across users and projects
Why Does That Matter?
Right now, every time you teach an AI a workflow — maybe scraping leads from a site or preparing a custom report — you’re doing it from scratch. Your AI doesn’t retain your style, your steps, or your nuanced processes.
With CleanSlate, you can save the whole workflow as a skill, so it’s ready to run again — or share it with teammates, collaborators, or clients.
Who Should Care?
AI Creators
- Build your own “library” of tricks and workflows, ready to drop into new projects.
- Experiment with new tool combinations without constant repetition.
- Assemble your custom AI stack using a “copy-paste” approach.
Marketers
- Create templates for analytics, content generation, social media workflows, and run them on repeat.
- Share working processes with colleagues, setting unified standards across your team.
- Implement features quickly, whether you found them elsewhere or built them yourself.
Entrepreneurs
- Standardize AI processes across your company — new hires or contractors don’t have to reinvent the wheel.
- Save time and training costs: one person creates, everyone else uses.
- Protect your company’s know-how by accumulating procedural knowledge.
How This Could Change the AI Landscape
1. AI Goes Truly Collaborative
Until now, AI assistants have been personal tools: you set something up, but repeating that for another team was tricky. Shared procedural memory lets you build real AI teams, where everyone can contribute to a collective “brain.”
2. Building Complex Pipelines Gets Easier
You don’t have to manually assemble workflows every time. Just grab the skill you need from your library. It’s like building with LEGO bricks instead of starting from scratch.
3. AI Adoption Gets Much Faster
Instead of slow rollouts and endless customization, companies can immediately use proven skills — whether from previous projects or shared by others — scaling best practices quickly.
How to Use This Right Now
Example: Marketing Automation
Say you’ve built a workflow: brainstorm post ideas, generate copy, pick images, publish to social. With CleanSlate, you save this as a skill — reuse it for different products, share it with teammates, or even sell it as a template.
Example: Team Training
Instead of lengthy manuals and verbal explanations, you hand new hires a ready-made set of AI skills. They start working to your standards immediately, with fewer mistakes and no wasted time.
Limitations and Future Potential
Right now, CleanSlate is early-stage: it needs some technical know-how and isn’t “polished” for everyday users. But the direction is promising:
- Expect more user-friendly interfaces and skills marketplaces soon
- Skills sharing between companies could become standard practice
- AI assistants will get genuinely smart — not just at “learning,” but at reusing knowledge
What to Do Right Now
- Explore CleanSlate’s source code — go to GitHub and see how procedural memory is implemented.
- Audit your AI workflows — identify which processes you repeat often and could automate as reusable skills.
- Map out a skills library — list procedures that you can turn into templates for your team.
- Experiment with skill sharing — try sharing your best workflows with colleagues, or adopt someone else’s.
- Keep an eye on the market — watch for new projects and tools that support reusable AI skills.
More hands-on AI tools and templates — in my courses.
Based on materials from Hacker News



