Personal AI Manifesto

The Personal AI Manifesto

What we want

1. We want an AI that works for us, free of any conflict of interest—present or future.

2. We want an AI that is resilient and won't stop working due to technological failure or geopolitical policy changes.

3. We want an AI that won't give us the same generic answer it gives everyone else.

4. We understand that owning our data is essential to our personal sovereignty.

5. We understand that what we see, read, and experience digitally shapes how we think—and we want to stay in control of it.

6. We recognize that technology has always required trading some personal sovereignty, but we believe this should be a matter of personal choice.

7. We believe that those who own at least part of their AI will stand out and gain professional and personal advantages.

8. We believe that open standards create economic value, and AI needs more openness to reach its full potential.

9. We believe centralized AI won't lead to the future we want—and we're here to challenge the status quo.

What I’ll do as pet project in 2026.

1. Help individuals own their data.
Build tools to download messages and contacts from all messaging apps and social platforms using official APIs or alternatives like scraping.
Build tools to download health data, financial data, you name it. Wherever data lives, we need to give it back to the individuals to own.

2. Help individuals make all their data AI-friendly.
Build tools that process various file formats and translate them into natural language so any LLM can understand them.
This will be the centerpiece of personal AI. All of a user's context in one place. This will enable new uses, new services, and boost innovation, competition, and economic value.

3. Help individuals control what they see online and when they see it.
Create tools that watch social feeds on their behalf, filter the content, extract the useful parts, and synthesize them. Make this content available as part of the user's data so their personal AI can use or react to it.

4. Help individuals think for themselves and get better AI than the generic ones.
Let individuals choose their trusted sources and define their own system of beliefs for their AI to use when answering questions.
Build tools to collect materials from trusted sources, make theses trusted source part of vaults.
Incorporate this context in every personal AI tools.

5. Help individuals build their own AI agents.
Create standards and infrastructure to describe and deploy agents, locally or in the cloud.
Encourage vibe coding—using AI on available personal data so someone can build the tool they need, when they need it.
Create a market for personal AI agents dedicated to processing personal data in secure and trusted environments.

4. Build resilient, personal AI.
Create tools that don't rely on a single technology or LLM provider, with local LLMs as a failsafe.
Build a system that fits on a portable hard drive, containing all your data in human-readable format, with AI agents able to run locally for the most important tasks.As AI becomes an important part of your life, you want to carry it in your pocket.