My Position on AI
I am actively exploring AI across three main areas:
- Personal productivity and workflow integration
- Societal impact, risks, and long-term implications
- Practical applications (automation, agents, service delivery)
This section captures both opportunities and concerns, with an emphasis on developing a balanced, usable perspective rather than reacting to hype.
Key Themes
AI: Opportunity vs Threat
My starting point for thinking about AI is captured in ai - opportunity or threat. This explores both the potential benefits and the psychological/social risks, including how engaging with AI can influence mindset and behaviour.
Human in the Loop (HITL)
A recurring theme is that AI systems require human oversight to remain safe and useful.
Relevant notes:
AI Slop and Quality Control
Not all AI-generated content is useful. I am tracking the rise of low-quality outputs and the need for filtering and validation.
Intentional Use of Information
AI sits alongside a broader goal of consuming information deliberately rather than passively.
Active Projects
AI-Enabled Workflows
I am designing a structured workflow for integrating AI into note-taking and execution:
- 0 AI enabled note taking
- 0.1 CODE workflow
- Current constraint: 2 - Organise
- Key challenge: constraints - cross vault context
Automation and Agents
Exploring tools like n8n to build AI-driven processes and agents that can perform tasks autonomously.
Initial areas of focus:
- Healthcare use cases
- Personal finance (PFA)
- Information curation (news, content filtering)
Tools and Models
There are many competing AI models, with a growing geopolitical dimension (US, Europe, China).
I am developing a working understanding of which models to use for different tasks:
Current tools in use:
- Mammoth.AI
- OpenRouter
- Gemini (as a secondary option, particularly where personal data integration is useful)
Prompting and Interaction
Prompting is a core skill for effective AI use and requires deliberate practice.
This is an area for continued improvement.
Open Questions
- Where is AI genuinely useful vs overhyped?
- How much human oversight is enough?
- What workflows are actually sustainable long-term?
- How can cross-vault context be effectively managed?
Index / Navigation
This note acts as a central entry point into all AI-related thinking.
Over time, this will include embedded summaries from key notes to improve visibility and retrieval.
Potential future structure:
- Embedded summaries of core notes
- Rolling updates on tools and models
- Consolidated views of projects and experiments