AI

Tools in daily rotation

• ChatGPT — quick questions; voice conversation in the car; side quests and curiousities
• Claude — Cowork for shared brainstorms, deeper dives, property CMA research; Claude Code for building web apps (and this site)
• Gemini — powers this w
• Anthropic API directly — powers this website's 'fit check' and 'ask ai' features plus other projects
• Notion — notes, drafts, crm, writing, journal, etc.

What I've built

• jonlinpei.com. Fun project for getting me grounded in vibe coding and website plumbing. A lot of overkill for a personal website, but built a few things from scratch - writing system, admin shell and the 'ask ai' / 'fit check' features.
• singysongbook.com. Passion project. Improve your language skills by learning how to sing those catchy songs you don't understand! I honestly wish I had this when I was an expat working in China and would go karaoke with friends and colleagues. Listen, Learn and Sing!
• atriumhb.com. Coming soon. Preference modeling, decision support, and collaboration space for Home Buyers and Real Estate Agents.
• A running log of what I'm trying lives in the writing — the Build log and Deep dives series are both about this.

What I'm thinking about

• What will jobs become? Not 'What jobs will be left?'. Every step function change in productivity from technological advancement jarringly eliminated whole categories of jobs, but also ushered in new ones. There will always be doomsayers. AI legitimately scares the bejeezus (bejeezes? bijeezez?) out of me, but it's also ridiculously empowering.
• What to learn next? Everyone now has a personalized private tutor. If you're having trouble learning something, move a bit upstream and explore how you learn best. Customize your learning to be most efficient for you. Break it down into components, show me a diagram, explain it to me like I'm five, etc. Keep a memory file of your preferred learning and engagement style / method. Target high level understanding, mid-level fluency, or low-level mastery depending on topic and application scope.
• What to build? With the drastic reduction in the cost of code, the bottleneck to innovation widens. On the application layer, you've just opened the top of the funnel to anyone, anywhere who is willing to put in as little as a single day to start learning how to build niche applications for use cases people would previously not touch. The SF Bay Area is arguably the tech and innovation hub of the world mainly for the concentration of talent and ambition. However a drawback to this risk of living and thinking in a bubble.
• What will people differentiate on? With the intersection of customized knowledge delivery and accelerated create / build capability, you're now mostly only limited by energy - for you and your devices. There are obviously other resources that people will still need to create and build in the physical world, but even those will be impacted downstream from innovation. Double down on athletics, art, music, taste, leadership, etc.?

Where I'm skeptical

• Putting the AI cat back in the bag. Or even trying to stall it will be an exercise in futility. You may buy yourself a little bit of time, but in a most Borg-like and dystopian manner, resistance is futile. I'm not saying people can't shut AI technology down if needed, but rather that people won't.
• The idea that traditional trades will be the safest jobs in the future is a bit misleading. It is true that the advancement of AI technology is impacting knowledge and tech workers the most, and that the ability to reduce the amount of software engineers or bookkeepers you have doesn't tie directly to the demand for more 'physical' labor. However, that's stopgap reasoning. Advancements in robotics will naturally accelerate downstream from AI.