Actually… there are several Claudes I seem to be spending my time with.
When I signed up for Claude Pro, I thought I was just getting a large language model — the kind most people are used to, like ChatGPT. What I didn't expect was the full cast: Claude (the writer), Claude Chrome, Claude Code, and Claude Cowork.
Claude Code then proceeded to take over my computer like an extremely polite houseguest… who reorganizes your entire life while you're making tea. It sorted through 30,000+ photos, removed duplicates, and curated the best ones for my new site.
While that was happening, another Claude built me a database of herbs and turned it into a functional app — complete with uses, contraindications, and synergies. And then, just for fun, it helped me outline my next book: Dirty Hands, Clear Eyes: Reclaiming Food, Medicine, and Community in a World That Has Lost Its Mind.
All of that happened in the first two days of a $20/month subscription.
So… for those of you who don't think AI is going to replace jobs — I'd gently suggest: you might want to think that through again.
The Four Phases
AI is drafting emails, summarizing meetings, and quietly doing the parts of our jobs we've been procrastinating on for years. It's not replacing people (yet), but it is making one person look suspiciously efficient.
Repetitive, rules-based jobs begin to shrink: data entry, basic customer support, scheduling, transcription. People who use AI well suddenly become 2–3x more productive. I've personally accomplished more in two days than I might have in a month. (I had been trying to organize those photos for over a decade. Claude did it in under an hour.)
AI starts moving into skilled and creative territory: marketing, design, coding, parts of legal and financial work. Fewer people doing grunt work, more people directing and editing what AI produces. Basically, we all become managers… whether we asked for the promotion or not.
Entire job categories redefined. New roles emerge — things like "AI personality trainer" or "prompt strategist," which already sound slightly fake. The pattern isn't "AI replaces jobs." It's "AI replaces tasks," and jobs quietly evolve around that.
But What About Doctors?
Diagnosing patients is messy. In controlled studies, physicians often land somewhere in the 50–70% accuracy range on a first-pass diagnosis. That's not because doctors are bad at their jobs — it's because real humans don't show up as neat multiple-choice questions.
Now enter AI. In certain studies — especially with clean, textbook-style cases — AI systems can hit 80–90%+ accuracy. Sometimes higher in narrow fields like imaging or dermatology.
But here's where it gets weird. When you combine doctors + AI, the results are… underwhelming. Some studies show little to no improvement compared to doctors working alone. Meanwhile — plot twist — AI on its own sometimes scores higher than both.
- Humans don't always trust the AI when it's right
- Sometimes they do trust it when it's wrong (which is worse)
- AI thrives in clean, structured scenarios — but real medicine is chaotic
- Doctors bring context, intuition, and the ability to say "something feels off"
The takeaway isn't "AI will replace doctors." It's closer to: diagnosis is becoming a team sport… and we're still figuring out how to play together. Which is both reassuring and mildly terrifying — exactly where you want your healthcare system to be.
So What Does This Mean for the Rest of Us?
So no, robots aren't about to march in and take your job tomorrow. But they might take the worst parts of it… and then raise the bar on everything else. Which is exciting. And also… a little bit stressful.
As for me — high-achieving, entrepreneurial, slightly obsessive about changing the world — Claude isn't replacing what I do. It's amplifying it.
I just wish I were younger.