Machine Proxies

I picked this up expecting a tech manifesto, but it's really a thoughtful look at how we've always outsourced judgment—just now to machines. The opening courtroom scene with Michael Cohen is a perfect hook. I do wish the author had spent a bit more time on concrete solutions or regulatory fixes; it's mostly diagnosis. Still, the writing is sharp and the examples (like delegating text messages while driving) hit home. Glad I read it.

This book nails the eerie feeling of realizing your voice assistant isn't just a tool—it's your representative. The central idea of a 'proxy' is well-argued, and I loved how the author connects ancient delegation (ledgers, calendars) to modern AI. My only gripe is that some chapters drag with legal and technical details. But the core question—'who reviews the machine?'—stayed with me long after I finished. Recommended for anyone who uses automation without thinking.

There's a good book buried in here about delegated authority, and the opening Cohen case is genuinely fascinating. But the author repeats the same point over and over: 'we've always delegated, but AI is different.' I get it. I wanted more depth on the actual harms—bias, accountability—and less philosophizing about the 'delegation instinct.' Also, the writing can get a bit academic. If you're new to AI ethics, you'll learn something; if you're not, you might skim.

This is the most important book I've read this year. The author takes a seemingly mundane fact—our AI assistants act for us—and reveals how it's quietly rewriting the rules of law, medicine, and even personal responsibility. The courtroom scene alone is worth the price. But it's the deeper argument about humans as 'delegators' that blew my mind. I've already recommended it to three colleagues. If you care about where technology is taking us, read this.

Really enjoyed the framing of AI as 'proxies' rather than tools—it changes how you think about every interaction with a chatbot or security AI. The historical context (how we've always offloaded tasks) is a nice antidote to tech-panic. The only reason not five stars: I felt the last third could have explored more real-world consequences, like the health chatbot example from the blurb. But overall, a smart, accessible read that made me reconsider my own delegation habits.

As a lawyer, I was skeptical about a book that opens with a court case, but the author handles the legal nuances well. The Cohen example is a perfect lens for the 'proxy' concept. I appreciated the balanced tone—not doom-and-gloom, but not naive either. My minor complaint: the chapter on health proxies felt rushed compared to the legal and security ones. Still, it's a thought-provoking read that I'll be passing to my colleagues who think AI is just a search engine.