Just had Claude Code draft a book proposal and it had a very similar voice to this post. I wonder if there's a way we should refer to this particular style.
I love the orchard and the fruit frame, and the ownership question underneath it which I'm interested in exploring too. If the economy becomes the AI's training environment, the location of the jobs matters less than their ownership. I think you're right about that.
Where I'd hedge, as someone observing vs forecasting: you look for the new ladders inside the knowledge-work stack that AI is eating, and I wonder if they could forming elsewhere. The buildout you open with (chips, power grids, construction) funds jobs whose apprenticeships are intact; nobody can skip the junior electrician or the fab technician or the civil engineer. Likewise, physical AI also needs experience-data from the world rather than scraped text, and someone has to build and supervise that collection/curation. Specialist science may see an explosion if I follow Demis and others who are bullish. If AI expands what's worth testing, the wet-lab jobs grow rather than shrinks, because the validation still has to be run by humans that are learning. And formation-native work like elder care, live performance, or edge roles can't be seniorised, because the formation is the point.
So my reading is that the ladder relocates rather than vanishes — off the clerical rungs you're rightly pointing to and onto physical, experimental and relational ones. But I hold this loosely, because your deeper point is valid even if I'm right about location: the new nurseries (data annotation, agent supervision) could be at risk of becoming gig work with better names, the learning flows upward to whoever owns the platform. I'm more hopeful than you on where the rungs appear. I'm not at all sure who ends up owning them.
Just had Claude Code draft a book proposal and it had a very similar voice to this post. I wonder if there's a way we should refer to this particular style.
https://summerlightning.substack.com/p/llms-pre-commodify-ideas?r=103&utm_medium=ios
After reading this I feel oddly connected to this simultaneous innovation vibe, there are worse parts of history to be part of
I love the orchard and the fruit frame, and the ownership question underneath it which I'm interested in exploring too. If the economy becomes the AI's training environment, the location of the jobs matters less than their ownership. I think you're right about that.
Where I'd hedge, as someone observing vs forecasting: you look for the new ladders inside the knowledge-work stack that AI is eating, and I wonder if they could forming elsewhere. The buildout you open with (chips, power grids, construction) funds jobs whose apprenticeships are intact; nobody can skip the junior electrician or the fab technician or the civil engineer. Likewise, physical AI also needs experience-data from the world rather than scraped text, and someone has to build and supervise that collection/curation. Specialist science may see an explosion if I follow Demis and others who are bullish. If AI expands what's worth testing, the wet-lab jobs grow rather than shrinks, because the validation still has to be run by humans that are learning. And formation-native work like elder care, live performance, or edge roles can't be seniorised, because the formation is the point.
So my reading is that the ladder relocates rather than vanishes — off the clerical rungs you're rightly pointing to and onto physical, experimental and relational ones. But I hold this loosely, because your deeper point is valid even if I'm right about location: the new nurseries (data annotation, agent supervision) could be at risk of becoming gig work with better names, the learning flows upward to whoever owns the platform. I'm more hopeful than you on where the rungs appear. I'm not at all sure who ends up owning them.