The declining potential for agency with LLMs

It’s fascinating watching Ethan Mollick get gradually depilled as each successive generation of frontier models minimises the potential role for human agency:

Importantly, it was just limited in how much work I did relative to the model, it was also limited in how much control I had over how the model did things, why the model chose particular approaches, or even how in-depth its results would be. The details of the AI’s decision making are not shown to me, and the process would be too long to even be worth following. The map required the AI to make judgement calls about hundreds of little choices, and it just made them, without me understanding the choices or having a chance to weigh in. In many ways, it is miraculous (I can always ask for edits at the end) on the other, it turns AI into the ultimate black box.

Last year I called this working with a wizard: you chant the spell and something happens. With Fable the spell has gotten powerful enough that I am no longer sure I am the wizard. I am closer to a patron. I describe what I want, I pay for it, and I judge the result. The conjuring happens somewhere I cannot watch, in hundreds of small choices I never get a vote on. The work has shifted from process to outcome. I no longer steer; I commission.

Honestly I guess I’m going through a similar process, with the exception that I was never as utopian about the underlying capabilities of the technology. But the process he describes here, which Milan Sturmer and I talk about as reducing the burden of articulation, should be seen as the key shift taking place in the models as they diffuse. It’s just not possible to exercise sustained agency over the emerging models in the way it was over say 2025 era Claude and ChatGPT models. This captures something very significant I think:

A patron commissions a single artist. Fable is closer to a whole studio, where I am the client who signs off on the final work without ever setting foot on the floor.

“Given the jagged frontier of #AIability, the complexities of #AIadoption, and the limitations of AI, I believe there is still a lot of room for humans to not just use AI, but to use AI to thrive. So, I decided to write a new book: #CoExistence, about how to work with AI that is sometimes, but not always, better than you.

The first question you might ask is how did I work with AI as I wrote this #book? I think the results are pretty indicative of the state of AI today. Sometimes I used AI a lot, sometimes very little, and sometimes I had to let the AI do what it wanted.”
#EthanMollick
https://www.oneusefulthing.org/p/co-existence-and-the-end-of-co-intelligence

AI 연구자 Mollick의 현재 진단, 지금이 형태를 결정할 마지막 창문

https://aisparkup.com/posts/10102

같은 모델인데 왜 결과가 다를까, 인터페이스가 AI 성능을 결정한다

https://aisparkup.com/posts/10745

New blog post: Co-working and 25 years of eggs – The Distant Librarian https://distlib.pival.me/co-working-and-25-years-of-eggs/

#distlib #EthanMollick #Co-intelligence #Co-working #OCR

Ethan Mollick이 말하는 에이전트 시대 AI 선택법, 모델보다 하네스가 중요한 이유

https://aisparkup.com/posts/9382

Ethan Mollick이 본 Claude Code, AI 혼자 스타트업 만든 74분

https://aisparkup.com/posts/8087

Helping people write code again

#ai, #ai-agents, #ai-assisted-programming, #careers, #coding-agents, #ethan-mollick, #generative-ai, #llms

An Introduction to Artificial Intelligence in General, and Generative AI in Particular