Discussion about this post

User's avatar
Naomi Thomson's avatar

My favorite rule of the above 3, the "Agency Audit". We have amazing tools in AI. I agree with your foundational statement, there are fewer things I can do better than AI. Therefore, it is necessary to structure the human role in each AI task. No one wants abdication.

Expand full comment
Hassan Abba's avatar

Interesting read Justin.

Fully agree that competitive advantage shifts from technicalities to purpose and accountability which are human traits.

Today I see two technical limitations that still need to be overcome -

1. Hallucinations - required in systems and outputs where precision is required.

2. World Model - LLMs don't operate like you and me. The know which words predict which other words in the context of words. They don’t have what people like Yann LeCun or Judea Pearl would call a world model. And I believe this is where Vision, Reputation and Will may sit.

So for tasks where precision is not required or there is no need for a world model (protein folding) Ai excels well beyond human capabilities. The world we live in is a gnarly place - humans since the birth of civilization have tried to standardise and control it.

Expand full comment

No posts