Most failures are decision failures. One essay a week on the discipline of deciding well: judgment, choice, strategy. The Decision Papers — a book in progress.
Our team @a16z calculated AI adoption per capita across the world.
The results were surprising. The U.S. leads AI development...but it ranks down at #20 in adoption.
At the top? Singapore, Hong Kong, the UAE, South Korea, and much of Europe 🤯
It is that, and it is also poisoning the data lake which in return due to Markov chain being vulnerable to feedback loops is biasing the future models. We really have to understand the Information Space better and the dynamics how feedback loops work in scale.
Policy makers have been focusing on regulating the algorithms; largely ignored importance of both the training data and the training methodology. In Reinforcement Learning, reward/penalty teaches the models to ignore reasoning paths which may be the best for the case in hand.
This also raises questions about credit and responsibility.
When a tool does part of the writing, does the “author” become an editor, a curator, or a decision-maker?
Tools don’t remove responsibility—they make it more visible.
On the surface, AI looks like creation.
We design architectures, collect data, train models, deploy systems.
That feels like sculpting something from nothing.