In the first entries in a multi-part series I look at how LLMs are grown, not engineered, and the consequences. I use this insight to develop a new way of thinking about science and the nature of progress
When consensus predictions don't pan out, it is always worth noting. That is why it is so striking how little attention it is getting that we had the future of technology dead wrong.
Mimicry is a complex and important topic across biology, markets, social networks, and other domains. It can lead to unexpected outcomes, as well as new sources of fragility.