Hacker News new | past | comments | ask | show | jobs | submit login

The data model question is - at best - a very poor choice of words. Implicitly assuming a data model is a bad idea as the data model is one of the major choices every program has to contend with. So implicitly assuming it isn't really a desirable goal. It is like types where explicitly having information on types in key places in a program is always desirable and the only question is how pervasive it should be.

I also liked "They will instead ask questions about data in a natural language". Humans can't answer questions about data in natural language - when data questions come up they have to be couched in formal terms and statistics. I look forward to the day an AI can answer "what is the best decision we can make with this data?" but it runs into a problem we already have in spades. If the asker of the question isn't thinking in formal terms I question whether they are going to be able to consistently pose meaningful questions and interpret the answers. Applied Math is not easy to implement in industry because a company's ability to perform is limited by its management's understanding.

It is easy to nitpick. I'm also glad we have people thinking about the big picture of how data gets handled.




Consider applying for YC's Spring batch! Applications are open till Feb 11.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: