Ask HN: What if all programmers in the world decided to contribute to build AI? | Hacker News Reader

Ask HN: What if all programmers in the world decided to contribute to build AI?

7 pointsbehnamoh7 comments

It's a simple [hypothetical] question: what if instead of spending their time on side-projects and hobbies, all programmers in the world would decide to contribute to a global AI project? Would we achieve true AI then?

View on HN

I don’t think we need programmers. I think for this task we’d need neuroscientists, philosophers, psychologists, mathematicians and engineers. Also, I don’t think we need more developers working on that problem. We need the right people with the right motivation and cognitive abilities to solve the problem of Artificial General Intelligence.

I think one of the major stumble blocks on achieving "true AI" is that there is no consensus over what constitutes "true AI".

I suspect that --As long as AI can be perceived as a potential threat to the human ego-- the goalpost will perpetually be moved to just outside of the state of the art in AI research.

Leaving aside the co-ordination problems of what needs to be done, and who does it, imagine all the time that would be wasted on debates over which environments, APIS, and formats to use...

Turn your question around. Hypothetical: at some point in the future, we'll be able to say "AI emerged." At that point, will everyone who contributed to internet architecture be partially responsible for that? Or did all human coding efforts contribute?

If this has already happened, would we know?

Coordinating work is hard. In addition to the programmers, you would need so many more people to coordinate and plan. And doesn't even include how to figure out what to produce in the 1st place.

Large projects are incredibly difficult. That's one reason why things like the space program and its achievements are such a big deal. It wasn't just the cost and knowledge but the coordination of 100's of thousands of people to get the job done.

A number of programmers prefer writing programs they do understand, and still these programs can do incredibly interesting and useful thinks: theorem proving or verification, finding optimal solutions to all kind of problems, ...

AI as it is now is merely brute force: huge datasets, huge computers. Of course it yields results: that would be a shame it would not using so many resource.

First, we need to define what "true AI" actually means. From all the discussions I've taken part in, and everything I've read, we don't have such a definition.

We don't even have a good definition for "intelligence", artificial or not.

Scale doesn't always work, 10 women can't give birth to a baby in 1 month.

Reply on news.ycombinator.com