24

Proposal: Artificial Intelligence

According to the blog post "No Artificial Intelligence in Area 51", there once was an AI site which was closed down on 2010-12-22.

StackOverflow generously describes the problems encountered and the reasons for the shutdown of ai.stackexchange.com and even writes

AI has become the new baseline for “failed to make it out of private beta.”

This is alarming, to say the least.

Looking at the blog-post descriptions of problems that were encountered back then… what exactly will prevent this proposal from crashing into the same problems again?

EDIT – Feb. 25, 2014

Just an update to note what happened… After 18 days in beta, the AI site didn't have enough activity during the beta, and has been closed. In the end, the private beta had 2.8 questions per day, 80% answered, 21 avid users, 434 total users, a 2.0 answer ratio, and a mere 111 visits/day. That’s what I was afraid of and the initial reason why I asked. Turns out, my worries made sense.

2
  • 1
    Regarding your edit: the questions seemed much better than in previous betas. Also, new AI questions kept coming simultaneously on SO and other sites. Perhaps AI wasn't advertised enough - with AI questions asked on these other sites it would look much better.
    – BartoszKP
    Feb 25, 2014 at 13:34
  • AI 2016 restart
    – kenorb
    Aug 10, 2016 at 11:15

7 Answers 7

Reset to default
18

According to the post No Artificial Intelligence in Area 51 the main reason for the failure of the last proposal was the lack of expert-level questions:

It wasn’t so much the lack of questions that was of concern — a site can stay in beta as long as it takes — but the conspicuous lack of expert-level questions. This was also the emerging opinion amongst the users:
[...]
70-80% of the questions didn’t run much deeper than “When will we have intelligent computers?” and “What is your favorite AI blog?”


What can we do to make it work this time?

I think the key to getting over that obstacle this time is to make sure we have lots of people actually working with A.I. on board in the early phase. We should:

  • invite all the people working with A.I. we know
  • ask expert-level questions ourselfs
3
  • 2
    I guess we should try to downvote and close those low-quality off-topic questions ASAP.
    – Simon Forsberg
    Dec 13, 2013 at 18:11
  • This is good as far as it goes but I think you should include the flip side of the coin: you should aggressively close questions that experts (at least not ones that you are trying to attract to the site) would not be interested in fielding.
    – Caleb
    Jan 29, 2014 at 13:31
  • You need a combination of Experts, advanced users, and beginners asking quality questions. Experts alone will not make the site, you need people asking questions for the experts to answer.
    – Chad
    Feb 25, 2014 at 15:55
4

As kdzia cited, the post that announces the end of the AI site details that the questions really didn't concern practical Artificial Intelligence projects. Considering how many people work on artificial intelligence related problems, their are plenty of people who know enough about AI to ask proper questions. If the people that get on board here participate well, this should not be a problem, especially if a lot of people get involved. Those who are legitimately working on artificial intelligence may have the responsibility of discouraging questions that are off topic.

4

What if we were to rename it into computational intelligence while doing so would exclude the expert programs used in many games the term would by virtue of being more technical prevent this kind of people from entering the website.

4

Don't give it a 'Hollywood' title like 'artificial intelligence'. Call it "Machine Learning and Intelligent Computation

3

There are plenty of places out there that discuss AI-related topics on regular basis. SO sites have a long-standing respect from the community and I believe that if we make a slight effort to advertise this start-up, it'll pay off in the end as a flourishing and informative information source.

My C++ : Data Structures instructor teaches AI at UCR, so I will ask him to spread the word as well.

1
  • Great… the more, the merrier.
    – e-sushi
    Dec 2, 2013 at 19:01
2

I'm advertising the link to this proposal in my SO profile, and try to remember to give a link to it every time someone asks a question related to AI on SO (EDIT: well I had it coming, and finally got chided for spamming, so I won't do this any more :) ). Perhaps this will help a bit.

On the other hand (referring to Gilles comment), maybe it's just a sad (?) fact, that scientific AI has just too small audience. If questions like "Is there any effective way to make an AI self-aware?" get 5 votes easily and occupy the front page then it suggests that "Science fiction" would be a better proposal to start (or let me guess - it already exists, and made it successfully out of the definition stage).

3
1

The new A.I. proposal is now restarting, so please join private beta.

We hope it won't be closed again.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .