At one point I thought I got it, but then I lost again.

Few highlights:

Obvious points are:

What about AI implementation and modelling? Above quotes are a bit contradictory.

So what's on-topic exactly, AI modelling or implementation, or none of it?

If none of it, what should be?

  • 2
    $\begingroup$ I think we should also be more clear on what we mean by "implementation", since it seems to me that not everyone can agree on a definition. In my opinion, programming questions and questions about the inner workings of basic algorithms (e.g. Viterbi, backpropagation, etc) should be definitely off-topic here. But on the other hand, if we want to be able to discuss AI, some basic level of knowledge about how these things work is certainly necessary. $\endgroup$ Commented Aug 4, 2016 at 17:04

4 Answers 4


I'm sorry, but we can't just go with a simple, blanket statement like "Programming, algorithm, modeling, math, philosophy, and history questions should the off-topic, as they are already on-topic in other SE, such as Stats and Data Science." Why? Because not all questions about "programming, algorithms, modeling...", vis-a-vis Artificial Intelligence, are on-topic at those other sites! But they are here.

And what's the distinction that should be in play? Well, simply, "programming, algorithms, modeling, math, etc. that are *specific to AI" are on-topic here. It really can't be any other way.

I mean, think about it... we claim to be a "science" site, but then try to say that "math" is off-topic? That's absurd. Science is math and math is science. Or to put it another way "math is the language of science".

If we keep pushing this idea that all hard technical questions are off-topic, all we're going to get are vague questions about speculative aspects of AI, with answers that are nothing to speculation and hand-waving.

What should be on-topic? Questions about Artificial Intelligence, full-stop. It's right there in the name on the marquee sign, as they say.


A for loop is not the same as gradient decent. Gradient decent is not the same as NN convergence, or generalization. You cannot do the latter without the former.

Biology is really Chemistry. Chemistry is really Physics. Physics is really math. If we required doctors to work through the math, then the quantum physics and molecular electronics, then the chemistry before they could do medicine they would die of old age first.

I think that we are trying to separate the fields of computer programming, data science, and such into layers of abstractions. Each layer has to be thick - to stand on its own and properly envelope its content.

Right now, Machine Learning and Artificial intelligence are "young" so there is going to be nuts and bolts. If we don't give clean bridges there then this area gets to be a philosophical wasteland - no engineering allowed. If we have some courtesy and we are willing to realize that the divisions are not clean yet, then we can make better mileage toward building a richer community.


Practical applications should be on topic here, at least at some extend to keep proportion between philosophy and implementations.

This is important in therms of building community. It is hard to change course later, and if you will this place to be one of fantasy-futurology-imaginary world building site of SE then strike out practical application.

this question could be asked on WB as well Is consciousness necessary for any AI task? people there are very exited in discussing consciousness.

This question is definitely low quality https://ai.stackexchange.com/questions/1869/deep-neural-network-for-not-so-popular-board-game , for obvious reasons, fresh enthusiastic member with practical question - holding it discourages people who where fresh attracted, attracted by super title Artificial Intelligence, not boring Data Science where you do not know what to expect as Science of Data is big field. But there is short and clear AI, yes I wish one, give me two.

At the moment there is no implementations of AI(in movie sense)(known, for me, not a expert) For that reason to keep track on the ground, practical application should be even if this site is intended to be subset of philosophy about AI. Maybe change name then to reflect that - based on AI.SE I had expectation to see there useful stuff. From PAI.SE I expect nothing, not interested to know will AI kill humanity or not, how it will change perception of humanity about world, whatever.

At least amateur level implementation should be, at least it is bare minimum what you need. Probably even that is not enough and will not work out. You should decide is that philosophy site or is that site about AI, if second then everything about AI(creation, theory, implementation, consequences) should be on topic here - all or nothing, have balls people, it is important for mankind, for our future.

You should discuss more which promises name promises - Artificial Intelligence - and whom it attracts.


Programming, algorithm, modeling, math, philosophy, and history questions should the off-topic, as they are already on-topic in other SE, such as Stats and Data Science.

Data science and the Stats SE already have a huge overlap (>~80%), and I am worried to have a third SE that also significantly overlaps with them. Personally, it would further demotivate me to write any answer, as it gets tiring to copy-paste content, and updating duplicated answers is a pain.

  • $\begingroup$ Maybe instead of writing answers, you can ask questions which you think are off-topic on the other sites (since you've good understanding what's on-topic there) to show some good examples of such questions for others. $\endgroup$
    – kenorb
    Commented Aug 4, 2016 at 18:15
  • $\begingroup$ @kenorb It's hard to come up with such questions :) I see only a handful amongst the existing questions that could qualify, but I am not even sure. $\endgroup$ Commented Aug 4, 2016 at 18:21

You must log in to answer this question.

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