Hey everyone, please excuse me for this thread being dead on my part. As someone stated it was dead for few months now and only got bumped up few days ago
I’m currently doing a lot of work on ongoing projects and have been checking the forum once per week or so. I will try to answer all of the accumulated questions soon. Really don’t wanna be asshole-ish it’s just that I seriously a bit busy ATM. A bit of patience guys and I hope I’m not pissing u off to much rn
How big does data need to be to be considered 'big data'
?
Hahahahaha well I would say it doesn’t really matter but the general idea behind approach to “big data” is that at some point if you don’t implement efficiently you run out of computational resources to do what you are doing (sure u can spend shit ton of money on more computational power but that method wouldn’t get us far
also if u don’t take it to heart to be efficient then you are setting yourself up for more problems. In some extreme cases you need to invent new methods as you are already running most efficient algorithms with supercomputers but I can’t help you with that cause I don’t work for NSA
) So as a rule of thumb if you don’t have to worry about calculating “in place” and runtime difference of heap sort vs quick sort you probably aren’t dealing with big data
.
Some less general knowledge (or maybe not idk it’s 2020 ppl are getting smarter by second):
It’s not only about how much data you have. You can do some funny AI stuff on small samples of data and have huge strain on resources. For example in machine learning you have the concept of a “vector” which is the set of inputs you use to teach your neural net (or something else, for example bayesian methods). As the vector gets longer and you use more inputs (for example: weight, height, blood type idk some wired name protein levels etc. of diabetics) the amount of calculations needed to be done explodes (the exact big O depends on the method you use but with many machine learning methods it’s n^k where k is the length of vector).
I hope this answer isn’t to trivial for you and gives you some interesting insight also sorry about using some many words I have trouble with getting to the point
This hasn't been active for over half year
@John Mahone
Anyway OP, what do you think hive mind will be incorporated to in the future? Any practical use?
Well idk much about hive minds sorry
But I do know a bit about related concept of swarm robotics and a bit about IoT. So I would say depending on human will... possibly every freakin thing
Fun fact there is already far more “intelligent” devices (I’m talking about WiFi connected fridges etc so that word is a bit of a stretch) in the world than there is humans. And swarm robotics is very dynamically growing field. Personally I consider it to have a lot of potential because it lets you do incredible things with very very simple robots that have laughable resources but are deployed in swarms. We are already capable of emulating social behaviors of insects without problems (we lack mechanically sophisticated robots that can be on par with a fly and be cheap tho, but it’s not really an issue cause in most applications that’s irrelevant. Like I said big capability with dirty cheap components). So yeah if you combine the two fields together (IoT and swarm robotics) it gets freaky. That’s what you meant by hive mind right? Swarms of robots doing shit thanks to mutual interaction? Because if you meant the hive mind as one Unit controlling multiple small devices then this is good old command and control server for you
I’m sure many ppl on this forum are familiar with that architecture when doing botting
