!dev
Unpopular opinion.

Mage or shaman is no different than data scientist or insurance agent.

Even more unpopular opinion: when technical people write anything about stuff like shamans or mages, odds are that it's the Dunning Kruger effect in action.

Why exactly?

@iiii cause of how probability works, we more and more tend to believe that something that have 99% of probability that it won’t happen - it won’t happen and than it fucking happens cause of this important 1% that can change everything at any given time but we don’t care and just try to overcome the failure with statistics that will prove that previous statistics was wrong

@Fast-Nop yeah well you’re over educated with technical stuff so more or less you’re to professional to give sane opinions about stuff.

@vane I really did not understand what was the idea you wanted to tell

@iiii that despite the statistics financial crisis and pandemic happens anyway so whatever you tend to describe with your fancy shitty charts and probability it doesn’t matter in long term.

@vane It's more that material cause and effect is one mechanism, and actually most reliable, but not the only one. The ordinary and the exceptional complement each other.

@Fast-Nop Yeah we don’t know what’s the cause, we only try to predict that if probability states that 99.9999% means that something will happen we also state that 0.00001% will happen sooner or later. Of course we prefer later but well - shit happens and we’re not masters of our life ( unless we figure out quantum alignment - I think ).

My main point is that statistics is good until it’s wrong so it’s no different then saying something as a mage or shaman 200 years later.

The only difference is the knowledge we have and probability of predictions. But the general meaning is the same. It’s only that the average statistican today is best mage 200 years ago.

@vane And that's a pretty deep misunderstanding of what magic and shamanism are even about. These work on the individual case, on the peculiarities, while statistics work on a large number of cases and try to remove all specifics to even get to large numbers.

One is art, the other is science.

@Fast-Nop yeah but at the end we just try to prove our thinking process, convince people that our rights are more trustworthy than others. We don’t make statistics about data we don’t want to process and we tend to make shaman/magic thoughts about things we are convinced that they’re truth.
At the end there is just past and present and when we spoke about magic or provide some statistics it’s more or less no longer truth.

@vane And that's the misunderstanding of equating "magical thinking" with "flawed logic".

Art isn't logical, it isn't reproducible, and it doesn't even have to be. In fact, if it were, it would be bad art. Art lives in the moment and cannot be generalised.

Art is what works this one time, and trying to make it reproducible actually would kill its essence and ruin it.

@Fast-Nop Our thoughts about statistics are no different than animals scared about survival - more food means more probability that we won’t be hungry. More heat, more coal, more money, more shit.

From technical point of view, artificial intelligence and machine learning is just precision and magnitude of data, math is still the same and no different than 100 years ago.

We are just blinded by logic that favors majority / average data but is it really in favor of how we want to live ?

Maybe, just maybe, just because how history proved many times that the minority is right not mainstream movement.
Maybe we’re wrong about how statistics work.

@vane the problem is not with probability and statistic and the way they work. That has been mathematically explained and proven as much as physics and optics. Its with not having a holistic preview of the data or people not being able to either understand them or compute them properly.

Taking an action based on a 99% probability would be correct 99 out of 100 times. Taking the off-chance you will be right or wrong depends on the person, the situation, and the environment.

@iiii I'm a data scientist. I agree with OP in a lot of cases what employers want from a data scientist or analyst is persuasive documents that support their existing philosophy. People in large groups need elaborate myths to reconcile their differences and work together, so leaders are always looking for a shaman class to perform a suitably impressive ritual to help with the persuading

@ArtOfBBQ now that's a comprehensive reason

but guys, hating on a data scientist for performing shamanic rituals is like hating a developer for creating exactly the useless, crappy software that his rich contract and/or manager asked him to make. Hate the contractor, not the contractee playas. Yes we can torture data until it confesses that his highness the CEO's hypotheses are all spot on, but we could also do actually useful things if asked to. Statistics, probability and machine learning are really useful tools that can do cool things, there just happens to be lots of demand for board room rituals and other even more evil stuff