Social media make us stupid. To be precise - they encourage production and emission of a useless verbiage as a from of virtue signaling.

The cultural change is that being “wrong” is ok for some talking heads, and nowadays it is even possible to argue that “there is no wrong”, just an “imperfect information”, you know.

The older cultures were better. They had a common sense notion of “you have no idea what you are talking about”. Yeah, this is ok for a bar room conversation, but something is wrong within the systems when “people of knowledge” produce an utter parrot-the-bird like bullshit.

One once reputable guy on the entertained produced an emitted an article, in which he asked a model questions like:

  • Will I make it to a NYC airport by 2pm on Saturday, the 24th?
  • Will Litecoin (LTC/USD) Close Higher July 22nd Than July 21st?
  • Will SBF make a tweet before Dec 31, 2022 11:59pm ET?

and so on.https://www.lesswrong.com/posts/CkhJAxHeyFCg2EcET/are-language-models-good-at-making-predictions

Notice that all the questions are equally absurd. They rise to a whole new level of absurdity given to what they have been asked.

Lets “unpack it”, as they used to say when they want to look smart and being an authority on a subject.

Predictions and even the notion of predictability are well-understood, unless you are have majored in humanties. Everything that mathematicians came up with is based on the notion of a probability of an event in very particularly constrained settings - simple games of chance.

The common characteristics of coins, dices and cards are that the set of possible outcomes at every state is finite, fully observable and the steps are discrete and well-ordered. In sort, it is a bucket-sorting problem (where buckets are of different sizes, but with a fixed total area).

In a dynamic, systems based on belief, each observed event changes the buckets (keeping the total area fixed).

Within such artificial settings (in the context of a well-chosen game) there are certain valid calculations could be made, which, however, have no direct connection (in principle) with an actual outcome of the next move. Moreover, each outcome (a coin toss, lets stay) is independent from the previous ones.

This is the limit where pure mathematical rationality could lead us. Beyond that everything else is a plain old bullshit.

A person who enjoys a high social status of being a “person of knowledge” or an “academic” should, of course, knew all this. Otherwise, I suppose, the well-known social mechanisms should move him form an expensive apartment on some campus into a cardboard under a bridge. At least this happens to an incompetence in the other competitive areas. But somehow it is ok with abstract bullshitters, exactly as it were within organized religions or sects. “There is no wrong” when it conforms.

Anyway, let’s continue the unpacking. There is another infallible principle about predictability. It comes from the nature of reality itself, but has been observed and validated even in artificial settings. If a single major factor is missed (not taken into consideration) the whole prediction is a guaranteed bullshit.

Another one, closely related is If some new factors are emerging (evolving) while we are trying to “predict”, the prediction will always be invalid.

In some settings, like weather, the new factors are rare, but when they actually emerge, all the predictions are fucked. Snow in Texas, the Indus river flooding etc.

People who understand the notion of being unpredictable (not just an opinion, but as a mathematical certainty) are the trading guys. The finance industry spent billions in order to “predict” the market events, only to be crushed against these infallible universal principles.

All your models are wrong and outdated the moment they have finished, not due to merely new data, but to the changes in the underlying causality.

A system is unpredictable (by definition) iff a new factors appear (emerge), old factors disappear and the “wights” and ratios undergo significant changes.

Another name for this is “an environment is not stable”, “partially-observable” and the information is always “incomplete”.

Now pay attention. Applying any statistical methods to such environment is a modern alchemy and astrology. Practitioners, therefore, shall be treated accordingly. Yes, yes, I know.

There is another fundamental principle though. A mere information can never be the source of a valid inferences and prediction. It will always be wrong as an uneducated, clueless person without any actual measurements.

The reasons and examples are trivial and underlay any sectarian or religious texts. It is as a reaction to this socially constructed and maintained bullshit the early Buddhism and then the Scientific Method has been evolved.

It is the Scientific Method (alone) which invalidates a mountains written verbiage.and what remains (cannot be discarded or falsified) is the basis of inferences or predictions, provided a system is stable (which, aside from simple games, is not).

Now the fun part? It seems that I should have a higher social status as a “person of knowledge” than these alchemists and bullshitters, shouldn’t I?