TL; DR google is not evil, blocking access to relevant information is just a side effect of its business model:
google silent autodafé rule is : any knowledge that is not generating a revenue stream does not deserve to be shown to you
Years ago, after a «burn out», because I am library rat I used my usual paper based routine to dig information on the topic.
Before I go for «new text», I always go in the past to see if there was no precedent to a bigger extent of a phenomenon we know. The rationale being if a phenomenon was extreme at this time hence it would have been easier to study. (How to do science with a Hammer (function)).
My topic was why am I feeling like shit after 60 hours work per week?
So instead of looking for burn-out in google I went for what were the reason to limit working weekly hours to 48hours.
Through an interesting journey on google I found something that I love! Figures and charts with close to unethical experiments on human beings.
It is called the scandal of the ammunition UK 1916. Long story short: WWI saw the prevalence of death per explosive during conflicts, hence winning a war was a question of optimizing the chain of production to maximize the throughput.
Until 1900 (or 2000?) there was a belief productivity was linear to the number of hours worked, and after a lot of patriotic propaganda, workers were pumped up to work willingly extra hours for the same money, and then all hell broke loose: productivity plunged, but more importantly quality; UK explosives were killing tommies because of defects.
So, the UK army did extensive unethical studies justified by war time on humans to scientifically establish the parameters influencing production and generating this mess.
Long story short:
- peak of productivity is 30 something hours;
- peak of productivity when accepting non lethal defects in production is 48hours;
- night works harm;
- the health cost due to burn out far outweighs the benefit in produced units after 48h weekly hours.
48hrs is not the optimum, it is the optimum degraded mode acceptable during wartime when it is okay to push workers above their limit to get a 7% increase in production in exchange for reduced quality.
Just so you know this was a thing USA adopted the 48hrs week in 1917 because they saw obvious business benefits. And O.I.T. later made it an international obligation that is on par in our states with Geneva convention on war law, and IP convention.
A lot of 1900's books reachable with google used to evocate fatigue professionelle (in texto) and the cost of damages and responsibility that lied on the employers.
Just for the record 1800 to 1900 have been quite intense in terms of work conflicts.
And google would give to the astute persons a way to uncover traces of this history. It would be in google books notably.
What pisses me off is that years after years this kind of knowledge disappears from all the search engine.
But where is it now on the internet?
It is gone, and all because of an algorithm.
What matters to me is not what matters to google. Google first reason to be is profit. They make stream of revenue on ads that are either on the contents you look for, or on the keywords you enter.
So their algorithm to maximize profits tends to favour when searching topics topic who have a paywall (academic, software or normative (like IEEE)).
Pointing at anything in the public domain is a clear cost sink for them.
If you try to visualize their algorithm, you will notice it favours fresh news other older one. And it is a good idea : I don't think a physical teaching of 1760 he is on par in terms of relevance with a 2017 teaching.
Basically, they modulate their relevance with a temporal curve. Probably and amortized exponential, that has a strong cut around anything that is not protected by patrimonial right.
It is probably not intended but a side effect of an algorithm optimizing the revenues.
No need to be a genius to guess the google engineers made the sin of any mechanical engineers : they put a positive feedback loop in their algorithm.
If you are good at "machine learning", why not teach your search engine to return the more relevant search results that maximizes your incomes? Resulting in the fact that any free content that doesn't generate revenue thus become irrelevant. They still are on the internet: but google will not give you the link to reach them.
Yes of course NEW materials tends to be more relevant than old one. But, it is far from being always the case. Like in the case of old established knowledge like: vaccination, Fourier transform, thermodynamic ....
You know in 1804 Fourier, one of the father of thermodynamics clearly stated human activity is responsible for global warming, and because of a small mistake in calculation, he concluded it was a good thing since according to him we were condemned to die in cold otherwise. The demonstration is still totally relevant. But it is not Fourier simple and clear demonstration that bubbles up in google : it is either the anti-climate BS or GIEC smoke-screen that comes up. The 1804 thesis of Fourier on global warming SHOULD BE in the top 10 choices when asking about global warming scientific basis.
I can also point you to "out of internet" knowledge that proves Fourier's calculation are not relevant anymore: trees and plants have moved up to 200kms/decade to colder places. But it comes from out of the web knowledge, people counting plants and insects around the world in (paper) notebooks.
It is funny that the anti-ageism and pro-tech bias of google seems to be at least a reality in its search algorithm.
Google silent autodafé rule is : any knowledge that is not generating a revenue stream does not deserve shown to you, hence any knowledge that is more than 70 years old.
Well, what we are doing is stupid anyway.
Imagine you have a public library, but do not want to have librarian because they cost to much.
So you give a company this mission saying just finance yourself.
The librarian goes to the bookshop and says I have business model idea: for every 10 search I submit to a reader I put one to a book you sell in exchange for a share of generated income.
This way you fund new knowledge while preserving the old one.
Readers also benefit from having updated knowledges.
The problem is that in medio stat virtue. Too much of any is stupid.
Only old news would leave us in a crusted world where no progress would be possible.
Only new contents condemn us to reinvent the wheel poorly by forgetting the basic knowledge.
Google does not care, they are not a librarian, they are a business. google is not a search engine, it is an ad based streaming revenue generator empowering any companies able to make people pay to access to data.
My conclusion is we all used so much google we forgot about Shannon definition of information: H = k ln (O) (where O is the ratio of relevant choices over total choices).
If google was useful to gather information we would have a rarefaction of choices to organically point to fewer and fewer relevant choices.
If google algorithm was information based it would tend towards losing revenues, and having its users lose revenues. At the opposite, google is in a position to create artificial revenues to make you pay for information that are publicly available out of the proposed choices. Google has a direct incentive to block you from publicly relevant accessible content so that the ads buyers can generate revenues. It does effectively hide information.
Google is NOT evil. Google is just like a rational good gear in a malfunctioning clock.
Sometimes there is no one evil but the rules of the game we play. Necessity may be evil, but there is still no necessity to leave under the realm of evil.
My conclusion is actual societies are immature thus sucking big balls at understanding the simple challenge raised by the con behind the Information Technology sector.
Information Technology is not about giving access to information, it is totally the opposite due to the incentive of the market.
No comments:
Post a Comment