Message:

All countries should block access to Facebook, Google, YouTube, Twitter, and AWS | Random thoughts | Who I am | Yaniv Hamo

All countries should block access to Facebook, Google, YouTube, Twitter, and AWS

Yes, I know, sounds like censorship, sounds like China… That cannot possibly be right, right? Wrong. Let us walk through why blocking the tech giants has become not only a natural extension of capitalism, but also necessary for protecting democracy.

Even though the tech giants were pushing a political agenda for at least a decade, it had only become clear for all to see in the aftermath of the 2020 US election. Twitter permanently banned Trump not for what his tweets actually said, but because of the “context of broader events in the country”. This is the language of Derrida’s deconstruction technique, post-modernism 101. Facebook followed suit, YouTube banned Trump, Google and AWS took down Parler, we were all witnessing.

The tech giants are busy for years pushing post-modernist agenda in much subtler, though highly effective ways. These platforms (publishers?) are collecting and organizing information, presenting it to their users in a seemingly objective manner. The magic happens in how they order the information displayed to the user. Information is digested by humans in sequence, we are bad at grasping multiple pieces of content all at once. Of course, what comes at the top of the sequence has higher probability of being read and internalized by us. Who gets to decide what comes at the top, be it Google’s search result page or your Facebook feed?

The tech giants will gladly explain to you that this is done by algorithms. Cold, rational, and completely objective algorithms sift through the data and find what is mathematically the best match to what you are after, ranked in order of relevance. If Google is showing a very unflattering (to the point of mocking) photo of Trump in the months following his 2016 election, it is only because that is what the internet thinks of him, Google is merely surfacing what is there, or so they would like you to believe.

Indeed, if you press them in a Congressional hearing and ask to see the algorithms – you’ll find that the algorithms are completely unbiased. There is nothing in the code that injects any bias into the results. The fact that Google top execs were crying on camera the night Trump was elected is completely unrelated, they are only humans, but the algorithms are not biased against him. You might say now “wait, I am not stupid, the algorithms are objective but their training data is biased!” – that is getting closer to the truth, but it is still imprecise enough to be proven wrong. Training data is obtained by sampling real-world traffic and user engagement signals, and that sampling is unbiased, it is dry statistics.

To understand how political bias is actually introduced, take a hypothetical case of a law suit coming Facebook’s way, because some inappropriate result found its way to the top of the feed. This happens all the time. Facebook cannot maintain that there is nothing it can do about it, because this is what its objective algorithms have decided – they will be grilled at court if this is a copyright issue, or child porn, or national security threat, and will be forced to take it down – and quick. They cannot re-train the algorithm now, such that it would down rank that bad result specifically, while keeping the quality of all other results: this would be both complex, and time consuming. Therefor Facebook, and all tech giants, must introduce a mechanism to block results from showing, subject to manual review; no longer an algorithm. These filters are where a bias is introduced, hidden from sight, without any traces in the code itself.

When it is a legal matter, the tech giants will naturally revert to this manual filter mechanism. But now that the mechanism is implemented, it is being used also for non-legal matters, and is very effective for pushing a political agenda. If Google showed a mocking photo of Obama at the top of the search result page, some engineer would file a ticket to complain about this, and the offending photo would be added to the filter in a matter of minutes. Another engineer filing a ticket to filter out Trump’s photo, will this time face internal objection which is very valid: “sorry, we are not interfering with the algorithm unless it is a legal matter”. The engineer would escalate, point to PR issues, and will eventually succeed, after several months, to filter out this image. But the entire world has seen a mocking photo of Trump for months, with all the subconscious implications it brings along. In addition, the mocking photo has collected a lot of clicks and engagement from users thanks to its prominent position, so an objective algorithm might decide that it is actually matching the query very well, and organically propagate it up in the future.

Here is the real beauty with these filters: they get to inject a bias into the algorithms without leaving traces, keeping the rest of the system completely objective, and even keeping the filter list itself objective. Imagine I filter out all images of oranges for the query [Orange]. Some other results will propagate up, there are always endless list of matches. Could be other fruits, could be anything. People will click these results, because people click on whatever is presented to them, let’s say they all click on images of mandarins. Next time the algorithm is trained, it will associate the query [Orange] with these clicks – it would learn that images of mandarins are what users engage with the most for the query [Orange]. After a while, you can take down the filter, but the result will stay the same: photos of mandarin for the query [Orange], when the entire system – training data, algorithm, and filter – are all unbiased and completely objective.

But these subtle ways are sweet memories from the past. Today, the tech giants just openly block whatever interferes with their political, post-modernist agenda.

By allowing their population to use services by these companies, countries are opening themselves up for political interference and cultural values that might not be compatible with the values of that country, or with its own best long-term interests. By allowing these services to run, countries expose their citizens to a deliberate brain wash by a foreign entity, which is pushing for a specific agenda. People spend vast amounts of their time in front of these services, and increasingly confuse the values and political views they are exposed to (well crafted and one-sided) to be representing of what is true and real. This will inevitably translate into effecting the results of democratic elections. It is bad enough that all democracies are vulnerable to internal propaganda, but this has always been the case with modern democracies. A foreign unidirectional and omnipresent propaganda is not an improvement.

To remain sovereign and democratic, countries have to protect their citizens from highly motivated foreign powers. China is at the forefront of blocking foreign political services in its territory. Most other countries are lagging behind, most notably, Russia. Russia was always vulnerable to western interference, with Lenin being the strongest attack on it by the west. But today you don’t need a Lenin to take down a country. Google, YouTube, Facebook, Twitter, and AWS are equally as powerful. How can Russia, France, Germany, Great Britain block the tech giants while still respecting the freedom of choice for their citizens? I do not have the answer. I only know that they will have to do it, or cease to exist as democracies.

Another reason for blocking the tech giants is good old capitalism. For those who haven’t noticed, the online technology world has shifted long ago from selling services and products, to selling the attention cycles of users. The big tech learned how to convert attention into money, and became extremely effective at that. Google search, Gmail, maps, YouTube, Facebook, Twitter, are all free thanks to the big tech companies selling your attention to the highest bidder. People are largely OK with that, because they save money they would otherwise have to pay for these services, and are spending more and more time in front of these services, enriching the big tech companies.

Attention cycles became the equivalent of a natural resource, they are a hard currency with a known exchange rate. Currently, almost all attention cycles of the entire planet are going to the US big tech companies, and money flows with them. Why would France, just as an example, agree to give up on this natural resource for free? Had they blocked all the US big tech companies, French companies would step up to fill the gap, providing equivalent (maybe even better, who knows) services for the French population. The attention of French users would be monetized inside of France, instead of freely leaking into a handful of US companies. Search engines and social networks are commodities these days, it is a lot easier than it seems to create an equivalent, if not superior, search engine to Google, or a better social network than Facebook, especially you tailor them to a specific population.

The one country that dared to make the step of blocking US big tech is China, and it shows in the valuations of their tech companies. Baidu, JD, Tencent, Alibaba, WeChat, Bilibili, are just some of the names, and their combined market cap is not too far off their US counterparts. No other country has such a thriving eco system of local innovation as China, all thanks to them blocking US services en masse. The combined market cap of Russian alternatives is ridiculously small, only because Russia still insists on putting personal freedom above national interest. I am not even talking about the almost non-existing French or German alternatives.

The benefits from blocking the US tech giants are great, and are increasing: protecting the rule of democracy, encouraging domestic innovation, securing financial resources, maintaining national security, keeping a healthy competition. The downsides of blocking the US tech giants are hard to see; there is nothing special or complicated in what they do, they bring no value over ever-improving domestic clones. I believe that these incentives will result in more countries following China’s footsteps. Time will tell.

back