Fighting algorithms with algorithms
Calling social media’s bluff…
With Miroslaw Manicki
E.U. Takes Aim at Social Media's Harms With Landmark New Law
The Digital Services Act would force Meta, Google and others to combat misinformation and restrict certain online ads…
The European Union is leading out in the regulation of social media companies. This is a timely and important development. The question is, how can this be done? The urgency of this is that much higher with the imminent privatization of Twitter by Elon Musk, all of a $44 billion transaction.
Hint on how to turn the tide on social media giants: They have a weakness: They didn’t finish school. This isn’t to say that they couldn’t get an education — most of them were elite kids at Harvard, Stanford, etc., who elected to not finish their programs — or even stay in them very long before dropping out. You have likely heard the stories. It seems these were the rule and not the exceptions.
Their rationale for dropping out? Why stay in school? They were making money. Why else would you stay? What else is school good for? Indeed, why would you stay in school and learn things, and complete some kind of program of study?
Here is a point: Perhaps, there might be something to learn there. If not how to program a computer, perhaps through school-oriented activity one could learn aspects of nature — how it functions, how its intricacies support our every breath and give us a firm foundation to live our lives.
If not this, perhaps there could be something to learn as to ‘how the world wags’ — the social side of things. How do millions if not billions of us can live in relative harmony most of the time? How does this happen? How can we keep this happening?
Could these be rationales for staying in school — especially if and when you have achieved level of power and influence? Or, is this simply the ramblings of ‘chumps’, people that are too stupid to understand that stupidity — or, at least, purposive ignorance — reigns?
Let’s find out answers to such questions by fighting fire with fire — or at least putting algorithms up against other algorithms — theirs against some new ones. As can be seen below, this could be a matter of comparing typical ‘if-then-else’ logic against more accessible and more comprehensible tree-based logical arrangements. This means what we can see below to the left in competition with what we can see on the right.
The ease of interpretation of the algorithm on the right — to do with health — compared to the arcane computer code on the left is the very point. Ironically, the easier interface provides a design environment much more suited to experts, who have less time to battle with computer code than programmers do and more interest in perfecting out their knowledge in environments that provide valuable insights and that can help to deepen their knowledge, particularly the knowledge of processes in their study areas.
One concept emerging from conventional social media banter is that there isn’t much knowledge worth worrying about. From this perspective, it is viewed that everything is trumped-up; nothing is intrinsicly true and valid. Accordingly, pretty much everyone just makes stuff up. This notion is reinforced by the billionaire dropouts: There really isn’t much to be known, much to learn in colleges. In their minds, “they” leapfrogged past the whole problem by making big money up front.
Oh, the technologists pledge allegiance to the kind of knowledge that put them in their positions of power: Semiconductors, nanotechnology, sensors, fiber optics, satellites, atomic clocks, display technologies, and on and on — those count as science, science that is used to undermine science in general. There is belief in Moore’s Law that supports improvements in smaller and faster devices, but little credence given to cooperation, collaboration, sharing, and egalatarian attitudes, which allowed for people to live in closer proximity as urbanization came to be.
They thus live in a simplistic world in which they were able to pull off a nice, little trick. As digitation was in the works anyway before social media took off — and the old modes were not particularly popular publicly — they were able to leverage the work that had gone before while ignoring lessons learned by their predecessors and by societal leaders since time immemorial.
Take religion, for example. Religions do represent things that people believe in that are problematic and implausible to others. More importantly, they represent lifestyles and habits that serve the public good in many ways. They establish and promote regular attendance in meetings, study, and prayer groups where understanding, compassion, humility, cooperation, and thoughtfulness are discussed and encouraged.
These are good things; they serve as the fabric of society, as it is said. Churches also encourage regular, time-oriented activities that improve mindfulness and health. These have a moderating effect; they enhance prospects for peace. They provide superstructures for meaningful, fulfilling lives. They assist in putting the ‘civil’ into civilization.
A lot of the ‘hotshots’ have little patience for nor interest in such things. They think they know better. Their financial winnings reinforce this. All of the social stuff — especially to the extent that it is ‘religious’ or ‘responsible’ — is for chumps. Aspects of science that they find inconvenient are similarly suspect. This feeds into the vast reservoir of wilfull ignorance that has glomed onto the prevalent social media sites.
Hints as to social media vulnerabilities
Here are a few things that the dropout social media barons do not know. These point to opportunities for more responsible, more fulfilling use of the technologies that virtually everyone now uses:
- Some aspects of what social media companies are doing will destroy society, which will then destroy what they are doing.
This was noted by Joseph Schumpeter, who pointed to the 1947 Railroad and Banking Disaster in England as evidence that extreme financial capitalism was destructive to society and simply wouldn’t work in the long term. Modern social media barons all know about Clayton Cristianson’s ‘disruptive capitalism’ stuff. They thus cheer on Google and Facebook’s conquest of the newspapers and Amazon’s demolishment of any other kind of retail. This isn’t that. Schumpeter documents hard core ‘destructive capitalism’, which is really ‘big boy’ stuff.
This isn’t unique to social media companies, but they turbocharged the process. Extreme financial capitalism refers to what happens when commerce is taken over by money people. Commerce is this great thing that can happen when people trade in goods and services and broaden their capacity to both give and receive. Extreme finance turns these things upside down.
We have the marvelous history of entrepreneurs that apply their talents to provide something unique to others. In some cases, this has to do with new products and services that had never existed; in other cases, it involves similar products provided with better quality, higher volumes, faster delivery, more value for the money in a variety of ways, often both scientific and creative. There is a virtuous cycle in this. With more business, costs can be cut, prices can be brought down, and all people benefit from high levels of productivity, better sourcing, and production improvements.
Extreme finance is different than this. It does not involve adding, but subtracting. The point here is, as with the fruit in the display below, profit from transactions needs to be continually increased using a variety of schemes.
First and foremost, the price can be hiked. That failing, costs can be decreased. How is this to be done? Product size can be brought down, perhaps with creative, ‘stylish’ packaging. Manpower can be decreased. This is a frightful aspect of acquisitions, which are a favorite vehicle for extreme finance to generate increasingly higher returns.
Inputs can be cheapened. Corners can be cut in a variety of ways.
If you are in the middle of such conditions, you will hear such things as that there are new programs that turn out to be simply shutting down of old programs. You hear of jobs being eliminated, part of which end up as new responsibilities for existing employees, including you if you happen to be among the fortunate ones whose jobs are not eliminated out from under you.
As Schumpeter noted, this is not sustainable. Eventually, everything just goes ‘poof’. It results in anomalies that are only too common in contemporary daily life. Where there used to be process, there is fractionalized momentum of sorts where holes were blown in processes and procedures of the past. Social media has caused a good deal of this, leaving fragmentary outcomes in its place.
2. The main lie is that there is a process when there is not.
Prevarification on this score is common. This is the essence of ‘fake news’, at least news that is inarguably fake. Real news is the result of rigorous journalistic activity, which involves multiple steps of information gathering, clarification, and validation. These are processes. Well-crafted processes, by their nature, are valuable, enlightening, and often motivating.
These are factors that render processes important. Understanding of and attention to a myriad of alternatives are needed to be sure that a question or a response to it isn’t just made up. The process requires skill, commitment, sacrifice, often risk, and specific kinds of specialized knowhow.
Such information gathering and transformation can be done in such a way as to be quick, efficient, credible, and satisfying. They can be designed to incorporate data from many sources — private to the person or persons in question, likely picking up context from prior activities and new data streams.
One piece of good news is that data need not be disclosed to the user to be useful and it may in fact serve to drive a process forward without even slowing down. Similarly, such data needn’t be transferred to or disclosed by the process in question. It can simply be ‘check’ for qualitative or quantitative characteristics and left where it is.
Typical page ranking, such as in the Google search engine, doesn’t work for this, as many pages, though popular, are sheer nonsense. Every time an open query is ordered up with such tools, contexts are tossed to the wind. Just as a great piece of music will lead to a crafted solution without a very loud air horn sound in the middle, online sessions need to be protected from outliers and interloapers.
Perhaps a process is well-served to look at ouside data — such as the availability of a product or another conditions as available on open networks, but in no case should that act override the logical process in process. Once context is lost, all bets are off.
3. Google page ranking and similar ‘popularity’ measures have played their course
Page ranking for Google was a brilliant product addition in its time — in the late 1990s. It was a neat trick; it won them the business in basically a week; it is now a problem. It doesn’t measure the efficacy of the page, just that it was popular. This lies at the basis of their algorithm problem — Duh!
So if you have a hundred people that say one thing, or repeat a thing said by someone that simply made it up, that counts more than statements of one person who has dedicated a lifetime of preparation and study, with different results and possibly a conclusive message. People have learned that under a page ranking algorithm, they can just shout down the sole expert — or even a bevy of experts — and that is the end of that.
Allan Bloom wrote in 1887 in the “Closing of the American Mind” that it was a travesty that the wonders of technology, the sacrifices and miracles leading to tremendous breakthroughs in computation and communications, would lead to a specious end: A sole teenage boy, alone in an apartment tower, playing endless videogames with little interruption and no engagement with others.
This is that only much worse. So many people have willingly embraced the void; they seldom if ever look up. Clearly for the forseeable future, they are going to go wherever the blinking cursors take them.
4. Artificial intelligence is an extension of the arrogance of the dropouts and their ignorance
Ardent technologists hold artificial intelligence out as their ‘ace in the hole’. ‘Soon’ people will not even need to think, as the brilliance of the technologists and their creations will result in cyber cognition that will render our own efforts at thought and decisionmakeing extinct if not irrelevant. The technologists themselves will thus be able to skip all of the awkward, low-pay ‘education stuff’ for real and forever. They will be able to skip to the head of the class by skipping class.
I have experience with artificial intelligence people dating back to a DARPA project from 1999 titled “Rapid Knowledge Formation.” It was an adventure in wonderment. Their quest was to extract what they called axioms, what you might call ‘factoids’, from experts so that they could embed them into their algorithms for computers to use to engage in what they called reasoning.
Such factoids would include things like ‘mothers are older than their daughters’ and ‘rain falls downward’ — things that are intrinsicly true, for the most part independent of context. The idea was once they had collected all of the axioms in the world, the computers would be able to blithely string them together in a series of queries, not unlike the decisions you witness when you do a Google search or type into your smart phone using error correction. The result would be magnificent decisions.
The problem they were facing was that experts weren’t cooperating. When they did, they were way too slow. The AI people insisted that they needed to collect axioms ten times as fast from the experts as had been the case to that point. In the meantime, they were trying to scan documents and do other things to pull their axioms from any and all sources.
To be clear, they didn’t really want to understand the thought processes of the experts, they wanted “Just the facts, Ma’am (or sir),” as television’s Jack Webb used to say.
Experts are like homecoming queens that have a tendency to know they are pretty. Experts know they are smart. As it turned out in the DARPA program, they were not very good targets for the technologists for axiom harvesting, as they would soon ditch the projects for greener pastures. Once again — they were experts. Such people have particularly low tolerances for nonsense and they are typically in high demand.
I presented an alternative to the DARPA artificial intelligence people. They could give experts the tools to design processes, including not only axioms, but logical frameworks to link them together. These would be processes.
This created a bit of a stir, because the AI people had a problem with the term process. I pointed out that since the algorithms were to be designed by experts, they could rightly be called expert processes, leading to expert systems, leading to artificial intelligence. They just looked at me in a funny, kind of squinting way.
My plan to them was, having the processes laid out by the experts — coupled with the context provided within the structures of their work, the AI people could pull the axioms out and piece them together as they would like. These could be pieced together in a tree format for the computers, what is called an ‘inference engine’.
They didn’t go for this, as it seemed to give too much capacity to the experts, who might thus just go out on their own and create more things. That wasn’t the plan at all.
They are very smart people, so there was an element of risk, as I was aware. So, as these things go, they tried to trick me. I was concerned that they might nail me with a potent zinger of some kind. In one of the sessions, after I presented my idea to a sea of demonstrably uninterested people, almost all men, came THE QUESTION. It was from a young man that I had gotten to know the night before — friendly to that point.
In the session with them all there, he said (I paraphrase), “How can you say that everything can be expressed using a tree? How would you describe a matrix using a tree — it just doesn’t fit that format.”
This was followed by a little bit of silence. I suppose this was a ‘mike drop’ event to put me in my place.
I responded, “Using a tree structure, you describe the rows of the matrix in a branch, you describe the matrix columns in another branch, and you lay out any formulas that determine the contents either in subsequent branches or within them. Next?”
There was no “next”. The point is, there is no way that they are going to open the door to legitimate experts. They are afraid of them. From that point, I was persona non grata to that crowd.
Hence the artificial intelligence ‘Hail Mary’ approach to gaining and retaining control of the Internet’s “switches”. They keep prophesying their ‘singularity’ up the road somewhere, when humans are eliminated from the equation (and the equations, for that matter) because they cannot keep up with the machines.
Meanwhile, their artificial solutions fail to deliver.
As an aside, when legitimate processes are available, designed by experts and authorities and readily available to be used, all of the benefits of cyber systems that serve questionable if not pernicious ends would work for the good. The weakness that humans bring to the table isn’t cognitive capacity, it is the ability to pull everything together to make decisions in a timely manner. This is where the machines can rise to the occasion. Quickly. Reliably.
5. There is no solution beyond rigourous management of context, which can only be done using trees
There is that sinking feeling in an online session that what you are learning makes sense, but for someone else. Take weight loss, for example. You read something online or you see a video and someone makes an impressive presentation on what you should or shouldn’t eat and how you should or shouldn’t eat it. Does it apply to you? Maybe, but the message likely came out in ‘blunderbuss’ fashion to a whole lot of people. There may have been some key words in the process, but these are highly suspect, having likely been entered as open text — very unlikely to have been data-driven or process-driven.
As a user, you are then left to guess as to applicability. You might try whatever it is, you might not. If you do, you may never know why it did or didn’t work. You don’t have enough information or you do not know how to piece it together. You likely have tried several things before. Will you buy a product? Will you sign up for a program? Do you have faith at all in likely outcomes?
This is the state of chaos brought on by loosely-coupled social media. The delivery mechanism is there, but it opens up into a grab-bag, if not a Pandora’s box. Frustration mounts. How do people respond? Some quit the platforms. Others keep trying, maintaining stoic resolve that eventually their quests for answers and solutions will be satisfied. In some cases, they might, but then followed by more trails in the dark, as the context problem will not go away without some kind of systemic solution to context matched with credible, applicable processes.
So, the way to beat them is to join them in the algorithm game. This isn’t to copy their simplistic page ranking and their rather dubious chains of AI queries. We need to use the technology to document what we know in the aggregate, organized in a way that sorts out the data and the issues so that we can get dependable answers when we need them.