Will Techno-Optimism Save the World?! Part I: Technology

Robert B. Kozma, PhD

This is the first of a four-part series of articles in response to Marc Andreessen’s Techno-Optimist Manifesto.? The series is: Part I. Technology, II. The Economy, III. Government, IV. Philosophy.

On October 16, 2023, Marc Andreessen published The Techno-Optimist Manifesto.? You may know Andreessen as the developer of the first widely used web browser, #Netscape.? But more likely, if you know him at all, you know him as the principal partner of #Andreessen-Horowitz, a Silicon Valley venture capital investment company with nearly $36 billion in assets under management.? Silicon Valley is, of course, the capitol of the technological world and with that much capital to invest, Andreessen’s words must be considered if you are involved in (or impacted by) technology, especially if those words are formulated as a “manifesto”.? It is likely that anyone who needs capital to develop their technology-based idea will shape their thoughts and projects around his words—indeed, that’s his intent.? And if everyone in that position shapes their ideas and products around Andreessen’s words, it shapes directly or indirectly much of the technology that affects all of us and our world.? This makes the Manifesto particularly important document.

However, his Manifesto goes beyond the mere promotion of technology, which financially benefits him and the companies he invests in, to articulate a more-or-less coherent philosophy of technology that positions it in the complex political-economic system that we all occupy in the modern world.? I believe it is best to understand the complex, systemic interconnections of technology, economics, government and philosophy in the Manifesto by examining and critiquing each of these facets individually, in turn, with the others lurking in the background.? I begin my analysis with the central element of his Manifesto: technology.

Technology

?Andreessen defines technology broadly, as do I.? It is not just the consumer products that we have come to associate with the modern technological revolution, such as computers, smart phones, VR headsets, and digital wearables.? In his view, technology includes the machines of the Industrial Revolution, the engineered grain seeds of the Green Revolution, vaccines, electric lights, indoor heating, and air conditioning.?

The question that Andreessen addresses with the Manifesto, and one we need to analyze carefully, is will technology save the world?? Much of the Manifesto is dedicated to answering that question in the affirmative.? Indeed, the world saved by technology is not a question for Andreessen, it’s a matter of faith.? But he goes on to claim that techno-optimism is necessary for that to happen.? In other words, he believes that technology is an inherent force for good and an unbridled deference to its development is necessary in order for it to save the world.

This article analyzes and critiques the arguments behind those claims.

Andreessen believes, for example, that technology is “inherently philanthropic.”? That it “is the spearhead of progress and the realization of our potential.”? That a “technologically advanced society has unlimited clean energy for everyone.”? With technology “we can advance to a far superior way of living and of being.”? And there is “no problem that can’t be solved by more technology.”? There is an element of magical thinking in the Manifesto when Andreessen says that, “We believe Artificial Intelligence is our alchemy, our Philosopher’s Stone.”

We’ve heard similar claims before by techno-optimists.? In 2015, Sheryl Sandberg, former COO of Meta, stated, “Social media has created a historical shift from the historically powerful to the historically powerless.”? Ten years ago, Bill Gates claimed that, “Technology is unlocking the innate compassion we have for our fellow human beings.”? In 2010, Carly Fiorina, former CEO of Hewlett Packard, believes “that technology is the great leveler.”? And a hundred years ago, Thomas Edison said, “I believe that the motion picture is destined to revolutionize our educational system.”

That these earlier techno-optimistic claims weren’t realized gives us reason to be skeptical.? But Andreessen doubles down.? He not only asserts his grandiose claims but rails again the “lies” and “enemies” who propose caution.??

Among the “lies” Andreessen identifies are that “technology takes our jobs, reduces our wages, increases inequality, threatens our health, ruins the environment, degrades our society, corrupts our children, impairs our humanity, threatens our future”.? And among his “enemies” are those “indulging in abstract theories … disconnected from the real world, delusional, unelected, and unaccountable – playing God with everyone else’s lives, with total insulation from the consequences.”? Enemies are also bad ideas, such as “existential risk”, “sustainability”, “ESG”, “Sustainable Development Goals”, “social responsibility”, “stakeholder capitalism”, “trust and safety”, “tech ethics”, “risk management”, “de-growth”, “the limits of growth” and the “Precautionary Principle”, which Andreessen claims is not just bad but “deeply immoral”.

Given limited time, it is difficult to address such a long list of lies and enemies individually.? So here, I test out his argument by picking one technology and a couple of the lies and enemies with the confidence that if his claims don’t hold in at least one case, we might want to be a bit skeptical about all of them.

Is AI inherently good?

Let us begin with an analysis of the technology that, of late, is getting the most press (and money)—artificial intelligence—a technology in which Andreessen-Horowitz is heavily invested.? Andreessen is passionate about AI and in a related article he states that it will, indeed, save the world.? Among his futuristic claims are that every child will have an AI tutor; every person, scientist, CEO and governmental official will have an AI assistant/coach/mentor; productivity growth will accelerate dramatically; the creative arts will enter a golden age; and warfare will improve.? He concludes the article in moral tones: “The development and proliferation of AI – far from a risk that we should fear – is a moral obligation.”? And in the Manifesto, Andreessen claims that “any deceleration of AI will cost lives. Deaths that were preventable by the AI that was prevented from existing is a form of murder.”

Let’s contrast these “optimistic” claims with some of the cautionary concerns expressed by others, claims Andreessen considers “lies”.? For example, that AI may take jobs, reduce wages, and increases inequality.? Andreessen’s response to this “lie” is that it won’t happen because it hasn’t happened with the introduction of any other major technology throughout history, despite the prediction that it would.? He argues that there is no reason that the introduction of AI technology would be any different.? Indeed, he goes on to speculate that the massive productivity growth spurred by the widespread adoption of AI may even cause “the most dramatic and sustained economic boom of all time, with correspondingly record job and wage growth.”?

I will take on Andreessen’s economic claims in Part II of this series.? But for now, let’s look more narrowly at the connection between the capabilities of AI and the component tasks of jobs.

A 2020 article by Stanford economist, Michael Webb, presents and validates a method for analyzing the impact of technology on jobs and wages.? Whereas Andreessen is, of course, referring to the overall number of jobs, Webb is concerned with certain classes of jobs.? He starts by listing the specific capabilities of a technology, as described in related patents, and maps them onto the skill requirements of different occupations, as list in the Department of Labor database.? He uses the mappings to predict which occupations will be most impacted by widespread us of that technology.

He validates the predictive power of the method by looking at how it explains employment displacement patterns of earlier technologies, specifically the introduction of robots and of software. ?Contrary to Andreessen’s claim, Webb found a reduction in both employment and pay for occupations that relate to motion and mechanical operation, such as forklift drivers and crane operators, compared to other occupations in the same industries, and the method explained this finding by mapping the mechanical capabilities of industrial robotics onto the tasks of the occupations most affected.? And for software, the method explained how the analytic and computational capabilities of various software productivity systems have reduced employment and pay for occupations that involve the processing of information according to predefined rules, such as parking lot attendants and plant operators.

Webb goes on to predict the impact of AI on occupations by looking at the capabilities of AI, as described in registered patent applications—mind you, the capabilities prior to the release of generative AI systems in 2022.? The method predicts that the occupations most threatened by AI are those that involve detecting patterns and making judgments, such as optometrists and chemical engineers.? Among those not impacted in Webb’s analysis are occupations that involve reasoning about situations not seen before or that involve interpersonal skills, such as teachers and managers.

?Based on these findings, Andreessen’s argument that job loss is not going to happen because it hasn’t happened before is flawed.? Yes, over all, the total number of jobs have increased in the past but with the introduction of new technologies.? But specific job categories were reduced or eliminated.? On a personal level, the fact that there are more jobs created is of little consolation when jobs with your skill set have been eliminated and new jobs require skills you don’t have.? If we are concerned about AI improving the human condition, as Andreessen claims to be, this personal level of disruption and harm needs to be considered as we move forward with AI.?

Another pattern that Webb found makes us think that the introduction of AI may indeed be different this time when it comes to over all job impact.? Webb found a creep of job elimination up the skill ladder with the introduction of robots, software, and AI moving from the displacement of manual jobs by robots, routine cognitive jobs by software and more-advanced cognitive jobs by pre-2022 AI.? With the added capabilities of post-2022 generative AI and the anticipated capabilities of artificial general intelligence, even occupations that involve reasoning about situations not seen before or those that involve interpersonal skills are likely in peril.? That is, the capabilities of AI, as it advances, are intended to include more and more of the skills that make humans distinctive; indeed, this is by intent.? This should give us considerable concern as we consider other skills we could develop that might allow us to move to other jobs not yet disrupted by new technology.? For example, all those auto assemblers who learned to code after being displaced by robots may again be displaced by AI and have nowhere else to turn.

Yet another pattern Webb found should give us even more concern.? He found that not only were a number of jobs eliminated with the introduction of these technologies but the wages were lower among those who remain employed.? This trend goes up the skill ladder as well.? This has the effect of reducing inequality among the large number of people.? But it is because in reducing wages of all, differences are compressed.? On the other hand, he found that wage inequality increased between this large group and the top group of very skilled people, those who, so far, are least threatened and benefit the most by the wide spread use of Ai.

This finding corresponds to a macro-economic trend in the U.S., over the past 30 years, decades in which various digital technologies have become pervasive, the top 10% of the wealthiest people come to own 69% of the nation’s wealth and the bottom 50% own only 2.8% of the wealth.? The widespread use of AI may very well make this situation worse.

So at least for this one technology—AI—the techno-optimistic claims of Andreessen must be moderated.? The “lies” of job loss, wage reduction, and inequality growth that he attacks turn out to be truths.

Technology and Morality

Andreessen frames our thinking about AI in moral terms: the development and proliferation of AI is a moral obligation; any deceleration is immoral, he says.

But are AI and other technologies inherently good?? No.? Nor are they inherently bad.? However, their design and use raise inherently moral issues.?

What do I mean by that?

Dutch philosopher Peter-Paul Verbeek considers technology and design “doing ethics by other means”.? That is, the way technology is designed and used has moral implications.? Guns, for example, are neither inherently good or bad.? But the fact that they are designed to maximize damage to the human body when used has moral implications and shapes the morality of their use.

Does this mean we shouldn’t produce guns?? Not necessarily.? But if we don’t consider the morality of the impact of the way guns are designed and used, that is immoral.

Is the deceleration of AI’s development inherently immoral?? No.? What is immoral is advancing the development and use of AI without considering the moral implications of its design and use.

One of Andreessen’s “enemies” is the Precautionary Principle, which he considers deeply immoral.? The Precautionary Principle has four central components: taking preventive action in the face of uncertainty; shifting the burden of proof to the proponents of an activity; exploring a wide range of alternatives to possibly harmful actions; and increasing public participation in decision making.? It does not mean stopping the development of a technology that holds promising positive impacts.? It means considering the potential negative, immoral impacts of it use and designing means to minimize that.? It also holds designers accountable, that they not be allowed to push the responsibility of all the negative consequences of its use onto the individual or society, at large.

What is morality when it comes to technology if it is not being concerned about the harmful impact of its use on people??

Will Technology Save the World?

Am I a techno-optimist or a techno-pessimist?? Am I a “boomer” or a “doomer”?

Let me put it this way: I think technology and particularly AI hold immense possibility for making the world a better place.? They also hold a variety of potential harms, from the elimination of jobs and growing wealth inequality, cited above, to anxiety and depression among teenage girls.? Whether it makes the world a better place or not depends on how wise we are in its development and use.

It is particularly unfortunate that in Marc Andreessen’s version of techno-optimism sustainability, ESG, Sustainable Development Goals, social responsibility, stakeholder capitalism, tech ethics and the Precautionary Principle are enemies, because it is the application of these ideas and principles and the moral values they represent that increases the likelihood that we will realize the positive benefits we hope for with these technologies.? As much as technologists think in binary terms, technological development and precaution are not antithetical.? You can both advance the life-saving capabilities of AI and reduce the likelihood that it will cause harm, indeed we must do both—that is the moral imperative!?

We can use technology to make the world a better place ... if we value people and their well-being more than we value the technology, per se.? But Andreessen and his Manifesto raise the value of technology development above all other considerations.? And as long as Andreessen-Horowitz impose this maximalist Manifesto on prospective developers of technology, it increases the likelihood that we will see the worst of its impact.