PSA: Man + Machine + Model

In their bestselling book, The Second Machine Age (2MA), MIT Professors Brynjolfsson and McAfee, invite “more novel and radical ideas – more ‘out-of-the-box thinking’ – to deal with the consequences of technological progress.” (pg.245-246)

We’re interested I hearing which ideas you like best, and others you would like to suggest. Contact us at www.SecondMachineAge.com to share your insights. (pg.247)

Here is my offering, from an area of interest to me – the global crisis in higher education.

2MA Claims

I’m not sure what to make of the reasoning found in 2MA. In its publicity, the authors, publishers and reviewers suggest the book offers cautious optimism for the near future. But by my lights, the reasoning is more of a confused optimism for the future of man and machine, in general and with respect to education.

Consider the following claims which the authors feel are supported by research:

1. Exponential improvement in computing, digitization and recombinant innovation of this age will continue to result in increased bounty, but also capital owner gains and superstar-winner-takes-all markets that produce increased spread with power law distribution. (Chapters 3, 4, 5 & Chapters 7, 8, 9, 10)
2. Relative performance drives consumers to select the highest quality product or service among markets to which they have access. (Chapter 10)
3. Complementary work shared between man and machine is a haven for human capital, along with non-routine mental work such as ideation, creation and innovation. (Chapters 12, 13, 14)
4. Thinking outside the box produces the best results when done in collaboration with a wide mix of people having varied experiences, knowledge, perspectives, values, etc. (Chapter 12)
5. Entrepreneurship should be encouraged, where this is understood as the introduction of unique, new products, services or processes. (Chapter 13)
6. Work is important to individuals and societies, as more than a source of income. (Chapter 14)

Next, the authors note one universal question asked by those with whom they discuss their work:

I have children in school. How should I be helping them prepare for the future you’re describing? (pg.188)

As part of their response, the authors approve of - are even keen on - MOOCs as a bounty of the second machine age. However, when taken together, claims 1-6 have implications that make their endorsement of MOOCs puzzling.

Mamas Don’t Let Your Babies Grow Up to be Cowboys

First, consider claims 1 and 2. The authors take the time to show how the (socioeconomic) spread is worsening due to superstar-winner-takes-all economics. By endorsing MOOCs in Chapters 12 and 13, they at least tacitly endorse application of this same economics to (higher) education.

We hope that MOOCs and other educational innovations eventually provide a lower-cost alternative to traditional colleges, and one that is taken seriously by employers, but until that time comes a college degree remains a vital stepping stone to most careers. (pg.201)

This has three implications, which the authors fail to address in their response to the universal question:

1. One nugget of advice they should be offering parents is: Don’t let your children grow up to be professors - since in the age of MOOCs the occupation is governed by superstar, winner-takes-all, power-law-distribution economics. Consequently, there will be little opportunity to make a living in academia. [Even less opportunity than exists in the severely depressed academic labour market of the current higher education model, since, to mix the authors’ examples, claims 1 and 2 cause Kahn Academy to bankrupt Kodak, leaving most college and university employees out of work.]
2. If there is less incentive to become a professor, then there will be fewer professors. Think of every burgeoning artist whose parents insist they stop wasting time at that easel and get back to their economics or calculus or anatomy homework. It is hard to know how many great artists or developments in art have been lost in this way. [Under the current higher education model, some have calculated the odds of becoming a professional athlete at equal to or better than becoming a professor. Of course, if claims 1 and 2 are correct, MOOCs will make it far less probable that someone can become a professor.]
3. Though there are reasons other than employment to get an education, when claims 1 and 2 are applied more widely to other occupations - as 2MA suggests is likely - there will be a more generalized dampening of incentive to seek (higher) education. If there are only 3 professors or lawyers or doctors or engineers or architects or whatever in the (global) market, then there is little point spending the time (if not the money) on the education necessary to attain such positions.

Entrepreneurs Without Ideas

Second, considering implications A-C with respect to claims 1-5, it seems confused to champion a combination of MOOCs, ideation, creation, innovation and entrepreneurship. As the authors correctly point out, good ideation, creation and innovation requires collaborative input from many people with diverse knowledge, beliefs, life experience, values, training, professional/personal investment, etc.

It is also known that professors tend to select graduate students who share their research interests, ideas and personality traits. As good as Sebastian Thrun might be at teaching AI, he will tend to gravitate toward collaboration with compatible students. And students will tend to gravitate toward a professor based on measures of relative superstardom performance. Together this will produce a self-selective system of incestuous education, even assuming there is still widespread participation in the higher education system at all.

With this in mind, to endorse MOOCs would seem to have the following further implication:        

       D. The depth and breadth of ideas, creation, and innovation is likely to diminish where MOOCs are increasingly used, because the diversity of cognition and character engaged in the provision and development of (higher) education is substantially reduced by MOOCs.

Surely this will lead to a high degree of homogeneity and stagnation in any given field. Just as diversity and competition among other products and services is desirable, so it is with ideas - while monopolies license the mediocre.

The authors open Chapter 5 on innovation, with the Linus Pauling quote, “If you want to have good ideas you must have many ideas.” (pg.71) Of course, this is not restricted to an individual, but appropriately and perhaps more effectively applies to groups of people. So, maybe homogeneity and stagnation will be overcome by the types of open, crowdsource, network systems the authors highlight as innovative connectivity (e.g., NASA, DARPA, Innocentive, Kaggle, MOOCs, etc.).

However, consider that for these types of systems to work, participation rests on two conditions: 1) socioeconomic incentive to answer the call for ideas and 2) the knowledge and skill necessary to generate good ideas. But arguably these conditions are threatened by claims 1 and 2, along with implications A-C. And even where these conditions are met it remains true that the proffered ideas are filtered by the few dominant superstars of the academic field or market.

In turn, implications A-D lead to another overlooked implication of the reasoning in 2MA:   
     
       E. To the extent that entrepreneurship depends on conditions of quality ideation, creation and innovation, the prescribed entrepreneurial mechanism for maintaining the value of human capital will be crippled by MOOCs.

The authors create confusion by, on the one hand, endorsing MOOCs that arguably stagnate these conditions and, on the other, calling for more support of entrepreneurship, as a positive response to the age.

But this confusion reaches further. 

It is well-established that the teaching and research performed in higher education has contributed to both basic and advanced knowledge used to help create the second machine age. It is also widely recognized that teaching and research effectively support one another in the production of this knowledge (including good ideas, creativity and innovation). MOOCs tend to divorce this mutually supportive relationship, resulting in the further degradation of the conditions necessary for entrepreneurship.

Also, the authors note that “the top performer in [Thrun’s graduate-level AI course] at Stanford, in fact, was only the 411^th best among all the online students.” (pg.199) This was presented as a point of praise for MOOCs. But seen through the wider lens of their book (claims 1-5), it is more a point of pity: a pity that thanks to machine age MOOCs and monopolies there will be no room for this talent in academia or other sectors of the economy. And so again, the development of diverse talent in teaching and research – something necessary for good ideation, creation and innovation – is thwarted by the power-law economics of MOOCs championed in 2MA.

A Bounty of Scarce Work

Third, consider claims 1-5 and implications A-E in relation to the final claim 6. The authors quote Voltaire:

Work saves a man from three great evils: boredom, vice and need. (pg.229)

They highlight the value of work to individuals and communities:

And just about all the research and evidence we’ve looked at has convinced us that Voltaire was right. It’s tremendously important for people to work not just because that’s how they get their money, but also because it’s one of the principle ways they get many other important things: self-worth, community, engagement, healthy values, structure, and dignity, to name just a few. (pg.234)

The evidence suggests that communities in which people are working are much healthier than communities where work is scarce, all other things being equal. So we support policies that encourage work, even as the second machine age progresses. (pg.236)

As the authors repeatedly state the second machine age will bring increasing bounty, but clearly it will also introduce increasing scarcity. Along with the loss of work for professors under the MOOC model of higher education there will be plenty of other lost jobs: administrators, managers, and support staff. This has occurred and will continue across other industries, which is one of the reasons the authors offer entrepreneurship as a consolation prize – which, as we have seen, is puzzling since MOOCs threaten the very conditions necessary for (technical) knowledge development and entrepreneurship.

But the threat of MOOCs is not just to the available opportunities for work in higher education or their effects on entrepreneurship and the demand for education in other industries. At least with respect to the occupation of professors, reducing the number of professors reduces production of the full spectrum of their work, which reaches well beyond the teaching facilitated by MOOCs. I have discussed this more thoroughly in a previous post, but suffice it to say that the loss of this work would be extremely detrimental to the health of society.

As the authors conclude,

Our success will depend not just on our technological choices, or even on the coinvention of new organizations and institutions. As we have fewer constraints on what we can do, it is then inevitable that our values will matter more than ever. (pg.257)

I couldn’t agree more, which is why I invented the PSA model for higher education. It strikes a healthy balance among the things we value as individuals and communities, including what some would say is the right to work.

Fear Not for the Future, Weep Not for the Past

Throughout 2MA, Professors Brynjolfsson and McAfee write with an optimism meant to mitigate the truth that eventually machines will have the capacity to supplant all human capital. Ignoring the disruptive cyborg middle ground between man and machine, where technology-enhanced humans increase spread and scarcity, the authors discuss a thought experiment that makes humans an economically endangered species:

Imagine that tomorrow a company introduced androids that could do absolutely everything a human worker could do, including building more androids. There’s an endless supply of these robots, and they’re extremely cheap to buy and virtually free to run over time. They work all day, every day, without breaking down. (pg.180)

They speculate that in addition to incredible bounty there will be,

…severe dislocations to the labor force. Every economically rational employer would prefer androids, since compared to the status quo they provide equal capability at lower cost. So they would very quickly replace most, if not all, human workers. Entrepreneurs would continue to develop novel products, create new markets, and found companies, but they’d staff these companies with androids instead of people. The owners of the androids and other capital assets or natural resources would capture all the value in the economy, and do all the consuming. Those with no assets would have only their labor to sell, and their labor would be worthless. (pg.180)

Entrepreneurism is a type of work and since androids could “do absolutely everything a human worker could do,” the experiment would seem to preclude the need for human entrepreneurs. Further, I assume the universal quantifier applies to CEOs, CTOs, COOs, on down the company structure. Truly, the only humans involved in the economy are owner-consumers.

Thanks to Moore’s law and power-law distribution, value would be passed among this absurdly small number of humans. At this point, capitalism as we know it has lost its appeal for all but a few and is radically adjusted, with (among other changes) a redefinition of what it means to be economically rational: Perhaps what is economically rational becomes what is socially rational.

If capitalism doesn’t lead to socialism in this future, it’s hard to imagine why a handful of humans who own capital assets and natural resources would bother with traditional capitalist measures of value in a post-scarcity, post-wage world. Certainly, it would be unimaginably narcissistic, even evil, to horde the bounty and anyway such a possibility ignores the roll of government in socioeconomic management, as the second machine age opens dialogue on the ultimate values of humanity.

I am excited for such a future. It seems to me a post-work era is a post-value era – or at least one where value is reduced to idiosyncrasies that don’t impact material needs and wants. For example, assuming the Pope has not been digitized, an audience with him might be of value to some. The same would be true for a meal prepared by a world-famous chef or a painting from a renowned artist. In each case, originality is the value, which could no longer be exchanged through traditional capitalist leverage.

As I read 2MA, there is a sense that Brynjolfsson and McAfee feel we have left to us a generation, maybe two, before the android thought experiment becomes a full-blown social experiment. But no matter the timeframe, their optimism is reserved for select interim work prospects. Outside of these domains, life will be miserable.

In response to their call for “novel and radical ideas” that mitigate this misery, the PSA model addresses the work prospects and values within and without higher education, while it avoids implications A-E.

The advancement of humanity requires a substantial reduction in the cost of higher education, combined with substantial increase in its valued human capital inputs (students and professors) and outputs (research, ideas, community service, etc.). However, the current higher education institution model of colleges and universities seems unable to accomplish this and, given implications A-E, it’s not clear that MOOCs will fare much better.

None of this reasoning should be read as Luddism. From pen to press to program, there is room for technology in education. But acknowledging this does not diminish our responsibility to use it effectively and morally.

PSA is not inconsistent with the use of technology or colleges and universities. Instead, the model adjusts their status from requisite to elective, allowing us greater discretion in their use. To see in full how this is so, read the PSA prospectus and blog posts, but in overview:

• It creates value for human capital through innovative organizational capital.
• It is a recombinant idea, not “something big and new, but instead recombining things that already exist.” (pg.78) It uses organizational building blocks from the first university to 19^th century professions, along with modern digital technology.
• It is a meta-idea that, as Paul Romer says, supports “the production and transmission of other ideas.” (pg.79) It accomplishes this by increasing the quantity of professors and students, improving the quality of their interaction, and implementing technological support as required (e.g., crowdsource assessment).
• It is an entrepreneurial model, in the sense of organizing, managing and assuming risk and responsibility, along with innovative product and service creation, within a private higher education practice. It mitres nicely with the increase in entrepreneurialism already underway in the academe, thanks to failure of the current model.
• It is a professional model, offering academic services as lawyers, doctors, accountants, psychiatrists, architects, etc. offer their valued services to the public through professional association, license, and practice.

Universities and colleges always have been nothing but inessential institutional middlemen. They facilitate the work of professors, including research and the student-teacher relationship. They facilitate the disbursal of public and private investment in higher education. They facilitate quality control. They are elective.

Professors are essential, at least until the machines reach full maturity. Without them there is no such thing as higher education. And their work can be facilitated, developed and overseen as it is within other professions. This is indisputable. The only question is whether it is to be preferred.

We gravitate to colleges and universities because they have a long tradition. We gravitate to MOOC technology because it speaks to the future. We grasp at both for desperate relief from the critical state of global higher education and because there seems to be no alternative. But PSA is a viable alternative and one that avoids the negative features of both higher education institutions and MOOCs, while it capitalizes on their positive features.

PSA will not stop the march of the machines. As I have acknowledged elsewhere, the day is coming when everyone has access to a digital professor, indistinguishable from the human version, but absent the shortcomings. In fact, by expanding student and professor access to quality higher education, PSA can help expedite introduction of professor-Watson.

And I agree with Brynjolfsson and McAfee, as much as possible we should not impede the march, though we must try to mitigate the misery and validate our values in the process. As a strategy, PSA is well-suited for this interval between the past and the future.