The Singularity

Every now and then, the conversation resurfaces: when will The Singularity occur, and what will we do when it happens?

A basic question about it is, if only the most talented of us are worthy of power, money, and recognition, and if those things are directly related to quality of life, is it ever in our best interest to enable programs to become better at what we do than we are? Of course this hinges on what people define as talent and worth. In business, value is often determined by rarity of skillset, where rarity is diffused by the availability of less expensive alternatives.

It's an important question for technologists, because most of us have spent our careers writing clever scripts or programs to do the tedious parts of what we do. This cleverness has led to business leaders and opportunists to encourage this trend as a means to reduce that rarity and pay less for technologists, or to justify having fewer of them.

It's a familiar argument to Infosec folks: the magic bullet application that protects all things so businesses can cut expenses on headcount.

So, with this observation, I have a proposal.

This trend has been most prevalent in technical fields because we have the capability, the knowledge of the systems involved. We started doing this because enables us to eliminate the parts of the job we like less so that we theoretically have more time to do what we enjoy more. But, seldom does this trend stop so that goal can be realized. Instead of having more time free to do things we enjoy, we are simply asked to do more that we do not enjoy under the assumption we are an infinite resource to produce infinite work, and under the fundamental desire that drives American businesses: get more, but pay less for it.

The drive to pay us less for this expertise so that businesses can enjoy more profit margins ultimately leads to products that promise to replace the need for people. And, in the end, all but the most fortunate, diversely skilled, long-working, cheapest, or most entrepreneur are able to remain employed in their fields while companies limp along with poorly configured products that no one understands and which may or may not be effective at their advertised function.

The development of tools to make work easier is a worthy pursuit. But, we have evolved past creating tools that we can use, and instead are using programs to replace people. And there has been tremendous focus in doing this within technology fields. The argument that people will retrain and find other jobs is a shaky one, because the theory that human achievement is limitless for everyone is proven to be false. And, because someone might be able to retrain himself to do some other task, that does not mean it is a task that person will like as equally. Nor does it mean that a replacement system will do as well. Consider for a moment, the Roomba. Sure, it will sweep your floors. But, will it do as good a job at it as a capable human who paid appropriate attention to detail? Of course not. But, which is cheaper?

This will continue to happen while it is appropriate for people to believe that 'other' people are interchangeable, or that what one capable person enjoys doing is worth less than whatever some other person might do, or while focus remains on short-term profitability while ignoring prospects of long-term stability.

So, I propose three things.

The first is, that tools should enable people, not replace them. Developers should consider this seriously in terms of the human impact of complex algorithms and self-sustaining systems. If there is no niche for displaced people to fill, then that must be considered part of the problem being solved, or else a true singularity will be disastrous for humankind. Consumers should consider this from the perspective that tools make jobs easier, but they do not make them go away.

The second is, we should put more focus into automating or creating tools to enable people without 'business sense' to operate more autonomously. We have put sufficient focus on reducing costs at the bottom; now we need to focus on reducing costs at the top. Let's identify ways we can enable technical people, or people with great ideas to work for themselves instead of for other people. Put some of our tremendous tech talent on that problem for a while, and end the trend of electronic feudalism before it's too late.

The last is, in relationships with your children, with your schools, with your peers, your subordinates, your superiors, your neighbors, and anyone else in your society as a whole; consider carefully how you make your value judgments, and see if you can find ways you can personally influence those judgments towards a human-centric point of view. We can all agree that it is fair to expect society's members to contribute in order to receive benefit. But, if the degree of contribution is directly linked to quality of life, we have a bigger problem when value judgements take place. Of course incentive to participate is important for many. But, is a janitor worth less than a systems administrator? Is a systems administrator worth less than a CEO? How does this play out when the only ones doing any actual work are the computers, and the only ones who can afford the computers are monolithic corporations with billions at their disposal?

No comments: