My attempt to synthesize the last several years of the emerging field of technology criticism into a set of recurring general principles. These ideas belong to many different thinkers. The contribution here is primarily distilling them down to their essential point and collecting them in one spot. My next step is to provide a “see more” section for texts and a “problems and examples” section.

  1. Data is never “raw,” immanent, or neutral. There is always bias and distortion in capture and modeling.

  2. The internet is not “a thing.” It is a distributed network of many layers. Treating it as its own monolith with a central cultural logic presents problems.

  3. Technology can never occupy a space outside of capitalism. With rare exceptions, every application, company, or innovation will have a funding source, a board, and a bottom-line; and in all cases the logic of capitalism will eventually supersede and control technical tools. What we identify as “tech” is just capitalism, but faster and worse.

  4. You can’t solve a social problem with a technical solution. Often, applying technical fixes only treat the symptom, and, in failing to address the underlying cause of the problem, makes it worse.

  5. If you are not paying for a platform, your data is the product. Attention is data and data is a commodity. If something is free and connected to a network, beware of the tradeoffs.

  6. Platforms are not institutions. Do not confuse them.

  7. Decentralization is an illusion. Even distributed networks enforce hierarchies of power and influence.

  8. Software is hard. Computing interfaces, rules, interactions, and protocols encode certain behaviors, and for that they should be scrutinized and interrogated as part of the body politic and the built environment.

  9. Algorithms are made of people. They are editors, they steer and privilege certain values, and are never objective.

  10. Beware of “open access.” Information may want to be free but beware of the consequences. Somewhere a new gatekeeper will benefit.

  11. Once a measure becomes a target it ceases to become a measure (Goodhart’s Law revisited). Or, when you over-optimize for a goal you’ll often destroy the thing or the market you set out to augment. Or, optimizing for a goal in a closed system will reinforce the production of that goal, and cease to deliver any insights.

  12. Information is the enemy of narrative. The more information, the more doubtful the narrative becomes.

  13. Crowdsourcing is a race to the bottom. Labor, knowledge, education, etc.. are all cheapened when forced to compete on a platform. Making it easier to perform a task has massive externalities.

  14. Your brain is not a computer and your computer is not a brain. There are things that cannot be automated, and intelligences that machines cannot have.

Mike Pepi @mikepepi (last updated 8/15/2018)

AuthorMike Pepi