Notes Towards a Definition of SLICCs: The NSA variant
“Solutions” that amplify, to a rentier’s profit, the very “problem” they claim to solve.
He gives examples of such in space exploration and the defense procurement.
As it happens, I read his post just after finishing a Harper's piece on William Bratton and "Intelligence-Led Policing" as well as Andrew Cockburn's Kill Chain. Both detail --albeit in regrettably empirical, a-theoretical form-- examples of what, after the break, I am hereby respectfully submitting for discussion as the "NSA form" of the SLICC, to distinguish it from the "NASA" and "DoD" forms already elucidated by lambert.
I. Define the problem (terrorism, insurgency, crime) as one that, self-evidently, would admit to a very simple, direct solution (kill, capture, fine) if sufficient data were available. Deprecate any existing approaches to the problem as messy heuristics, rough-and-ready mental shortcuts whose biases and inefficiencies have been tolerated only because the unfortunate computational limitations of the human brain seem to leave society with no alternative.
Invoke stuff like Moore's Law and quantum computing to posit that now, at last, technology enables the kind of large-scale, centralized data collection and analysis that makes the self-evident, rational solution a practicable policy alternative.
Independently of the nimbus around technology in contemporary society, such solutions have a strong chance of being adopted because they take decision-making away from unskilled or semi-skilled labor (i.e., individuals with no- or low-prestige college degrees, such as Army NCOs, A-10 pilots, or beat cops), vesting it instead on labor so skilled its no longer labor at all, but rather the much nicer "human capital" (i.e., people with postgraduate degrees or at least undergraduate degrees from elite schools). This latter category will usually comprise three critical sets of individuals
(i) You and your rentier friends.
(ii) High-ranking officials in the relevant bureaucracy.
(iii) Policy- and decision-making elites in control of budget allocation.
as well as two highly-useful ones
(iv) Academics at prestigious schools developing techniques for the data collection- or -analysis.
(v) Young mainstream media journalists in expensive cities fretting about their product and platform.
II. Such things as the Curse of Dimensionality and GIGO guarantee that your system will unable either to process all the data collected or to sort useful from non-useful (here speak knowingly about "signals" and "noise"). This may or may not aggravate the existing problem, but scandals will inevitably arise involving transparently terrible data quality or analysis. These scandals are strictly necessarily to your SLICC, because it is only through them that you will begin to control the production of the problems you are being be paid to solve.
Re-package these second-order problems as being --also-- the result of a simple lack of computational capacity. As such, they cannot, self-evidently, be resolved by devolving decision-making back to people with degrees from crappy schools, but, rather, must be addressed by adding further human-capital-intensive layers to your system. (Pro-tip, any layer-addition that invokes "algorithms" can usually be presented as layer-reducing. "Neural networks" will silence any remaining critics, since neither you nor your critics are likely to be among the small number of people who actually know what they are.) The layer additions not only increase the demand for your product and updates thereto, but also create additional funding/hiring opportunities for people in the [ii-v] categories, further increasing your constituency in the bureaucracy and respectable public opinion.
Comments or examples welcome, etc..