Cryptanalysis, the Turing Test, and the Language of Business
Unless you have done so, you just can’t fathom the elation of breaking code or reconstructing a communications network. I did both for the first time when I was 18. My first radio network was a Soviet era one, probably from the early 70’s. My first cracked code was a simple substitution. In each case, my reaction was, “I now know you. I can figure out what you are doing.” Yes, from brain straight to ego, the ultimate mental orgasm. The experience changed me forever, as I became a lover of language. Language changes us, yet we control its evolution. Put to meter, language thrills. However, language always adheres to logic and rules. Like its offspring, math, language conveys knowledge. Unlike math, language is at its best when it creates its own new rules. Structure is the yin that allows us to use language as a way to convey knowledge, while whimsy is the yang with which language derives its greatest value. The whimsy has also made language hard to master algorithmically. In this thread, I posit that we are at the beginning of a new language based business paradigm. With this new paradigm–to ultimately understand and perform projects at anything near a six sigma success rate–we have to use the project team’s own words as our metrics.
How we got here: the story of cryptanalysis
Whispered words in a confidant’s ear were the earliest and simplest method of obfuscation to the words we spent so many millennia to utter and share. Over time, our warring ways led us to devise obfuscation at a distance regardless of time. The scytale was the first tool that could encipher information. The message could be carried afar, and deciphered much later. Of course, with every whispered word, there were listeners wanting what was denied. As soon as the first simple substitution ciphers were developed during the rise of the Muslim faith, letter frequency cryptanalysis was documented in tandem.
One only has to look at the threaded discussion posts leading up to this, and you will be reminded why the Renaissance was such a pivotal time in human knowledge. Just before Pacioli documented the double entry form of accounting, Cicco Simonetta wrote a deciphering manual for Italian and Latin text. Manual systems of increasing complexity were the mainstay until the beginning of the 20th Century. Up until that time, reading that which needed encryption was considered sometimes necessary, but always distasteful. Secretary of State Henry L. Stimson once stated, “Gentlemen do not read each others mail.”
However, the Zimmerman Telegram aftermath and the strategic advantage for the US during the Washington Naval Conference convinced many of the need for a full time cryptanalysis organization. One of my heroes, William Friedman, was a key player for the US on both the encryption and cryptanalysis fronts. However, if Newton and Einstein were the smartest to have ever lived, it would be Alan Turing who could sit at their table and join the conversation. From his Banburismus to the playfully named Turingery cryptanalytic processes, Turing used his deep understanding and contributions to probability theory to break the German Enigma during World War II. In doing so, he became the father of intelligent computing and the rise of natural language processing.
The Quintessential Intelligence Test: The Turing Test
After the war, Turning wrote an article for the Mind journal called Computing Machinery and Intelligence. It was the first public discourse of the Turing Test, the base meme of the Geekhood. The test basically postulates if a computer can converse with a human evaluator where the evaluator cannot tell the difference from a human conversation, the computer has reached intelligence. Turing’s key insight was that language—namely, the conversation–is the currency for knowledge and ideas. With this, Turing gave us the new accounting tool for the intangible economy.
Ideas inside our head have no value and are only useful to one. When wrapped in clear language with soaring adjectives, ideas can be instantly monetized as intelligence. Many can do things with ideas in the wild. It took a man steeped in probability, challenged with breaking encrypted words to save the world, to give us the amazing insight to the power of language. From the first sounds during the prehistoric hunt to organize, language has evolved to truly run the world. Now, we must tame language with algorithm so that we can measure our business.
The Language of Business
If the double entry form of accounting valued and accounted for tangible goods and things, it is language that values and accounts for intangible ideas. The double entry form of “where did I get it, and where did I put it,” is replaced by the ‘what’ am I asked to do and ‘how can I accomplish it’ algorithm.
As featured in a previous webinar on this blog, project is the instantiation of work needed to accomplish a goal. Regardless of how we want to articulate the work in aggregate–the current Agile-waterfall debate–each granular object of work, an activity or task, is expressed as a verb-direct object combination. Develop requirements. Write code. Test code. The implied actor is the resource assigned to the work.
Ok, so stay with me here. Just like the daily ritual of brushing our teeth, humans like the comfort of routine. Individually, routine is called habit. In the collective business world, habit is called process. And what is process? Process is a series of related actions that transform an input to an output. A process step then is equivalent to a project task, no? So basically, no matter how one looks at it, work in business follows the basic grammar structure of one of the five basic sentence patterns, subject-verb-direct object. Patterns, they are everywhere!
Since Turing, we have become increasingly adept at manipulating language patterns with computing devices. Siri, while not ready to pass the Turing test, is so very close to a usable personal assistant that we can all now picture the day when she will. Inside Siri are deep learning algorithms and pseudo knowledge algorithms based on graph theory. Siri is a generalist. What would a Siri like capability look like when trained in the language of business?
I have one more topic to cover before I wrap all of this together in what business is about to become, and needs to become. The next thread is about people and their need to tribe.