Prelude to Awakened Enterprise (2/2)
Digital = Magic
Harnessing advancing computing technologies to surpass legacy limitations.
Evangelizing “digital” as a game-changer is tired by nearly three decades. However, we should not under-utilize the transformational capabilities the ride on the exponential advances from Moore’s Law has made abundant and affordable in modern digital processing and storage technologies. We envision revolutionary human-machine collaboration harnessing modern information infrastructure to complement deficits of the human mind.
The following discussion builds on “Despite Uncertainty” to lay key foundational concepts as a prelude to the Awakened Enterprise vision.
Warping Energy, Matter, Space & Time
Simply qualifying “nearly” to terms such as: instantly, reliably, infinitely, and cost-free would allow us to make incredible but true assertions, such as:
In stark contrast to our familiar experience with physical goods, digital objects can be made, replicated and globally transported — instantly, reliably, infinitely, and cost-free.
We can exploit parallel processing and communications to tackle and tame massive datasets, concurrently processing and re-synthesizing these into results that are indexed and cross-correlated in (seemingly) infinite archives able to be searched (nearly) instantaneously.
Digital processing, memory, networks, and robotics can virtualize physical materials & processes and perform cognitive & physical labor, at blinding speed and inconceivable scale — altering our perception of space and time.
Magics of Digital Technology
Commonplace uses of present-day digital technology applications perform tasks that are only familiar to us as magic in literature; perhaps now more constrained by human imagination and cultural hesitation than the need for technical invention.
Consider these supernatural powers:
- Telepathy: instant non-verbal (text/media) communication,
- Scrying: audio-visual interaction across great distances,
- Clairvoyance: far-seeing and knowledge of the future/unseen,
- Enchantment: awareness in physical objects, or invisible agents to task,
- Omniscience: perfect memory of all human knowledge.
Digital technologies put these magical powers within our grasp:
- Our pockets hold devices from which we can instantly, anywhere consult vast human experience and expertise (Wikipedia), step by step instruction (YouTube), and navigational geography (maps + GPS).
- We can consult and socialize with communities based upon group interests, affinities, and ideology rather than our local geography.
- Our extended environment can be captured through distributed sensor networks and crowd-sourced data to track weather, traffic, or illnesses.
- Innovative media allows digitally superimposing data and images on the physical world in real time (rendering mixed realities) — highlighting tumor cells for a surgeon, labeling parts for a mechanic, or envisioning architectural plans for new construction.
- Through mixed reality and high-speed networks, we can even project centralized core expertise to field agents through two-way live interactive media — for medical clinics or technical repair service calls.
- We can conceive, build, and maintain products and services with greater efficiency, quality, and agility — sharing data and models (such as digital twins) across physical distance and organizational boundaries.
- Digital agents perform tasks on our behalf (repetitively, tirelessly, consistently, and obediently) — tasks of increasing complexity via AI.
Complementary Strengths
Several common human fallibilities can be mitigated by enlisting the capabilities of computational machinery. Our biological state is susceptible to fatigue and boredom, whereas machines are purpose-built for continual and repetitive tasks. Human minds are prone to subjectivity and bias, while logic machines follow rules with formulaic consistency.
Brain memory suffers notorious unreliability, inaccuracy, and limited capacity, while digital storage technology has flourished — providing vast volumes of data accessed with high-bandwidth, low-latency, and essentially perfect recall.
While some people boast to be gifted “multitaskers”, studies show this aptitude more accurately described as selective attention. In contrast, computer processor architectures contain independent “cores” explicitly dedicated to independent task execution. Multiplied over racks of servers in a datacenter, computing systems can focus on and simultaneously execute billions (and billions of billions) of concurrent tasks.
All these tireless and consistently logical digital brains can then be aimed at superhuman tasks — beyond human perception or comprehension. For example: interpreting entire slide — rather than consecutive views of what is in focus of a microscope or employing precision concurrent control of thousands of orchestrated drones.
Conversely, “Framers” by Kenneth Cukier et. al. offers examples of how human skills and aptitudes provide complementary strengths to machines. For example, improvisation in adapting to emergent situations employs tactics that are inherently non-formulaic. Models rely upon human sense-making, causal inference, and bounding by asserting constraints — driven in purpose through the imagining of hypotheses and “what-if” scenarios
Machine Awakening
Affordable and abundant digital storage and processing has resulted in mind-boggling volume, velocity, variety, & veracity of data. The new data created every day is estimated to be roughly equivalent to every US citizen (~330 million) entirely filling the storage on a new laptop (~1 TB).
To query such vast data, we rely upon search engines such as Google and Bing to index these vast data — as well as services like WebMD, arXiv, and Justia for specialized topics (such as biomedical, scientific, and legal publications). Yet even employing such tools, our manual, ad hoc means of searching taxes the limits of human perception and cognition.
Artificial Intelligence innovations offer promise to tame these vast digital data (while subsequently stimulating further demand for even more data). Machine learning (ML) techniques have evolved rapidly in recent years, exploiting the richness of this “Big Data” itself as the basis for training ever-increasingly “intelligent” models. Automated software agents, instructed through semantic queries can persistently query and monitor updates to the information indexed by these sites.
In 2023, extraordinary chatbots captured the public’s attention, thanks to LLMs (large language models) built on clever techniques (GPT) and monumental data and processing. One prominent use of chatbots taps into generative capabilities (writing text, computer code, and visual media).
Another chatbot capability leverages conversational interaction, now approaching the semblance of human beings — allowing non-technical users to task powerful computing machinery without learning languages for computer programming. This feature, however, is flagged as a risk by Cassie Kozyrkov, noting:
Think of [ML/AI] as a proliferation of magic lamps. […] It’s not the genie that’s dangerous, it’s the unskilled wisher. […] The scary part of AI is not the robots. [It’s the] teeming multitudes of well-meaning bumblers.
Of course these benefits are not without cost — the price for these powerful capabilities through the tremendous energy consumption of data centers. The energy (and emissions) to power the magics we tap into through AI result not only from their massive training efforts, but in each and every query performed.
Digital Dangers
Our literature is abundant with cautionary tales of runaway automated machine servants unintentionally escalating complexity to surpass the pragmatic needs of the enterprise or individual. A modern version has been captured in the “paperclip maximizer” thought experiment.
There exist sufficiently plentiful additional concerns and caveats to merit independent threads of discussion. Four are listed here for reference:
- Obstacles to implementing enterprise Digital Thread — from functional architecture to organizational culture (including misaligned incentives where one division incurs costs, risks and resource burden for value awarded to another division.)
- Employing the vast information from web searches to second-guess in hindsight — on the mistaken belief that something was known all along rather than rapidly assimilated through the agility of indexing agents.
- Placement of the human in automation strategies — Tom Limoncelli’s “Automation should be like Iron Man, not Ultron” perspective contrasts the “Leftover Principle” (automate everything possible, assign humans whatever is left over) vs. “Compensatory Principle” (collaboratively apply complementary strengths).
- Artificial Intelligence explainability, transparency and what Jonathan Zittrain terms intellectual debt — loss of autonomy in knowing what we do not know, and thus ability to adjust to the unusual, and to grow.
Awakened Enterprise
The Awakened Enterprise leverages increasingly capable automation, abundant and affordable data repositories, and complexity-taming machine intelligence to overcome limitations in human span of attention, fatigue in tediously repetitive vigilance and unreliable ad hoc memory.
Recognizing preparedness and opportunity in these environments relies upon taming VUCA characteristics of internal operations, the supply chain, and the market. Leveraging strengths of digital technology and AI to complement human cognition, we can devise knowledge systems to overcome traditional human limitations — which evolve to collaborate with human organizations through a degree of awareness (“awaken”) .
Digital Technologies can build systemic awareness and enable a coherent frame of reference to make better decisions and actions despite spans of time, organizational obstacles, and other sources of uncertainty. The form this will take could revise how we organize our knowledge assets to be indexed and labeled so crucial conditions within archives are reliably and efficiently findable for consideration in future decision-making.
© 2024 All Rights Reserved
