Lend me your Fridays

“உயிரெதற்கும் வதையின்றி உண்மைக்கு புறம்பன்றி வரலாறு படைப்பேன்”

in my poem, வைகையில் காலூன்றி

Saturday 13 SEP 2025

IT System Administration is a human thing that can not be AIed/MLed/NLed

I am a regulatory die-hard. Conflict of interest is a is a fundamental principle in all regulations that finds manifestation in the way the civilised world has organised itself in all entities across the spectrum of interests such as government functions, various industries, industry peek-bodies, academic institutions, charities(though I don’t quite like the word because of its inherent ability to miscommunicate intent and practise) and so on.

Conflict of interest manifests in system design as not allocating specifications for mitigation of potential failure of a subsystem to itself- self correction is difficult to design and in most cases its easier/less burdensome to conceive and implement an additional subsystem or atleast a controlled interface to another subsystem to implement the mitigation. If the potential failure is anticipated in a software subsystem, one would have to necessarily assume any mitigation implemented in the software subsystem is also prone to not dissimilar mode of failure; its easier to implement mitigation for software failures in hardware or all together eliminate the functionality from the system and assign it a human user; specifically note, assigning a mitigation of a software failure to a user through labelling is at best naive in critical systems or systems whose failures may potentially lead to permanent human harm (death or serious injury).

Identifying and addressing conflict of interest is a skill that can not be derived from decisions based on rewarding past history, because, most conflicts are triggered by unique situations encountered(unknown/unhandled exceptions) in the field, completely out of the controlled scope of the system designer. Specifically, its my insight that autonomous solutions incorporating AI/ML/NL for administering IT systems whose failure has a potential for human harm can not foresee and prevent human harm; nor can one isolate the root cause upon eventuation of human harm in a way that the system can be corrected to prevent recurrence- its much more practicable for the business to hire good system administrators dedicated to understanding the intent and scope of use of the system, set up measures for assurance of services commensurate with the organisational need. Rule of thumb from my experience with good system administrators is, for every 150 human months of IT use in localised IT systems, you need atleast one human month of system administration efforts- a cost I would budget for in digitising- or there is always the non IT option for critical information needs of organisation of any size - last time I checked , I could pick up a 46 page note book for fifteen cents Australian and a pen or pencil for an additional dollar or two.

publishing prior to review,

Sincerely,

SumathyRamesh

Saturday 12 JUL 2025

Specifying the person in the objective to accomplish security of person: Strength and Weakness

Almost all humans have the cliched attribute that would win their interviewer over: its my insight that every human’s weakness is actually also their strength.

Almost all roles a human is engaged to do make the most of the human’s capacity to play their strength- and rarely give a thought to the situations where that strength manifests as a weakness.

And almost all employments are broken at the first sight of such manifestation, typically driving mediocracy in the work place because the organisational structure does not have the necessary framework to enable its human employees to play their strength in adversity, advancing organisational capabilities or loosely described as organisational learnings**, preventing not dissimilar adversities and thus striving towards perfection. Its also my insight that complex multinational-organisational frameworks assume a geographic-monolith of its operational priorities feeding cavitations that would burgeon unattended to until it pushes the organisational learning frameworks to sub par mediocracy -often irrecoverably. Given businesses exist to make money through serving their offerings, the offering leaving their lines a precursor for the inward flow of its money, business decisions reach out to external businesses forgetting that the external businesses exist to resort to make money through serving their offerings, the offering leaving their lines a precursor for the inward flow of its money, often recursively depending on the extent of the customer base, the complexity of the offerings and the presence/absence of a sugar-daddy-conglomerate exposing themselves as takeover targets- at this point the priorities to perfect the offering may potentially have lost its place- often irrecoverably. AI/ML/NL based solutions dig sink holes of consensus reality by amortising human engagement in one or more businesses cementing mediocracy as a worthwhile pursuit. அஞ்சுக்கு ரெண்டு பழுதில்ல is ok forஜாதக பொருத்தம்,* its completely unacceptable for solutions that have consequences to human life, liberty and security of person.

Sincerely,

SumathyRamesh

Founder, Writer, and Operator

lzerobzero® a sole trader entity registered in NSW, Australia

eatwell@lzerobzero.com.au | livewell@lzerobzero.com.au | sleepwell@lzerobzero.com.au

* horoscope matching is a practice for arranging marriages in the culture I hail from- it has a lot of practical merit if applied sparingly, mostly it would be a goer, if the couple fulfill two out of every five criteria, given:

a) its an imprecise application of validation of a fairly decent compatibility checklist

b) they have consciously engaged in understanding the unmatched criteria, and decided - potentially on their objective knowledge of the capacity of the couple to navigate potential adversity they may encounter

** I’ll write more about it later, its my insight that strong middle management is key to the strength of a mature organisation that knows to learn and grow