Systemic Omens - AI and Work

I want to draw parallels between two species-wide disturbances which really shook faith in existing systems. Systems is a fine abstract word which covers all we may be interested in - from social and political to technical - and as John Gall has aptly articulated, we live immersed in a new age of faith, the age of 'faith in systems'. Both of these two disturbances which I concern myself with involve prodigious comets in the 'outer world' of humanity and automation in the inner. As schizos may attest, perhaps there indeed are celestial omens signalling planetary scale changes - but I am merely correlating and constructing an arguement about technology and the human condition, and (most certainly) not implying any cosmic causation. 

The two comets in question are: the great comet of 1811, and 3I/ATLAS which just passed around us. Their stories are written alongside the histories of radical technological displacement. The 1811 comet was shadowed by the beginning of industrial automation - which is a corollary to the beginning of the age of systems - mass production, mass education, mass media, "the masses" and politics that centered around "the worker", the pedestalisation of working in and for a system, and consuming "the slop" that linked status and identity with "the job" which provided for that consumption. Familial and even socio-biological fabrics needed to be "optimised" in pursuit of the vision of a wage (for its own sake) for all.  

Steam engines had turned man's pursuits from understanding nature to dominating it. Electricity connected these ambitions and pursuits like an interwebs of the yore. Transportation and logistics made possible global empires, and eventually world wars which set in motion technology developments - starring artificially intelligent automation in information societies - that challenge some of the sociopolitical realities which had become possible only by that original industrial revolution. This brings us to the second comet, 3I/ATLAS, which has foreshadowed the AI era. Like its erstwhile brethren, this outer world phenomena also has come at a time where "the worker" - the IT coolies this time - are going to be fighting for the billable hour, the "mythical man-month" as Brooks Jr. described it (and Indian IT sector lived it). Those who survive this will do so with increased monitoring and higher productivity demands. Also, because AI wraps up those hours of "work" in minutes - the impacted industries would shift to low-billing to protect their margins in a competitive sector.

But, there is a silverlining.

If androids could cry, will humans not?

This disruption forces a pivot. If the machines handle the execution, the only thing humans could do is high-level strategy and accountability of systems. There is a movement of value from "the hands" to "the architect" because strategy and logistics are much harder (and more situated) than production. Therefore in the disrupted markets, value will migrate, not dissapepear. If your job was just making some "digital thing" you are now in a race to the bottom on price. But at the same time a new intentionality layer emerges for those who who don't lay the bricks - but deeply understand and specify how the building should function in its environment. There will be some "labour-saving" - yes, but the silverlining (and the creative destruction) is here.

A recent benchmark cutting through the hype of "tool-level" code completion to assess real "agentic" software development affirms these ideas. Instead of asking the AIs to design a function or write some code, researchers game these an end-to-end software engineering brief - design the architecture, manage the dependencies, maintain conceptual consistency across files, and build a testable installable library from scratch. What this tells us is (and also why AGI is farther than the SamAltmans tell you) that long-horizon task planning is still a human domain as the AI starts losing the plot when the going starts getting tough (or a few thousand lines of code). Thus, below I frame and exemplify how "The Architect" emerges from decimation of the old world legacy labor.

Role	Old World Value [Task Based]	New World Value [Outcome Based]
Copywriter	Writing 50 taglines	Defining the psychological hook that makes a tagline work
Art Director	Spending 10 hours retouching an image	Synthesizing disparate cultural references into a formal visual language
Planner	Researching market data	Modelling consumer behavior for specific activities and creative "vibes"
Software Engineer	Writing boilerplate code and debugging	Designing the logic flow and guardrails for an AI agent to build the repo
Data Scientist	Cleaning datasets and manual feature engineering	Defining the hypothesis and "reward function" that makes data meaningful
Cyber Security Analyst	Sifting through logs and CTI to identify malicious objects	Threat modeling, anticipating sophisticated attacker psychology and hardening the system's behavioral logic

In the new digital-first world, intent and curation drive the new creations. Just as those two awesome comets passing through "our system" have altered our understanding of nature and reality, AI automation is also passing through our economies and workplaces - collapsing skill-gaps and stripping away the existing understandings, meaning and value of human labour. There is some tension here between the automatic and the autonomous - you don't want your stenographer setting your minute workflow and daily agenda - but you want the frictionless execution of same from them without the tool overstepping its bounds and making executive decisions, even minor ones, without agentic consent.