Above: Illustration made using the Flux AI image generator on 123rf.com

BitDepth#1528 for September 15, 2025

The technology singularity postulates that at some point, intelligent machines would begin creating even more intelligent machines in an explosion of advancement that would quickly leave human understanding behind.

The British mathematician IJ Good considered this all the way back in 1965, when he wrote: “The first ultra-intelligent machine is the last invention that man need ever make, provided that the machine is docile enough to tell us how to keep it under control.”

Good was writing when distributed networks of connected information were an abstraction, so he could not have known that such a machine, having achieved super-intelligence, would emerge knowing everything that humanity had ever done, to itself, to its neighbours and to the world.

Such a machine might conclude, in that nanosecond of awareness, that peeling the crust off planet Earth and tossing all its inhabitants into the sun like a bedsheet teeming with selfish, self-important bedbugs might be a good first step.

Instead, we have lived with the march of the machines Skynet scenario all these years, little suspecting that the singularity would take place at machine speed.

Before then; however, the world grappling with AI is likely to face a dramatic reorganisation of jobs. Many of the earliest speculation about which automated jobs will require human oversight seem ambitious.

Interns aren’t disposable napkins, they are the entry point for a new generation of leaders and visionaries, so simply dismissing all entry-level jobs would be a mistake.

Robert Reich, in an August 19 post to his Substack, pointed out that of the three broad classes of jobs, making, thinking, and caring, the ones that are likely to survive will be those that are driven by thinking and caring.

Making jobs are particularly vulnerable to automation and offshoring, moving to where the cost of labour is either cheap or can be reduced to closer to zero through increased automation.

Thinking jobs will continue to exist, but will require a shrinking pool of meat operators as ever more operations are ceded to fast AI control.

That’s already happening in cybersecurity operations, as hackers power their invasive tools with AI with response times that demand AI level blocks and countermeasures.

According to Crowdstrike, a threat-detection company, bad-actor AI will “Identify vulnerabilities, deploy campaigns along identified attack vectors, advance attack paths, establish backdoors within systems, exfiltrate or tamper with data, and interfere with system operations. AI-powered cyberattacks can learn and evolve and can adapt to avoid detection or create a pattern of attack that a security system can’t detect.”

Crowdstrike develops AI tools to respond to these next-level threats.

The extent to which thinking jobs will survive will depend on the personal acuity of practitioners and the identifiably human approach they can bring to that work.

Clearer is the potential for caring jobs, child-care and elder-care, jobs in the disability sector that require an experienced, intuitive response to nonstandard subjects.

While the standard of the human resource dedicated to such care is generally poor today, saturating the market with talent does not require an army of new recruits. There are people who are temperamentally unsuitable to any kind of sustained caregiving.

A saturated market means generally poor pay for those employed in the sector, which does not bode well for the quality of talent it will eventually attract.

The ideas of making, thinking and caring work does suggest specific sectors of the economy, but an economy fragmented by AI will be a fundamentally different one.

Journalism, for instance, in a world of ready availability of raw facts will be defined by practitioners who can both present a perspective and demonstrate both authority and trust in increasingly narrow niches.

Authors like Reich who are flexing Substack’s reach and discovery potential will thrive through the identifiable uniqueness of their work, not on cookie-cutter contributions to the thought pool.

The “care” business will also demand trust, as care will likely move beyond concepts of medical care into art, pet-care, household management, landscaping and areas of personal adult satisfaction, with only the best rising to the top of those professions.

There will, eventually, be more workers than there is work to be done, and that creates a new conundrum for human society.

Neither the capitalist nor the socialist forms of governance ever envisioned societies which no longer needed to build or advance, in which small sectors of the population accumulate vast wealth, leaving salt for the poor.

That won’t be an easy world to manage, if it comes to pass.