Career planning for the Singularity

Yet again, I attempt to ring the warning bell

Photo by julien Tromeur on Unsplash

Video summary

But you can't make people listen. They have to come round in their own time, wondering what happened and why the world blew up around them.
Ray Bradbury, Fahrenheit 451 (1953)

Top AI experts (Dario Amodei, Elon Musk, Sam Altman) all agree: In just 2-3 years, AI will surpass human capability in every computer-based task. Not in decades. Not in 10 years. In 2-3 years. That’s nothing… We’re no longer in the realm of sci-fi. We’re in the realm of short-term planning. Said differently, AI is a freight train heading straight for every knowledge worker's job and now is the time to react. The longer you wait, the harder it is to adapt when the real changes start happening.
Yet, most people aren’t responding—not really.
Michael Simmons, Blockbuster Blueprint with Michael Simmons, 19/2/25

Yuval Noah Harari, who wrote Sapiens, was asked if there was any book he would recommend that people could read in the present to understand the future that’s coming. He said no he couldn’t think of such a book because change was going to be so incredibly fast.
Douglas Murray, ARC speech, 19/2/25

Smart writers aren't competing with AI. They're "conducting" it… If you want to thrive in the next 3 years, stop trying to outwork the machines. Instead, learn to direct these powerful tools while focusing your creativity where it matters most.
Ed Gandia, 23/2/25

In mid-2020, I wrote an article warning my erstwhile legacy media colleagues that their industry was entering the final phase of its death spiral. This was not a popular observation to make in Fourth Estate circles. Unfortunately, it turned out to be an accurate one.

For the last three years, I have been warning my Professional-Managerial Class (PMC) confreres that if they persisted in their efforts to erase fundamental boundaries, such as those between man and woman and citizen and non-citizen, they would generate a tectonic backlash. (As an aside, it’s not a “reactionary” backlash, because it’s hardly reactionary to insist, no, there aren’t really 58 genders and, yes, “It's just obvious you can't have free immigration and a welfare state.")

These have been, to say the least, unpopular observations to make. At least in the PMC circles I once aspired to move in.

I share this trip down memory lane to establish that I have runs on the board when it comes to telling people things they very much don’t want to hear but would very much benefit from acknowledging.

Past performance is no guarantee of future returns. But given my track record, dear reader, you may like to fight your desire to dismiss out of hand what I’m about to try to convince you of.

Here it is: We are no longer in the realm of sci-fi. We are in the realm of short-term planning. AI is a freight train bearing down. The longer you remain in denial about this, the worse your position will be 6/12/18 months down the track.

In short, it appears the Singularity is nigh.

Singularity studies 101
To quote Wikipedia:

The technological singularity—or simply the singularity—is a hypothetical point in time at which technological growth becomes uncontrollable and irreversible, resulting in unforeseeable consequences for human civilization. According to the most popular version of the singularity hypothesis, I. J. Good's intelligence explosion model of 1965, an upgradable intelligent agent could eventually enter a positive feedback loop of successive self-improvement cycles; more intelligent generations would appear more and more rapidly, causing a rapid increase ("explosion") in intelligence which would culminate in a powerful superintelligence, far surpassing all human intelligence.

Tech-industry types have been aware of the Singularity for quite some time. I seem to remember them making humorous remarks about its imminent arrival in the relatively recent past. Of course, these remarks were only funny because everybody assumed it wasn’t happening until at least 2045 and quite possibly was never going to happen at all.

As Mr Simmons correctly notes, Altman, Amodei (Anthropic’s CEO) and Musk are now on the record saying AGI, or something very much like it, will be here by 2027. That’s sooner than the end of Trump’s current term.

My (purely intuitive) sense is that plain old Generative AI still has plenty of room to get much better before it runs into any serious constraints and that this type of AI will be more than sufficient to replace a big chunk of the workforce.

If the boffins make the leap from Generative Artificial Intelligence to Artificial General Intelligence (AGI), we will really be in Skynet achieving self-awareness territory.

As an aside, the Terminator franchise had this occurring in mid-1997. That may turn out to be out by precisely three decades. This is perhaps an appropriate juncture to note that AI is already being integrated into missile defence systems.

I’m not talking my own book
Unlike Mr Simmons, I have no AI-related wares to promote.

I’m not trying to plug any product or service. If I were concerned with furthering my economic self-interest, I’d be frantically pumping out LinkedIn posts insisting that AI is overhyped and that it will never ever replace human workers, especially the creative ones.*

But after having failed to accept my own industry (print journalism) was in a state of irreversible collapse back when I first should have, I hate to see others making a similar mistake.

It's possible that all those tech-industry heavy hitters who are predicting substantial, short-term advancements in AI are wrong.

It’s possible that the hundreds of billions that have been ploughed into AI since the first iteration of ChatGPT launched in late 2022 won’t result in significant advances and that AI technology is about to plateau for one reason or another.

It’s possible that the recent Deep Seek breakthrough, which demonstrated it was possible to bring the (previously daunting) costs of providing a high-performing AI way down, is a sinister CCP psyop.

It’s possible those AI agents that shareholders and C-suiters are salivating about ‘augmenting’ their human workforces with will turn out to be incompetent.

All those things are entirely possible.

They are not likely.

Gradually then suddenly
Just as a thought experiment, think about how the world, and your own little world, changes if AI starts making broad swathes of the workforce redundant.

Struggling to imagine such a scenario? Here’s Simmons – admittedly, a not entirely disinterested party – sounding the warning bell:

Many top AI experts are lying by omission. They are penning thought pieces of all of the utopian scenarios where AI lowers the price of everything and discovers new science. But here’s what they’re not talking about. They are not talking about what will happen when 1B+ knowledge workers are unemployed because AI is significantly faster, cheaper, and better. They aren’t talking about the fact that governments around the world are so in debt that they can’t afford universal basic income.

Over the next 10 years, I imagine AI will render almost all of humanity redundant. In the next five years, I’m confident those who recognise the world is being turned upside down and who adapt accordingly will (a) remain employable/hireable far longer than their denialist peers and (b) potentially be able to leverage AI to turbocharge their productivity and earnings before they too end up applying for whatever UBI governments provide for their now non-working citizens.

(Unlike Simmons, I assume a UBI will need to be introduced to prevent flaming-torch-wielding mobs of would-be Luigi Mangiones coming after the alpha tech bros, but I wouldn’t count on it being particularly generous.)

After scaring the shit out of his potential customers, Mr Simmons does throw them the following bone – “In the next 2-3 years, there will be an opportunity for proactive knowledge workers to 10x and even 100x their productivity.”

On that point, I’ll think about how ‘early acknowledgers’ can now begin to maximise their employability and earnings and share my thoughts next week.

*I’ve read many such articles from my colleagues over the last couple of years and have failed to be convinced by any of them.