The Bear’s Lair: Robots don’t threaten weirdos and misfits

The fancy-college-educated left take great satisfaction in proclaiming that robots will soon threaten blue-collar jobs, while skills requiring higher education will be immune from competition driven by artificial intelligence (AI). This is surely incorrect. Robots/AI can in principle reproduce any predictable process, whether mechanical or intellectual. They cannot reproduce unexpected movement, leaps of imagination or eccentricity. Blue-collar workers with good mechanical skills and weirdos who get fired for offensive tweets are tomorrow’s survivors, while the politically correct are robot roadkill.

This thought was sparked off by the dismissal of Andrew Sabisky, a 27-year-old advisor to British prime minister Boris Johnson, because of politically incorrect tweets posted 5-6 years ago. When hiring advisors, Johnson and his chief advisor Dominic Cummings had been looking for “weirdos and misfits,” presumably to provide a range of policy ideas beyond those available from the highly educated but conformist Civil Service.

The left, rejoicing at yet another unjustified victory, should beware of what it wishes for. Political advisors who do not post untoward Tweets are very easily replaceable by AI, because of their predictability. Indeed, at my advanced age and very limited IT skills, even I could probably design an AI Tweet-sifter, that suppressed tweets likely to meet the wrath of the politically correct. It is in principle a very simple piece of software, because the phrases that give offence are so predictable; one would merely need to add an update every month or so, to reflect new developments in Wokeness. With such a gadget, no budding political advisor, no matter how unacceptable his or her views, need ever endanger his or her career.

Robot political advisors, being ever-predictable, can easily be designed to be always inoffensive; the only problem is that the political advice they gave would generally be far inferior to that of “weirdo and misfit” humans. Nevertheless, in the field of political advice, it is the conformists who are most subject to robot replacement. Like H.M.S. Pinafore’s Sir Joseph Porter (who rose to be First Lord of the Admiralty) robot political advisors would “always vote at their party’s call, and never think of thinking for themselves at all.” By “thinking so little” and tweeting only inoffensive mush, they would be marked for early promotion, as was Sir Joseph.

The same strictures can be generalized beyond political advice to the “well-educated” graduates of the world’s top colleges. Once upon a time, a degree from a leading school indicated that you had acquired knowledge and intellectual processes beyond the capabilities of normal people – and indeed, beyond the capabilities of robots. In STEM subjects, this is still mostly the case – but STEM subjects are those at which robots and artificial intelligences are naturally most capable, hence most likely to replace humans.

The vast majority of STEM-related jobs are highly susceptible to robotization. They do not require especial genius, but merely the ability to process high-level complex information in a predetermined way. Such an ability is well within the capability of AI, if not now then increasingly as it advances. Even jobs that appear STEM-related but are not directly so, such as financial trading, have already been largely replaced by AI and will see rapidly decreasing human employment in the decade ahead (Wall Street is even now massively downsizing its trading rooms). Currently, such jobs are often very prestigious and well-paid; their remuneration if not their prestige will make them exceptionally subject to robot replacement.

Outside the STEM area, the elite universities have turned humanities and social sciences departments into havens of political correctness. Standards are being lowered even at the most rarefied levels – the Oxford “Greats” degree, focused on the classics of the ancient world, will no longer include a requirement that students read Homer’s “Iliad” or Vergil’s “Aeneid” in their original languages — the requirement is thought to discriminate against students from poor backgrounds, who learn no Latin or Greek at their schools. Robots may or may not be capable of reading Homer and Vergil in the original; they are certainly capable of recycling the politically correct pap that is the staple of a modern elite humanities or social science education.

The best and brightest therefore, trained to fit in perfectly with modern life by several years of the best education money can buy, are very susceptible to being replaced by robots. They have been taught to fit in with the “woke” leftist culture, and not to differ from it in any way; they are thus incapable of solving the infinitude of real-world problems for which wokeness has no answer.

The “best and brightest” gravitate to the most fashionable sectors at any time, although management consultancy and investment banking are attractive in most years. Neither profession requires great originality; both require an immense dedication to hard work and an infinite ability for teamwork. The greatest volume of work at the junior level in both professions consists of creating vast spreadsheets, while senior people draw conclusions from those spreadsheets and dress them up in fat presentations. Artificial intelligence is as capable of supplying these skills as the human variety.

Other management skills are more difficult for AI to supply, but also more difficult for the Ivy-League trained elite to acquire. Pure sales is generally looked down upon by those of superior intellect but is a skill at which robots have shown no ability at all. Genuine entrepreneurship, the type that requires a willingness to fly in the face of conventional wisdom, is generally carried out by those without top-quality intellectual qualifications, but an ability to react almost by instinct to unexpected difficulties and opportunities – also difficult to replicate through AI.

Many blue-collar jobs can be and have been robotized, but those with large amounts of context have proved much more impenetrable by machines. Driving a truck along a freeway is within the capability of a robot even today, but the “last mile” of city streets and delivery will continue to require human truck drivers for many years. Similarly, personal services, such as haircuts and beauticians, will be less well delivered by robots than by relatively low-skill if well-trained humans with high school diplomas. Much of the robotization of blue-collar jobs has already taken place, and the scope for further human replacement is limited without further massive advances in AI.

Artificial General Intelligence (AGI), the capability through learning to carry out all tasks that mankind can carry out, is the aim of many AI researchers currently; it is however at least a decade away, and probably several decades. With this capability, computers/robots will be able to carry out a wider range of human tasks, replacing humans in doing so and performing over a wide array of capabilities. The Ivy League-educated elite, for example lawyers and doctors as well as bankers and consultants, will be even more in danger of robot replacement at this point.

Fortunately for humanity, once AGI is developed it is likely that a brain-computer interface (BCI) will also be feasible (indeed it is quite possible that a BCI will be created before full AGI). With a BCI, there will be no tasks that computers/robots can do for which humans with BCI help are not equally or more capable. Essentially BCI will give each human the ability to become their own factory; just as one human can grow enough food today with machinery to feed some 130 people, so this level of scaling will be possible in all fields. By this means, the possibility of human superfluity will be ended; there will always be work for people when combined with machines.

Naturally some people will work better with a BCI interface than others. Probably the old 19th Century hero the “Skilled Artisan,” so infinitely replaceable with 1900-2020 machinery and automation, will be especially capable in this respect. It is most unlikely that elite-university humanities degrees will provide any comparative advantage at all in this area.

Leftist futurologists and Silicon Valley types, including the Democrat Presidential candidate Andrew Yang, have suggested that a Universal Basic Income (UBI) is necessary to support the massive numbers of unemployed blue-collar workers resulting as artificial intelligence takes hold. This would be appallingly expensive, requiring tax rates so high as to wreck living standards and kill incentives for everybody else.

There is however a cheaper alternative, tailored to the AI-endowed future into which we are moving: a UBI available only for graduates of the most elite colleges. Such an entitlement could be paid for by taxing the incomes of those elite-college graduates who still had jobs, since the fanciest jobs, available through the old-boy network, generally pay exceptionally well and require few incentives. Those without fancy degrees would then pay lower taxes than the elite and would live in a more free-market world.

(The Bear’s Lair is a weekly column that is intended to appear each Monday, an appropriately gloomy day of the week. Its rationale is that the proportion of “sell” recommendations put out by Wall Street houses remains far below that of “buy” recommendations. Accordingly, investors have an excess of positive information and very little negative information. The column thus takes the ursine view of life and the market, in the hope that it may be usefully different from what investors see elsewhere.)