In What World Is It OK to Pay a Twenty-Something Employee a $100 Million Salary Package?

(image:ideogram.ai)

Early last week, the technology world was startled to learn that Meta had poached a few AI researchers from OpenAI with first-year compensation packages of up to $100 million. These were more than salaries - they were like signing bonuses, because they are reputedly made up in part by stock options with early vesting periods, meaning that the new hires can turn them into cash within the first year if they choose.

The names of the AI researchers are Shengjia Zhao, Shuchao Bi, Jiahui Yu, Hongyu Ren, Lucas Beyer, Alexander Kolesnikov, Xiaohua Zhai, Trapit Bansa. By the end of the week, this new cohort at Meta had grown to 12, poached mostly from OpenAI, but some from other prestigious AI companies.

Who are these enviable beneficiaries, all still in their twenties and thirties? It turns out that, while they are all highly respected researchers in various AI specialties, they are not even on the very top rung of the AI researcher celebrity ladder. They are simply part of a tight cadre of smart PhDs from prestigious universities like Stanford and MIT who have published impressive papers and who have toiled in important areas of AI development like language and vision and brand-name large-language models.

The Wall Street Journal reported that they were recruited from 'The List', a highly secret compilation of names assembled by Zuckerberg and his team, a veritable trove of the highest rated AI researchers in the world. Zuckerberg personally reached out to some of the researchers, inviting them to his various homes, in a charm offensive which seems to have paid off.

Apologies for repeating myself, but that's $100 million in the first year for some of these researchers. Up to $300 million by the end of the fourth year.

There are many strands to this story - from the reaction of OpenAI executives (really pissed off) to the flurry within the industry to re-adjust top AI researcher salaries and perks (accompanied by the whooshing sounds of top researchers re-polishing their CVs), to the question of why Meta did this (their AI offerings do not have an ‘edge’ above others). But there is another matter, which is perhaps more interesting, at least from a sociological perspective. A system that rewards a single individual to that extent, at least compared with the average worker’s salary (even that of a well paid tech worker), seems to me to be out of whack on some profound human metric.

There are few people for whom this news was unsurprising. We are used to seeing massive pay packages for some CEOs. Most of us tut-tut at the numbers and wonder whether a particularly well paid CEO is worth the money, whether their company made the sort of returns to justify it, whether the gap between their pay and that of others in the company who contributed is not too wide.

There is a well-trodden field of remuneration science that ponders these matters. Cogent arguments about the structuring of executive packages have been made. The same applies to sports stars, actors and other celebrities. An individual's remunerative worth is usually carefully considered, at least within the boundaries of a given system - in this case, the US brand of capitalism. When a highly paid superstar loses their sheen, they are usually bounced out quickly - and sometimes ignominiously. But they can generally afford to salve their wounds with their bounties.

In any event, we are used to this sort of thing, even if we sometimes raise an eyebrow. But here we are talking about some unknown techies a few years out of college who have no particular skill - other than genius. Genius is in short supply, of course - that is its defining ingredient - but there are many unsung geniuses out there whose financial rewards are unremarkable. They exist everywhere: teachers, plumbers, interior designers, poets, chefs, game rangers. Being a genius is not a sufficient condition for $100 million in the early years of your career. Other matters must intrude.

The other matters are scarcity and perceived value.

The field in which these few geniuses operate is, of course, AI. It’s a new field, requiring all manner of smarts, from computer science to maths to stats to linguistics, as well as out-of-the-box thinking about how we think, reason, create, remember. There are not many geniuses in this AI world yet because it has only been around for a blink of an eye. AI geniuses are scarce in a way that other geniuses are not.

Then there is perceived value. Those who peer into the future and provide the capital to get us there are utterly convinced that AI represents one of the single most important developments in human history, if not the most important. For them, AI's promise is that it will be the greatest accelerator of human development ever, and in the shortest amount of time. Their belief borders on religious zealotry.

We read constantly of AI hallucinations and job losses and inauthenticity and ethical concerns and the danger of bad actors, but for those with billions to spend these issues are irrelevant. For them, these are merely engineering problems, small matters, minor hiccups on the way to human upliftment and general abundance.

Spending $100 million on a first-year salary package for a rare genius in this time and place?

The recruiters clearly think it’s a bargain.

Others may shrug and mutter something about willing buyers and willing sellers.

But, for some of us, it feels like the concept of human value is being rewritten and debased.

Steven Boykey Sidley is a professor of practice at JBS, University of Johannesburg and a partner at Bridge Capital and a columnist-at-large at Daily Maverick. His new book "It's Mine: How the Crypto Industry is Redefining Ownership" is published by Maverick451 in SA and Legend Times Group in UK/EU, available now. Copy edit by Bryony Mortimer.

---