Whoever thinks the big money is in picks and shovels
Doesn't have a clue about gold mining or today's AI.
It has been repeated all over the place that to make money in the AI gold rush you need to be in the picks and shovels business. This a useless and dangerous metaphor in both AI and in mining.
The people who made the real money in mining were the people that owned, financed and operated the mines.
Added bit: (Thanks to Holger and LJ for the comments) I realise that most Americans when they think of gold mining, think of the 1849 California gold rush. And yes, the people who made money there were largely in the supply chain, rather than in mining (i.e Levi Strauss). There was a lot of rush, but not much gold.
When real diamonds or gold are found, in quantity, like in Kimberley and Johannesburg in the late 19th century, the people who made serious money weren’t selling shovels. They were buying land and digging big holes in it. Rhodes, Wernher, Joel, Cullinan, Phillips, Sauer, Eckstein, Robinson, Barnato, Beit, Oppenheimer, Rudd and others made massive bets on land, hired the best engineers, invented new mining and extraction techniques, designed innovative financing models, partnered, squabbled, amassed staggeringly huge fortunes and some of them changed labour supplies, redefined national borders, rewrote laws, developed new forms of human oppression and exploitation, overthrew governments and kingdoms, established monopolies, started wars, fought boxing matches, funded colonization, co-opted leading academic institutions, and developed some rather weird, grandiose theories about society, intelligence and culture. If you would like an introduction to the history of the Randlords, Geoffrey Wheatcroft’s book of the same name is brilliant, I’ve read it several times.
Can we stop with the picks and shovel metaphor, please. It is worse than useless.
The large AI companies are building the virtual equivalent of the mining industry. The people owning and running these companies, are for better or worse, the Randlords of today.
They are not mere purveyors of picks and shovels, they are shaping our very world. They are data mining, in a most visceral way. They have agency, and the responsibility comes with it. Their work creates good, and it creates harm, at scale. Calling them shovel-makers absolves them of that agency responsibility, and that does not serve society well.
In a 100 years I would hope that history judges today’s AI magnates more positively than it has done the Randlords, but hope is not a strategy. As a society we have not yet figured out AI responsibility, and I’m not sure that leaving it to the market or the wisdom and good nature of entrepreneurs is entirely wise, but neither is poorly enforced regulation. I’ve read widely and thought deeply, but I have no clear answers.
Joseph Weizenbaum is for me the greatest computer scientist, his work in early AI and computer ethics should be compulsory reading of anyone working in our field. We build stuff that has great power to do good and harm.
The computer programmer is a creator of universes for which he alone is the lawgiver. No playwright, no stage director, no emperor, however powerful, has ever exercised such absolute authority to arrange a stage or field of battle and to command such unswervingly dutiful actors or troops.
— Joseph Weizenbaum
I’ll leave you with some jazz this time. The one and and only Hugh Masekela’s song about the train bringing migrant workers from across Africa to the mines.
‘Hope is not a strategy.’ Kaboom!!
I always enjoy reading Work in Progress, this article is one of the best. Someone has to challenge general AI ignorance and shake the regulatory tree by its roots.