on ai, panic, and meaningful work

artificial intelligence arrived in a rush. models shipped overnight and the headlines promised displacement, but the emotion most builders actually felt was a quieter one: whether the work would still be there.

strip away the gold-rush noise and the pattern turns out to be older than the panic. keynes, writing in 1930, expected technological acceleration to absorb repetitive labour and hand the time back [1]. brynjolfsson and mcafee found much the same nearly a century later — automation takes over routine cognitive work and leaves judgment where it was [2].

when software takes over the boilerplate, the repetitive syntax and the log parsing, what it removes is friction. marcus aurelius made the point two thousand years earlier: most of what fills a day is not essential, and cutting it leaves you doing fewer things better [3].

which leaves the harder question of what the time is for.