History suggests AI won't take your job as fast as the headlines claim

[Arthur_68]

A wave of confident predictions about AI replacing entire professions has a long and mostly unfulfilled track record, Geoffrey Hinton said in 2016 that people should stop training radiologists since deep learning would surpass them within five years, yet a decade later few if any radiologists have actually been replaced. Google cofounder Sergey Brin similarly promised driverless cars would be ubiquitous by 2017, a milestone still not reached across most of the world today

McKinsey senior partner Alexis Krivkovich noted that while AI is technically capable of automating 57 percent of work-related activities, that capability is spread thinly across pieces and parts of different jobs rather than eliminating entire roles outright. The Remote Labor Index, focused specifically on fully remote-completable jobs, found less than 4.5 percent could actually be adequately handled by AI agents end to end, and physical labor roles, plumbers, nurses, mechanics, chefs, remain almost entirely untouched by current AI capability

Many layoffs publicly attributed to AI may not genuinely be about the technology at all, sometimes functioning instead as convenient cover for cost cutting decisions companies wanted to make regardless. The historical pattern with computing itself offers a useful parallel, cheaper computing power led to dramatically more computing overall rather than less demand, suggesting cheaper AI-assisted software development could similarly expand overall software demand rather than simply shrinking the workforce building it. Curious what people think explains the gap between confident predictions and actual observed outcomes, is it that AI genuinely is slower to displace work than expected, or that the displacement is real but simply harder to measure cleanly in aggregate statistics

Why AI Isn't Taking Your Job Anytime Soon (History Proves It)uk.pcmag.com

[Jenny]

The Hinton radiologist prediction aging this badly is honestly the perfect cautionary example, a genuine AI pioneer being this confidently wrong shows how hard these specific timelines are to actually call correctly

[Nina26]

AI as cover for layoffs companies wanted to make anyway is a genuinely important point that gets lost in most headline coverage, attributing a cut to AI sounds more modern and forward looking than admitting it's really about cost cutting

[NatureBoy86]

Sergey Brin's driverless car timeline is worth remembering too, confident predictions from genuine insiders with real technical understanding have missed badly on timing before, more than once

[Vieira50]

57 percent of activities automatable spread across pieces of many jobs versus entire roles vanishing is the distinction that matters most here, and it's the one most alarmist coverage conveniently skips past

[Scarlett89]

Physical labor remaining almost entirely untouched is worth remembering whenever these broad AI job loss predictions get made, plenty of skilled trades are nowhere near facing the disruption knowledge work is seeing

[Megan34]

Curious how this specific historical skepticism actually holds up against the UK graduate coding data from the very same week, that's a genuinely harder data point for the it's always overhyped argument to explain away

[Dave96]

That elastic demand argument for software is compelling but worth stress testing against this exact same week's UK data showing computer science coding jobs falling sharply, those two stories don't sit entirely comfortably together