The Future of Work Is Getting Out of the Way — Julien Reszka

← Blog

The future of work is not longer hours. It is learning when to step back, and designing systems so that when humans interfere, the damage is limited.

-65% Reduction in Swedish road fatalities since Vision Zero launched in 1997 Swedish Transport Agency

The coming shift in work is not about working more, it is about working less in the right way, staying out of the path of autonomous agents long enough for them to be useful, while taking on a different kind of responsibility: designing systems that assume human error and contain its consequences. Sweden's Vision Zero road safety programme is the clearest model for this: rather than expecting drivers to be perfect, engineers redesigned roads so that mistakes could not be fatal:

roundabouts instead of intersections

barriers between lanes

speed limits set by physics rather than optimism

AWS formalised the same logic for cloud infrastructure with its shared responsibility model, which draws a hard line between what the platform guarantees and what the user must handle, so neither side is surprised when something breaks. The worker of the future is less an executor and more a system designer who asks, before anything goes wrong, how bad can this get if someone does the wrong thing, and then makes the answer as small as possible.

Myth: The future of work means working harder alongside smarter machinesSwedish Transport Administration, Vision Zero

For any process you own, ask: if someone does the worst plausible thing, how bad does it get? Then redesign so the answer is recoverable. That question is your new job description.

Post on X

Discussion

Are you trying to figure out how to design your processes so that human mistakes are contained rather than catastrophic?

Post on X

Olga K. Warsaw, Poland 2026-04-02

Yes. We've been trying to fix human errors with more training and it never sticks. Redesigning the process so the bad outcome is impossible is a completely different conversation.

Reply

Will T. 2026-04-03

Yes. Trying to draw the shared-responsibility line between our team and the autonomous agent we deployed last month. Clarity on who owns what when it breaks is missing and it's hurting us.

Reply

Olga K. 2026-04-03

Write the line down explicitly. The act of writing it surfaces every gap.

Reply

Sana O. Seoul, South Korea 2026-05-31

Hard agree on the second half. The hardest part of my week is figuring out what to do, not doing it. AI tools that help me decide what's worth doing would matter ten times more than tools that help me do it faster.

Reply

Marco R. Milan, Italy 2026-04-04

The Vision Zero analogy is imperfect in one key way: road engineers can enforce physical constraints. You literally cannot drive fast through a roundabout. Software has no equivalent: any guardrail I design can be bypassed by someone with deploy access. The damage-containment goal is right, but the forcing functions available in software are much weaker than in road design.

Reply

Your name *

Your email * (never published)

City, country (optional)

Tell me about your situation *

Send<br>All comments are manually moderated by the author.

Subscribe to get new posts by email →

See also

RSS Is Back. AI Agents Are Reading It.

Less Work, More Worry

Automation Takes the Easy Jobs and Leaves You the Hard Ones

Browse all figures →<br>Browse all myths →