What I got right and wrong in 2017
In 2017 I wrote my first ever post. I said automation was coming much quicker than we thought. Nine years on, I thought I'd mark my own homework.
What I got right
Ai in hiring. The recruiters in those comments were sure a machine couldn't do their job. I wasn't so sure. In a 2025 Resume.org survey, 57% of companies said they already use Ai in hiring. Some now let an Ai agent run the first-round interview. That's further than anyone in those comments expected.
Driverless taxis. I mentioned self-driving Ubers "in limited form". Waymo now runs robotaxis in 14 US cities. Not in Manchester yet, sadly.
Driverless lorries. I guessed automated lorries were on the way. In May 2025, Aurora started hauling paid freight between Dallas and Houston with nobody in the cab.
Working from the car. I got excited about Volvo putting Skype for Business in the 90 series. In 2025 Mercedes put Teams video calls into the new CLA, with Microsoft 365 Copilot on board too. (Whether anyone wants meetings in the car is a different question.)
What I got wrong
Uber building its own driverless cars. Uber sold its self-driving unit to Aurora in 2020 and partnered with Waymo instead. Now Waymo plans to launch its own app in the two cities they share.
Checkout-free shops. I listed the Amazon store with no tills as the future of shopping. In January 2026, Amazon closed almost all its Go stores. The tech lives on in over 360 other shops, so half a point maybe.
Insurance disappearing. I said car insurance would become almost redundant. Most cars are still driven by people, and my renewal still arrives every year.
Driving lessons. I said I'd never have to pay for my daughter's. Jury's still out on that one. She's got a few years yet.
What I missed completely
In 2017 I was thinking about robots, tills and cars. I didn't picture Ai that writes, codes and talks back.
That's the bit I work with every day now. I build tools with it, I write specs with it, and it helped me check the facts in this post.
My 2017 self also asked whether new jobs would replace the ones automation takes. Nine years later, I still don't have a full answer to that one.