“So there I am, nine o’clock that night, still annoyed, blank page and a very strong urge to write the kind of letter you regret by the morning.”
The Wrong Door
The fear is real. You’re just facing the wrong way.
In the US, roughly half of adults now use AI chatbots regularly. Only 16%think AI will be good for society. That’s not a technology adoption curve — it’s a trust collapse happening at the same time as everyone signs up.
FOBO — the fear of being obsolete— is a legitimate fear. But this episode argues it’s aimed at the wrong door. The door people are watching is instant, total replacement. The door that’s actually opening is quieter: entry-level hiring frozen, graduate roles cut, the bottom rungs of the ladder sawn off while everyone stares at the top of the stairs.
The two waves
— vs just 16% who think it’ll be good for society
Pew Research Center’s most recent national survey (fielded Feb 17–23, 2026, published June 17, 2026) found that about half of US adults now use AI chatbots — up sharply from a third in 2024. But belief that AI is good for society sits at just 16%. Usage and trust are moving in opposite directions.
Entry-level roles and graduate hiring are contracting in the US even as the broader economy holds up — boosted, in part, by AI-driven stock market highs. Stephen shares the story of a friend in California, thirty-plus years in IT, whose contract ends this month as his employer redirects budget toward AI investment.
The story his employer is telling investors — “we’re investing in AI” — is a different story than the one being told to the people losing their jobs.
up from 35% in 2025
The British Chambers of Commerce’s March 2026 report (with Atos), Powering Productivity: AI and the Future of UK Work, found AI adoption among UK firms has grown fast — but it’s the small group going beyond generic tools (ChatGPT, Copilot) into bespoke integration where the job cuts are actually showing up.
About one in ten firms have moved into that deeper, bespoke integration — and of those, roughly one in five report AI-attributable staffing cuts.
using AI for 12+ months
Morgan Stanley’s January 2026 analysis of firms using AI for at least a year found the UK’s net job loss rate — 8% — is the highest among major economies studied (also covering the US, Germany, Japan and Australia), roughly double the international average.
UK firms saw productivity gains similar to their US counterparts, but unlike US firms, didn’t turn those gains into net new hiring.
move fast enough to cause civil unrest
A King’s College London Policy Institute survey (fielded April 2026, published mid-May 2026) found that university students — the group using AI most heavily — are also the most pessimistic about where it leads: 34% think the job losses could move fast enough to cause civil unrest, against 22% of the general public.
More than half of the students surveyed believe the resulting job losses would be worse than a normal recession.
The reframe
The door everyone’s watching is the one where AI rolls out at your organisation and you’re simply done. That’s not how it’s played out so far, in the US or the UK. What’s actually happening is quieter and slower: a hiring chill. Entry-level jobs frozen. Graduate roles cut. If you’re early in your career, that’s the door that’s actually closing — not because a robot took your desk, but because you never got a foot on the ladder in the first place.
Further up the ladder, the story is different again: less about replacement, more about burnout. The mundane, recoverable parts of the job — the shuffling, the collecting, the admin — are getting automated. What’s left is constant critical thinking, with none of the built-in recovery time the boring bits used to provide.
Left is the fear everyone is watching.Right is the door the data actually shows opening. The hatched cells aren’t missing data — the loud fear simply has no real-world column to fill.
Start where you’re not afraid
Lauren’s partner, Billy Lyme — a self-taught artist with an upcoming gallery placement as a stockroom artist — spent a couple of hours with a Claude skill originally built for personal branding. It helped her articulate who she is as an artist, reviewed whether her website’s language and colour choices actually matched her philosophy, and acted, in Lauren’s words, like a coach.
“This was mind blowing. I didn’t expect this at all.”
Stephen used the same skill on his own personal brand and found it dug into where he wanted to go and why — including calling out things he’d built that were, in his words, “the antithesis” of what he was trying to achieve.
Available at his own pace, at 9pm after watching TV if that’s when he was ready for it.
Stephen’s assistant coach on his daughter’s under-14s team — no professional coding background — vibe-coded (using Codex) a small app to help manage game-day substitutions: tracking who’s played, what position, and giving a fair rotation across the bench.
Built in a couple of nights of casual chatting with the tool. Something that, as Stephen puts it, “would never have gotten built in the past” — there just wasn’t the time.
Lauren used Claude Cowork — prompted to try the Claude Chrome extension — to pull together account summaries and expenses for a financial review that would normally have taken weeks. “What it could do in about five minutes was extraordinary.”
Her caveat: know your own risk tolerance before granting that kind of access, and be deliberate about what you let a tool see.
Back to where the episode started: the 9pm, still-furious letter about a badly refereed under-14s match. Stephen used AI to turn the angry first draft into something measured, firm and fair — pulling in the relevant rules, stripping out the emotion, keeping the facts.
It worked: a proper referee is being arranged for the next game.
You can’t think your way out
This isn’t about being obsolete tomorrow. At least not here, not yet.
You can’t think your way out of FOBO. You can only do your way out of it.
Start outside work, where the stakes are low and the upside is real.
A brand reflection. A finance clean-up. A substitution app for Sunday-morning football.
None of it needs the paid tier — the free tier covers everything in this episode.
— Stephen & Lauren