TL;DR
Happy Indigenous People's Day / Columbus Day / 3 Day Weekend for! I hope that you have the day off from work.
It’s been quiet weekend, so here’s a roundup of four stories:
Software developer job postings are rising even as AI writes more code, though the gains are going to senior engineers.
Palantir’s CEO says only trade workers and neurodivergent thinkers are safe.
Amazon is reportedly shopping for a $7B AI startup.
And a major literary prize faces questions about whether its winner was written by a machine.
WHAT TO KNOW
THE WEEKEND IN 4 STORIES
1. AI is writing more code, so why are developer jobs coming back?
US software developer job postings have risen since February 2025, when Anthropic launched Claude Code, even though overall job postings fell. That cuts against the prediction that coding jobs would vanish first. The catch is who’s getting hired: 71% of the increase from May 2025 to May 2026 came from senior roles, and 37% from AI-titled jobs. Meanwhile, employment for younger developers has fallen. The government still projects growth: software developer employment rising 10% from 2025 to 2035, from about 1.72 million to 1.89 million jobs. (CNBC)
How worried?: 🟢 2/5 — Worth Watching (🟡 3/5 if you’re entry-level)
2. Palantir’s CEO names the two kinds of people who’ll be fine
Alex Karp says the only sure signs you have a future in the AI era are vocational training or being neurodivergent. His argument is that trades are hard to automate, and people who think differently will do the original work AI can’t. Palantir has put their money where their mouth it; its fellowship for high school grads who skip college pays a $5,400 monthly stipend (and top performers can get full-time offers at Plantir).
Not everyone agrees though. Anthropic cofounder Daniela Amodei has argued back in February of this year that humanities study will matter more than ever. Time will tell who is correct. It also might be a blend of both mindsets. (Fortune)
How worried?: 🟢 2/5 — Worth Watching
3. Amazon in talks to buy AI startup Decart for about $7 billion
Decart is an Israeli startup whose software claims to make AI models run up to eight times faster on the same hardware, including Amazon’s own Trainium chips. The deal would be a big step up from the $4 billion valuation Decart raised at in May.
Anthropic was in talks first, but walked away from a $6 billion deal after due diligence. The bigger signal is that big tech is now paying billions to make AI cheaper to run, not just smarter. (Wall Street Journal)
How worried?: 🟢 2/5 — Worth Watching (While this is an efficiency deal (and not a direct threat),cheaper AI will speed up adoption.)
4. Did AI write a prize-winning short story?
Zimbabwean writer Simbarashe Steyn Kundizeza won the £10,000 2026 Caine Prize for African Writing for “The Blue Fighter.” The Atlantic now argues the story shows signs of being AI-written.
No finding has been made, and this is the second prize controversy this year. In May, a Commonwealth Short Story Prize winner was flagged by readers and AI detectors, and the publisher said the truth may never be known. The real problem is that today’s AI detectors aren’t reliable and can cast suspicion on innocent writers. (The Atlantic)
How worried?: 🟡 3/5 — Meaningful
OUR TAKE
The jobs story this week is more hopeful than the headlines usually are: AI is creating demand for people who can direct it. But the pattern from last week’s McKinsey report holds. The gains go to people who already have experience, and the squeeze lands on people just starting out.
Meanwhile, the Caine Prize story shows a different kind of risk: when we can’t tell human work from AI work, every writer gets less trust.
How worried?: 🟢 2/5 — Worth Watching
WHAT TO WATCH?
So what should you keep an eye on?
Junior developer hiring. If senior and AI roles keep growing while entry-level roles keep shrinking, the industry may run short of experienced engineers in five years.
How worried?: 🟡 3/5 — Meaningful
WHAT DO YOU THINK?
Which path looks safest in the AI era?
Hit reply: are we too worried, or not worried enough? We read every response.