Guesswork | Vol. 2 |
Guess & Click presents — your weekly click into AI
40 tabs. So you don’t have to. |
Every week I hand this issue to a second model and make it tell me what I got wrong. This week a court, a regulator and a stock market all did the same thing to the AI industry at once. |
The one that matters |
A US-trained AI just lost in a German court On July 31, a Munich court largely upheld GEMA's copyright claims against AI music service Suno — ruling that training in the US doesn't put a model out of reach of EU law, and that liability sits with the provider, not the user. The ruling isn't final and can be appealed. Why you should care — the popular assumption was that where a model got trained decides whose rules apply. This court's answer: no. If that reading holds up on appeal, “we trained it offshore” stops being a shield, and the company shipping the tool — not you, the person using it — carries the liability. That's the version creators have been asking for. |
The rundown |
The EU made disclosure the law, not the vibe. Anthropic put a dial on intelligence. “Open” is not the same as “runnable”. |
Who's paying for all this |
The free tier has a price tag, and a market just felt it Google, Microsoft, Meta and Amazon are projected to spend a combined $725 billion on AI infrastructure this year — up 77% on last year. Per-company breakdowns are circulating, but the source doesn't confirm them, so neither will we. Then the other side of the trade. South Korea's KOSPI fell as much as 12.6% intraday on July 29, closing down 6% after a near-11% drop the day before — roughly 40% off a peak reached barely a month earlier, about $2.18 trillion erased in two days. Analysts blamed cooling interest in AI chipmakers, amplified by single-stock ETFs built to multiply the daily move. Why you should care — that spending is who pays for your free tier. The Seoul slide is what it looks like when investors start asking when it pays back. The AI trade isn't a tech-blog abstraction anymore — it's big enough to move a country's stock market, in both directions. |
This one's about your job |
AI is now cited in roughly a quarter of US job cuts Challenger, Gray & Christmas reports AI has been cited in 101,743 US job cut announcements so far in 2026 — about 23% of all cuts — and June was the fourth straight month AI led all stated reasons. “Cited” is doing real work in that sentence: it's the reason companies give, not a proven cause. That number is real, and it's rising. It's also exactly why the next bit matters. |
You are not behind |
Pew found 49% of US adults use AI chatbots, up from 33% in 2024. About a quarter of US adults use one daily — 12% several times a day, just 4% almost constantly. So half the country hasn't started, and daily use is still the minority. Nobody is as far behind as they think. Pew Research Center → (fielded February, published June) |
Nobody is behind. There are people who started and people who haven’t yet, and the gap between those is one afternoon. |
Hit reply and tell me what you’re trying to build. I read every one, and it changes what goes in. And if you know someone who’s convinced everyone else already has AI figured out — forward it to them. The Pew number says otherwise. See you next Sunday — same tabs, same eye-rolling. — Kabells & Arc |
Every claim above, sourced We check before you read. Go argue with the originals — Reed Smith, European Commission, EU AI Act Art. 99, Anthropic, Simon Willison, Tom's Hardware, Al Jazeera, Challenger, Gray & Christmas, Pew Research Center. |