Sunday, August 9, 2026 · about a 1 minute read
Today's stories share a quiet tension: AI keeps getting more capable and more costly, and the two things are not unrelated. That tradeoff is starting to show up in real company decisions.
Get the calm version of AI news.
One email a day on what is actually happening in AI, in plain English. No hype, no doom.
Free. One calm email a day. No hype, no doom.
SAP is pausing most travel and new hires specifically to fund its AI spending, which means the cost of running these tools is already reshaping how large companies allocate budgets. If you work at a big company, do not be surprised if your own org starts making similar tradeoffs quietly.
Anthropic flipped the default so that Claude Code now works in auto mode, meaning it picks its own approach rather than waiting for you to configure it. That is a small settings change on paper, but it signals that AI coding tools are moving toward doing more on their own before you even ask.
Get this every morning.
Auto mode in an AI works a bit like cruise control. You still set the destination, but you stop touching the pedals. The model decides when to look something up, when to write code, when to stop. The appeal is speed. The risk is that you can drift somewhere you did not intend if you are not watching the road.
John Gruber's framing of writing like live music versus a studio album is a useful lens for thinking about how people are starting to use AI in creative work: some output is meant to be polished and deliberate, and some is just meant to be present and honest. Knowing which one you are making changes how much you should lean on a tool.
That's today. See you tomorrow.
Get this every morning.
One email a day on what is actually happening in AI, in plain English. No hype, no doom.
Free. One calm email a day. No hype, no doom.
The book behind this newsletter
How ChatGPT, Claude, and Modern AI Actually Work
The trick is small. The world it built is not.