It's been a couple weeks. Sorry for the quiet.

I'm not going to pretend I was off building something heroic. I just went heads down and the newsletter slipped. But May turned out to be one of the wildest months in AI I can remember.

Three big labs swapped their default model inside thirty days. Most of it won't change your Tuesday. One small thing might.

— Khadin

In today's issue:

Hit reply and tell me what you think. I read every one.

When you type a prompt, you summarize. When you speak one, you explain. Wispr Flow captures your full reasoning — constraints, edge cases, examples, tone — and turns it into clean, structured text you paste into ChatGPT, Claude, or any AI tool. The difference shows up immediately. More context in, fewer follow-ups out.

89% of messages sent with zero edits. Used by teams at OpenAI, Vercel, and Clay. Try Wispr Flow free — works on Mac, Windows, and iPhone.

Start flowing free

The Scoop: Anthropic shipped Claude Opus 4.8 on May 28 with a new "Effort" dial that lets you trade speed for care on any prompt.

Until last week, you got one Claude. It thought as hard as it thought, and you took what you got.

Then Anthropic added a slider. It sits right next to the model name on claude.ai and in Cowork. Turn it up, Claude thinks harder and catches more. Turn it down, Claude answers faster and saves your usage limits for later.

Here's why this matters for your actual job. Most of what you ask AI to do is small. A subject line. A quick rewrite. A summary. You don't need maximum brainpower for that, and burning it there means you hit your limit by 2pm. Save the high setting for the one thing that actually matters this week.

Everyone treats "smarter" as always better. It isn't. The skill now is knowing when to dial it down.

So what for you: Stop running every prompt at full power. Match the effort to the task and you get more done before you run out.

The next wave of millionaires will be people who figured out how to make AI work for them.

The window to get ahead is still open. But not for long.

Here are 200+ proven ways to make money with AI in 2026.

Sign up for Superhuman AI, the free daily newsletter read by 1M+ professionals, and get instant access to all 200+ ways to profit from AI this year.

Claim your free list

The Scoop: At I/O on May 19, Google gave Gemini a "Daily Brief" that pulls your inbox, calendar, and top tasks into one morning summary.

Picture opening one screen at 7am and seeing the three things that matter today, already sorted. That's the pitch.

It reads your Gmail, your Google Calendar, and your task list, then hands you a short digest instead of three open tabs. If your morning starts with you staring at a full inbox wondering where to begin, this is aimed straight at you.

I'm not sold yet on letting one app read everything. But the direction is clear. The assistant that owns your morning owns your day.

So what for you: If you live in Gmail and Calendar, try the Daily Brief for a week and see if it actually saves you the 10-minute "what do I do first" scramble.

The Scoop: Federal Reserve research clocks generative AI's average time savings at 5.4% of work hours, roughly 2.2 hours a week. Stanford tracked 200,000 US households and found ChatGPT users spend most of that saved time on leisure, not on skill-building or extra output.

The productivity story gets messier here. Yes, AI saves time. About 2.2 hours a week for the average user, and 9+ for heavy users. But Stanford's data on 200,000 households shows the saved time mostly goes to leisure (TV and scrolling, mostly). Which is fine for individuals.

It just isn't the "AI is making the workforce dramatically more productive" narrative being used to justify capex and cuts.

So what for you: Time saved by AI isn't a productivity gain unless someone reinvests it. Most people don't. The layoff press releases skip that part.

Quietly, the most useful change in Opus 4.8 isn't speed. Anthropic says it's about four times less likely than the last version to let a mistake slip by without flagging it. Early testers said it owns up to what it isn't sure about instead of bluffing.

That's the thing nobody markets but everybody needs. An AI that says "I'm not certain here" saves you the bad afternoon of trusting a confident wrong answer.

So what for you: Next time the answer looks too clean, ask "what are you unsure about?" The newer models will actually tell you.

Before your week starts, do this once:

Ten minutes, and you'll know which tasks deserve the good setting for the rest of the month.