The daily brief you'll actually keep reading

Something that works while you're on the floor

Both Claude and ChatGPT can now run a task on a schedule. You write it once, pick a time, and it goes on its own whether you're at your desk or in the middle of a lift.

Most people use this for reminders, which is a waste of the technology. The better use is for learning. A research update that's waiting for you at 6:00am. A daily walkthrough of one topic you're trying to get better at. A weekly scan of what's new in an area you can't keep up with.

One problem is that a scheduled task runs on time no matter what. It doesn't know the difference between a day with three studies worth reading and a day with nothing. If you haven't told it how to respond in that scenario, it'll fill the space anyway. That's how you end up with a daily brief you stop opening after two weeks.

You're not writing a prompt. You're writing a job description.

Think about how you'd actually train a new intern to do this job.

You wouldn't say "keep me updated on research" and walk away. You'd tell them what you're trying to get out of it, which sources you trust and which ones you don't, what to do on a slow week when there's nothing worth reporting, and what format you want it in so you can skim it in a few minutes.

That's what separates a scheduled task worth reading from one you'll let sit without opening.

Six things that belong in it

• Say what the job is for. Not the topic, the purpose. "Help me become a better practitioner" produces different work than "send me sports science news."
• Set the source bar. Name the journals, organizations, and platforms you actually trust. Say whether practitioner content counts, and if it does, require that it gets labeled as practitioner insight rather than evidence.
• Tell it what to do when there's nothing. This is the one everybody misses, and it's the difference between a brief you trust and one you don't. Some days there is no meaningful new research. Tell it to say that plainly and point you toward something older worth revisiting instead of inflating a weak study to fill the space.
• Make it label confidence. Strong evidence, emerging evidence, practitioner insight, opinion, worth monitoring but not proven. When every claim carries a label, you can read fast without losing track of what's actually established.
• Lock the format. Same sections, same order, every day. You learn where to look and reading it stops being work.
• End with action. A section that asks what you could test this week, or questions worth thinking about. Otherwise you've built a reading habit instead of a coaching habit.

Inside the Lab: setting it up

First thing worth knowing: this is a paid feature in both tools. ChatGPT requires Plus or higher, and Claude's scheduled tasks run inside Cowork on any paid plan. If you're on a free account, this one isn't available to you yet. However, everything about writing the instructions still applies whenever you get there or if you'd rather just run it manually.

In ChatGPT there's a Scheduled page in the sidebar on web, mobile, and desktop. You can create it there, or type the whole thing into a chat with the cadence stated up front and confirm when it asks. In Claude, you set it up in Cowork, and it runs remotely, so the task fires whether your laptop is open or not.

Either way, the setup is the same: write the instructions once, state the schedule inside them, and let it run.

The instructions are the work. Expect yours to be long, and expect to revise it for the first week or two, because the first few runs will show you what you forgot to ask for.

If you want a working example rather than a blank page, I built one. The Daily Sport Performance Brief is the exact prompt I run, weekdays at 6:00am, covering research, applied ideas, nutrition, recovery, return to play, and tech, with sources cited and evidence labeled. It's free on the site. Copy it, run it as is, or use the structure and adjust the wording for whatever you're actually trying to learn.

What it looks like in the real world

The version I run covers everything from new research and applied coaching ideas to nutrition, recovery, return to play, and sport tech. All of that is useful, but there are two sections that leave me with an actual next step.

One is a further learning queue. Five to ten resources with a line on why each is worth my time and what question it answers. That turns a wall of links into a reading list I'll actually get through.

The other asks what I could apply this week, and gives me three practical ideas I could test or bring to staff. That's the part that moves something off the page. Everything above it tells me what's out there, and those two sections are what turn it into something I can use.

Where the coach stays in charge

Check the links, at least early on. A brief that cites a study is not the same as a brief that reads it correctly. The summary can be right about the paper's existence and wrong about what it found. If something's going to change what you're doing, go read the source yourself.

If you find yourself skipping the brief for a week straight, shut the task off or rewrite it. ChatGPT will auto-pause tasks you ignore, which is a pretty good hint. A scheduled task you don't read is worse than no task at all, because it feels like you're staying current when you're just stacking up output nobody opened.

Start with one. Run it for two weeks and adjust the instructions before you add another.

The Coach's Scouting Report

• Signal. A scheduled task runs whether it has anything useful to say or not. Write the instructions so it knows how you want it to respond in different situations.
• Noise. Scheduling five things at once. Reading briefs without acting on them.
• Next Rep. Pick one thing you're trying to get better at. Write the job description for it: purpose, sources, what to do on a slow day, how to label confidence, the format, and a section that ends in action. Schedule it for a time you'll actually read it.

See you next Sunday.