A year ago I stood in the Arctic on polar bear watch, and it changed what I do every morning.

I was on a multi-day trek with a group in Svalbard (shout-out to Peter Zaitsev of Geeks Go Peaks). Someone had to stay awake around the clock in case a bear wandered into camp. I took the 4am-7am shift. We were in the middle of nowhere. I'd take a walk up into the mountains to observe our small camp from above. Just me and a very long horizon. It was polar day (24 hours of sunlight), so 4am looked like 4pm.

One morning, I noticed I felt completely at peace. No angst. None of the low-grade dread that usually hums in the background.

So I tried to figure out where that dread normally comes from. It wasn't my work. It wasn't my family. It was the news and the constant information cycle I typically processed each day.

When I got home, I cut off all “news”. Not a detox weekend. Off. A year later, here's what I've found: if something actually matters, I hear about it anyway. Somebody mentions it, I ask, and I'm caught up in two minutes. Everything else was noise I was paying for with my mood.

There was one exception. I kept my AI newsletters, because following AI is my job. Then I started writing one.

So I'm in an awkward position this week. I'm about to tell you to read less AI news, in an AI newsletter.

Every couple of weeks a primary care physician or ER doctor calls me about starting a preventative health clinic. They all ask some version of the same question: "How do I use AI to differentiate?"

My answer surprises them. Every smart physician in America is opening ChatGPT right now and asking it to design their next clinic. The prompts are equally good. The answers are equally available. AI won't set you apart. It's the minimum.

The better question is what you can do that has nothing to do with AI. Something physical. Something built on relationships. Something a competitor can't copy by typing a prompt.

I learned this the expensive way. In 2022 I co-founded a preventative health company, one of the first to automate the doctor workflow with GPT-4. We had chat-with-your-health-data before almost anyone. What did patients actually value? Time with the doctor. The exact thing we were trying to automate away.

The distinction I use with clients: table stakes vs. edge. Table stakes is anything every competitor can get by next quarter. Edge is what's left after they all have it. AI tools are table stakes now.

The same thing applies to how you follow AI.

Every executive I know reads the same launch announcements, the same benchmark charts, the same "this changes everything" posts. Being well informed about AI is table stakes too. Reading more of it won't give you an edge, because your competitors are reading the same feeds at the same pace.

And most of it doesn't affect your business. My sixth grader doesn't need Einstein for his math homework. A decent tutor is plenty. Most company work is like that. Reviewing a contract, summarizing a report, drafting an analysis: today's models already clear the bar, and the next release won't change your outcome.

Sometimes reacting fast to launches hurts you. I use Claude all day, every day. For my work, Opus 4.7 was worse than the version before it, and Opus 5 was worse than 4.8. Both improved eventually, but for a stretch the newest model was the worst option I had. Anyone who switched on launch day just to stay current got worse results.

A simple filter: signal changes a decision you're making this quarter. News changes your mood. Most of the AI feed is news.

I can't fully follow my own advice, and you shouldn't either.

Some launches really do matter. Earlier this year a client spent more than $600K building a reporting platform the traditional way. Over the summer, new capabilities shipped that let us replace it with a system built on Claude in a fraction of the time. That architecture didn't exist when they started. Somebody needed to be paying attention when it did.

So the answer isn't to ignore AI. Someone in your company should be watching closely. It doesn't need to be everyone, it doesn't need to be daily, and it shouldn't be done in a panic. One person, a weekly filter, and a short list of the decisions that are actually open.

Hold this newsletter to the same test. If an issue doesn't change something you're deciding, delete it. I'd rather you skip one than read it anxiously.

The calm I felt on polar bear watch wasn't from knowing less. It came from noticing how little of what I was taking in had anything to do with my actual life. For most companies right now, that's true of the AI news too.

Homework: List every AI source you or your leadership team checks weekly: newsletters, podcasts, feeds, vendor emails. Next to each one, write the last decision it actually changed. Drop the ones where you can't name anything. Then answer one more question: if every competitor got the same AI tools tomorrow, what would still set you apart? That's where your attention should go.

Two things I'd genuinely like back from you. First, hit Reply and tell me which part of this matches what you're seeing. Second, if there's someone in your world who should be on this list, forward it to them.

- Robbie