I’m an avid reader. Not in the polite sense, in the compulsive sense. I consume a lot across formats and topics, whatever device happens to be in my hand.

I find things everywhere:

• Social media puts something in front of me

• A newsletter lands in my inbox

• Someone sends me a link on WhatsApp or Slack because they know I’ll actually read it

None of that stops. I’ve made peace with that part.

What I actually love isn’t just the reading itself. I am absolutely obsessed with the workflow around it. Save it, come back to it later, on my terms.

I read across devices too:

• Phone, when I’m out

• iPad, when I’m on the couch

• Mac, when I’m at my desk

I also stopped calling it a reading list. It’s a consumption list. Reading covers tweets, articles, the odd ebook I keep meaning to finish. I’m also watching and listening to things I’ve flagged to get to. All of it goes into the same holding pattern. All of it needs to be prioritized, tracked, revisited.

What I’ve tried

I’ve used a lot of apps trying to hold this together. I liked capturing something the moment I found it, from any app, through the share sheet. What none of them handled was the rest of it.

• Pocket. Worked quietly for years. Mozilla shut it down in July 2025. No warning that mattered, just an export window, then gone. That’s the risk of building a habit around someone else’s product.

• Readwise Reader. The best reading experience I’ve used. Also the one I’ve canceled and come back to more than once. Bugs that shouldn’t exist at ten dollars a month, and it occasionally just doesn’t sync between devices, which defeats the point when my whole habit is switching screens mid-read.

• Raindrop. Still on my phone, but I’ve stopped pretending it’s a reading app. Good at one thing: organizing links into categories I might open again in six months. Continuous reading was never what it was built for.

Every one of these asks for the same thing: a login, a subscription, one more piece of my attention that now has to live somewhere new.

I’ve spent years trying to shrink the number of apps holding pieces of my life.

I already live inside Reminders. I already live inside Notes, iCloud, Mail, every day, without thinking about it. So why was my consumption list the one part of my life that lived somewhere else entirely.

This wasn’t my first attempt at fixing it, either:

• A couple of years ago I built a barebones link saver in Shortcuts. It grabbed a URL and dropped it somewhere. Didn’t check if the page was readable, didn’t handle files, just saved a link and hoped.

• I gave up on it fast and went back to paying for Readwise. Yikes, looking back.

• Outside of that, my Shortcuts use was small stuff: highlight text into Notes, download a file into a folder.

Nothing that felt like it could hold real weight. I’d basically filed Shortcuts under toy.

What I actually built

Here is the entire spec for those interested.

This time I specced it properly, with a lot of back-and-forth with Claude. What I thought I needed was simple: save something, get a reminder to come back to it in a couple of days.

Here’s what actually happens now when I share something, from any app, on any device.

If it’s a link:

• Cleans it first, following redirects and shorteners to the real destination

• Pulls the page’s own title, so the reminder is descriptive, even for a tweet

• Checks the host. Sites like X can’t be turned into a clean article and will crash an extractor if you try, so those skip straight to a link-only reminder

• Everything else gets a real attempt. If it’s a genuine article, it builds a PDF from the body and saves it offline. If it’s not, homepage, dashboard, whatever, it falls back to a link-only reminder instead of failing outright

If it’s a file:

• Reads the file’s name and uses it as the title

• Timestamps the filename so nothing overwrites something else

• Copies the file into my Reading List folder for offline access

• Hands the reminder a tappable link straight to the copy

One rule sits above all of it: the reminder always gets created, no matter what else fails. If iCloud is down, if the PDF step breaks, the reminder still exists. The PDF is a bonus. The reminder is the constant.

It works the same on my phone, my iPad, and my Mac, because that’s actually how I read, whichever screen is in front of me.

I didn’t ask for any of that complexity. It came out of actually thinking through what happens when things go wrong, which turns out to be most of what building something properly requires.

The part that surprised me

Apple Shortcuts can call external APIs. It can hold real nested logic, several layers deep. It can manipulate data, not just move it around.

I had never imagined the app I mostly used to copy stuff was capable of any of that.

The tool that finally solved this was free, already installed, and never once asked for a login. I’d spent years paying other people to almost solve this, when the thing capable of actually solving it had been one tap away the whole time.

Here is the Shortcuts link if any of you wish to fork it.

I remember the early 2020s vividly, the narrative was shifting. Marketing wasn’t just lead gen play anymore. It was accountable for closed-won, retention, expansion, the whole customer journey.

It also spun a lot of new and funny titles. Pipeline Marketing Manager. VP of Customer Marketing. Full Funnel Lead.

Fair enough. That push wasn’t entirely wrong.

But I think it sent a lot of marketing leaders in the wrong direction.

The instinct it created was to design every program, every dollar, every campaign around moving the full funnel. And if you couldn’t show how it connected to the aggregate metric, it felt hard to justify. Especially when the board and the investors push for LTV over anything else.

Here’s where I think that breaks down. Most of your dollars are going into one specific stage. They should be. A webinar for mid-pipeline prospects is an acquisition play. Run it like one. The moment you start asking how it contributes to retention or long-term customer economics, you’ve muddied what you’re actually trying to do.

I know what you are thinking. Yes, early decisions have downstream effects. Good acquisition feeds better activation. Sharp onboarding feeds retention. Those patterns are real. But noticing them isn’t the same as designing programs to move everything at once.

The aggregate metric is the output of individual stages done well, over time, in parallel. I wouldn’t go at it directly. I’d find the stage that’s most broken or most opportunistic, put focused dollars there, and run it long enough to actually see results.

In my experience, that’s always where the returns have come from. Not from trying to attribute everything to everything.

I’ve sat across enough founders to know what borrowed confidence looks like. The words land well. The framing is tight. But ask one follow-up question and you can feel the floor give way. No scar tissue underneath. No story. Just a clean answer that nobody actually owns.

This is happening more. And I think AI is why.

We’ve seen this before

This is what happened to corporate executives in the 90s when management consulting became a default. McKinsey would come in, synthesize everything, and produce a beautiful recommendation. The executive would present it. The team would nod. Everyone would mistake the clarity of the output for the quality of the thinking.

It worked until something broke. Because when things break, frameworks don’t hold. Conviction does. And conviction is the one thing you cannot outsource.

We’re doing it again.

AI is a better consulting firm than McKinsey ever was. Faster, cheaper, always available, and it argues in your voice if you prompt it right. Ask it what you should build, who you should sell to, where your market is going. It will answer. The answer will be good. Good enough to feel like something you figured out yourself.

That’s the trap.

The gap you can’t see until you need it

A founder’s understanding of their customer isn’t a research output. It forms through repetition and discomfort. Through a hundred conversations where someone said no and you paid attention to exactly how they said it. Through deals that closed for the wrong reasons and ones that didn’t for reasons that turned out to matter. You don’t generate that kind of understanding. You earn it slowly, and it lives somewhere below the surface.

AI compresses that into a paragraph. And founders take the paragraph and run.

What makes this harder to catch than the McKinsey problem is that executives knew they were outsourcing. They said it openly. With AI there’s no distance. The output arrives in first person, shaped to your context, sounding like a thought you had. Founders think they’re thinking.

That gap is invisible until you actually need what’s underneath.

You need it when a prospect says “why would this ever work for us” and the deck isn’t enough. You need it when your first big product bet is failing and you have to decide whether to push or fold. You need it when you’re recruiting someone who has options and they’re trying to figure out if there’s something real here. People can feel conviction. They can also feel its absence.

The input has to be yours first

I’m not saying don’t use AI. That’s not the point.

The point is that there are a few questions, just a few, where your input has to come before anything else. What do I actually believe about who needs this? What am I building toward, and why that and not something else? What would I hold onto if everything else got challenged?

These aren’t research questions. They’re belief questions. AI can sharpen a belief. It cannot form one.

The executives who got real value from McKinsey were the ones who already knew what they thought and used the consultants to pressure-test it. The ones who got the least walked in with an open mind and walked out with someone else’s strategy.

The founders who will get the most from AI are the ones who bring something real to the conversation. Already formed, possibly messy, maybe half-wrong. But theirs.

The ones who walk in empty and walk out confident are building on borrowed ground.