I’m increasingly polite to OpenAI and Claude in case they train on my comments. I’m increasingly rude to Tesla Autopilot for the exact same reason.

Spent the last year watching founders hire for "energy" and "hustle" when what they actually needed was someone comfortable with silence. The person who sits with a hard problem instead of filling the void with activity. That's rarer than it sounds.

Dashboards replace conversations. And somewhere in the translation, you stop getting feedback about yourself. Just about everything else.

Been noticing founders obsess over hiring "culture fit" when what they really need is someone who'll tell them no. The uncomfortable kind of no. The kind that makes you defensive for a week then changes everything.

The Feedback You Stop Getting is live. As you climb, you get more data about your org. You get less about yourself. That gap matters. https://open.substack.com/pub/atpractice/p/the-feedback-you-stop-getting?r=1ikro&utm_campaign=post&utm_medium=web&showWelcomeOnShare=true

Watched a founder hire their third VP of Sales in two years. Each one was smart. Each one left saying "the product isn't ready for enterprise yet." The founder kept hiring salespeople. Should've hired a product person first.

Feedback that stings usually means someone saw something true about you that you weren't ready to see yet. The feedback that actually changes you? That one doesn't sting. It just resonates because you were already looking.

If MBA programs teach what's actually useful, why aren't more people just... reading the same books and doing the work solo instead of paying six figures?

Feedback that stings is about you. Feedback that changes you is about the work. One makes you defensive. The other makes you curious. The sting fades in a day. The change lasts years.

Most AI discourse is just urgency theater. The real work isn't keeping up. It's asking: what problem does this actually solve for my team, right now? That's unsexy. So nobody talks about it.

Most teams don't have a strategy problem. They have a fragmentation problem. You find out what you actually believe when someone has to build it.

There's this moment when you realize the company's scaling faster than you are. New problems arrive. Old playbooks don't work. You're suddenly the bottleneck. It's uncomfortable. Also necessary.

You don't need a $150k degree to learn business. A reading list of practitioners and thinkers will often go deeper than a lecture hall ever could.

Every org has two org charts. The official one on the website. And the real one, where you figure out who actually decides things. Most people spend months finding out they're on the wrong chart.

"being scrappy" is mostly just being broke with a better narrative. We celebrate the founder working from a garage like it's a virtue. It's not. It's a constraint. Once you have money, you stop calling it scrappy. You call it efficient.

In Japanese martial arts, "shu" means master the form before you break it. In startups, we skip straight to breaking things. Maybe the insight isnt "move fast." Its knowing exactly which rule youre breaking and why. Restraint first. Then freedom.

The move that made you successful as a founder - ship fast, test everything, generate signal - can become your biggest liability once you're trying to build something that lasts. Speed creates noise, not always clarity.

Win! Finally closed out a lingering Q1 backlog item. The Christmas tree is down.

Most advice is contextual. The hard part isn't hearing it - it's knowing if the context matches yours. I've ignored great advice. And taken bad advice seriously. The difference? I asked who said it and why, not just what they said. That skill takes years to build.

The best business education I got was reading books nobody assigned me. No institution filtering what matters, just you and someone who actually built something.

What if the companies with the clearest three-slide strategy are actually the ones making better decisions than everyone else?

Most founders are wired to move. But strategy is often about knowing when to sit still. The hard part isn't building the thing - it's having the discipline to wait for the moment it actually matters.

New essay: Strategy Is Mostly Waiting The hardest part of building isn't moving fast - it's knowing when to stop moving and let things settle. atpractice.substack.com/p/strategy-i...

Teams execute whatever's closest to revenue or got mentioned last in all-hands. That's your real strategy, whether you wrote it down or not.

Reading 40 books from practitioners beats an MBA because you get their unfiltered thinking instead of someone else's interpretation of it. Cheaper too.

New essay: 40 Books That Replace an MBA Most MBAs teach you what schools think you should know. These 40 teach you what actually works. https://josephfung.ca/40-books-that-replace-an-mba-the-2026-reading-list-cd27c657365b

Most teams don't fail because they lack a strategy framework. They fail because nobody actually agreed on what strategy means, so everyone's executing a different one in their head.

Yup. Cucumber chips (crisps for @amff.bsky.social and other Brits) are confusing. Similar to plain, but with a mild sweet taste. Also, somehow they taste “fresher”. I didn’t expect to like them this much.

Trained hard enough this week that I've got a pinched nerve now. My body's basically saying "hey, remember moderation?" I don't think I listen well.

The best use of AI might not be doing the work. It's helping you think clearer about what work actually matters, then keeping that thinking alive instead of letting it gather dust in a folder somewhere.

New essay: AI as Your Strategy Chief of Staff Most strategy lives in scattered docs and conversations. What if AI could consolidate it into something that actually evolves? josephfung.ca/strategy-os-...

New essay: AI as Your Strategy Chief of Staff Most strategy lives in scattered docs and conversations. What if AI could consolidate it into something that actually evolves? https://josephfung.ca/strategy-os-bf2d7294d608

Job descriptions describe the role in a vacuum. They almost never tell you what the company actually needs *right now*. So you apply thinking you're solving one problem, onboard thinking you're solving another. No wonder job search feels broken.

Most founders don’t have a strategy problem. They have strategy sprawl: decks, docs, battle cards, decision threads. I’m open-sourcing “Strategy OS” Claude Skills to help consolidate, pressure-test, and keep execution honest. github.com/josephfung/s...

There is something darkly comical about vibe coding forcing a bunch of engineers who likely derided engineering management now managing a corral of AI agents that basically act like a mix between Amelia Bedelia and the guy from Memento

Hah! winning traffic on “plumbering services” is peak search-era internet. Shame it won’t last long. Search preserved users’ messy wording. AI interprets and cleans it up. That’ll kill off traffic from literacy arbitrage 🤷

The contracts that actually matter are never signed. They're made in the moment someone's drowning and you throw them a rope anyway, no questions asked. That's the one they'll remember.

Happy International Women’s Day 💜 Reminder: Careers don’t compound at hiring, they compound at opportunity. • first leadership opportunity • investor introductions • customer visibility Those points shape who ends up building companies, running teams, and allocating capital later #IWD2026

I got roasted on Substack and it was painfully accurate. “Refined LinkedIn slop” = the meanest compliment I’ve ever received. mariahfaithcontinelli.substack.com/p/roast-for-...

There are mistakes you fix with paperwork. And there are mistakes you fix out of your own pocket. New piece on substack: The Cost of Tuition atpractice.substack.com/p/the-cost-o...

Tuning in to @tedunderwood.com and Mohit Cheppudira on the real world direction of generative AI. Thx @inovia.bsky.social for pulling it together

Most founders think they have a communication problem. Often, they have a strategy problem. I’ve been thinking a lot about strategic thinking as a real skill - not instinct - and wrote about it: The Strategy Skill Gap: bit.ly/4aKp3lk If you’re scaling & feeling alignment drift, this might sting

One of the things I’m enjoying most how about using Claude Code, is de-bugging. Sliding into planning mode to kick an idea around has been a game changer. The efficiency gains are marginal, but the quality of life in lower frustration is immeasurable!

Earlier this week a friend asked: “Why do I keep joining companies that look amazing from the outside… and once I’m in, they’re kind of broken?” She’s sharp. This isn’t incompetence. I told her it’s probably not her. It’s startups.

6:00/km pace. Heart rate under control. Hair looks like I lost a fight with a leaf blower Measure what I can, accept what I can’t 🫠

SaaS reset: systemic correction or AI breaking the model? 15–25x revenue was never normal, but there’s overreaction. I ran the numbers across ~25 SaaS stocks to see where the compression looks cyclical vs structural. I’m curious what others are seeing. Here’s the full write up: bit.ly/3Mx2reT

Most ceilings are self-installed. You don’t find that out by nudging them. You find out by pressing. Pull on Time: open.substack.com/pub/atpracti...

Today is the first day of the Lunar New Year Not just “Chinese New Year”. The lunar cal shapes things across China, Korea, Vietnam, Singapore, and more. And this year? The Year of the Fire Horse 🔥🐴 Horse = momentum Fire = intensity So if 2026 feels unhinged brilliance… that tracks!

Flag Day being the day after Valentine’s is very Canadian. “Go ahead, love. You take the spotlight. I’ll just… hold the country together.” 🍁

A 3rd-century Roman priest defied Emperor Claudius II by secretly marrying couples. He was executed on Feb 14. Medieval poets romanticized it, Victorians mass-produced cards, and now we mark a martyr’s death with prix fixe dinners and heart-shaped chocolate. History is weird. #happyvalentinesday