Weekly digest #92
Why I’m not worried about AI causing mass unemployment, How inconceivable startups happen, Write your obituary before your resume
We (Matt and Shray) read hundreds of articles on company building, venture investing, and self-management and curate the best ones into a weekly digest to help founders and operators stay on the top of their game.
Better thinking
The gut decision matrix (4 minute read)
Most people use “gut feeling” as if it’s one thing, but there are actually two very different things going on. One is instinct, which is basically your survival wiring kicking in. The other is intuition, which is your brain quietly drawing on everything you’ve learned. They feel similar, but trusting the wrong one at the wrong moment can lead you badly astray. The trick is learning to tell them apart, and the piece gives you a simple way to do it. Ask yourself whether you’re in actual danger, and whether you have real expertise in the situation. If it’s a genuine threat, act. If it’s complex and you’re out of your depth, treat the feeling as a hypothesis rather than a verdict.
Why ATMs didn’t kill bank teller jobs, but the iPhone did (20 minute read)
Long but worth it. The piece takes a comforting parable that economists love (ATMs didn’t kill bank teller jobs, therefore technology creates as much work as it destroys) and shows it’s only half true. The iPhone quietly finished what the ATM started, not by automating teller tasks but by making the whole branch model obsolete. The broader argument is that task automation within an existing paradigm rarely displaces workers, because human-shaped workflows create endless friction. What actually displaces workers is when a new paradigm makes the old one irrelevant entirely.
Why I’m not worried about AI causing mass unemployment (12 minute read)
Marc Andreessen declared in 2011 that software would eat the world. It mostly nibbled. Airbnb dented hotels, Uber improved on taxis, and WeWork was a fiasco. The argument is that the same friction will slow AI down. Yes, white-collar workers sitting in front of computers all day will feel the disruption, just as they felt the internet. But most of the economy isn’t that. Plumbers need to climb stairs. Starbucks customers are partly paying for a human to make their coffee. Most jobs are stubbornly physical, and that’s not going away.
Operational tactics
Claude Cowork might be the most consequential piece of software you’re not using (10 minute read)
Most people open Claude Cowork, ask it a question, and close it. This piece is for anyone who suspects there’s a lot more under the hood. Over two months, one person turned it into a full working environment, building plugins that triage his inbox, skills that fill out his kids’ medical forms, and a pipeline that decoded his own DNA. The piece walks through exactly how he did it.
What founders get wrong about selling to enterprises (8 minute read)
Enterprises buy the least risky path forward, full stop. This piece is for founders who keep losing enterprise deals on merit they think should be enough. The gap is rarely technical. The real buyer turns out to be the legal, compliance, and risk teams sitting behind the curtain with quiet veto power, and most pitches never reach them. The piece walks through how to find those people, speak their language, and design a path to yes that feels boringly safe.
The Post-IPO PM playbook is being rewritten (58 minute listen)
There’s a lot of conversation right now about the PM function being rebuilt. Most of it comes from two ends of the spectrum: AI-first startups that are slim, flat, and often still finding product-market fit, or big tech companies with thousands of PMs operating at massive scale. But there’s a third category that doesn’t get enough attention: companies that are far older, successful, and have gone from startup to public company in the past few years. Nikhyl (three-time founder, CPO, and product exec at Google and Meta) sits down with the CPOs of Hims & Hers, Rubrik, and Figma to understand why it’s never been a more exciting time to be a PM.
Refer and we’ll send you our favorite books as a “thank you” for spreading the word.
Private investing
Ben Horowitz on what makes the best fundraising pitch (5 minute read)
Ben Horowitz thinks great fundraising hasn’t changed much. The best pitch is simply the one you genuinely believe. Founders make the mistake of trying to reverse-engineer what an investor wants to hear, but that backfires because the argument was never really theirs. If you can convince yourself this is the greatest idea in the world, that conviction is what actually lands.
Institutional AI vs. individual AI (10 minute read)
So good. Anyone working in a company of >50 people will feel this one. It argues that while Individual AI makes people more productive, “institutional AI” makes organizations more valuable and coherent. Most companies only focus on the former (giving employees tools), but don’t address the latter (changing workflows around AI entirely). Drawing on the 1890s textile mills (which electrified but didn’t redesign the factory floor, and lost to those who did), it makes the case that the real opportunity is building AI around how organizations actually work, with coordination, signal-finding, objectivity, and domain-specific edge, rather than just giving individuals faster tools.
How inconceivable startups happen (7 minute read)
Doing the inconceivable feels petrifying but once you’ve done it, the next inconceivable task feels less so and your gains materially compound. Sequoia’s David Cahn has spent considerable time pondering about identifying founders who establish inconceivable companies and boils it down to three key ideas: 1. Knowing something others don’t 2. Believing that it must happen 3. Experimentation. He suggests that for investors to train their mind to see the inconceivable, they must attempt inconceivable outcomes themselves (and verify it’s possible) or engage in deep reading to subdue their own subjective biases and expand their critical thinking toolbox. But truth be told, it’s much easier to invest capital behind the “sure thing” rather than the “crazy idea that could put a dent in the universe”. As long as humans fear being wrong, there’s returns to be made in backing the inconceivable.
Career management
AI tooling for software engineers in 2026 (20 minute read)
This is a paid post, but you get a lot of info without having to pay. Gergely’s Survey of 900+ engineers reveals that Claude Code is the #1 AI coding tool just 8 months after launch. 95% of engineers use AI weekly, with 75% using it for half their work. Staff+ engineers lead agent usage, and smaller companies favor Claude Code while enterprises default to GitHub Copilot.
Making Claude Code my Chief of Stuff (7 minute read)
We haven’t dug into this ourselves, so we can’t fully vouch for it. But I’m seeing more genuinely smart people do this kinda things. This piece walks through how one person built something closer to a personal chief of staff, a system that syncs meeting notes, preps her calendar, drafts emails in her voice, and organises her files, all before she sits down at her desk. It’s technical in places, but the real value is in seeing how someone thought through their own workflow carefully enough to actually automate it.
Write your obituary before your resume (4 minute read)
A Stanford professor asked his MBA students to write their own obituaries before their careers had even started. The idea is simple but quietly unsettling. Most people optimize for motion (the next job, title, promotion) without asking what kind of story they’re actually writing. Starting from the ending and working backwards turns out to be the clearest way to figure out whether the life you’re building is the one you actually want.