Tools for progress #66
Think about what to focus on, OpenAI: Building the "Everything Platform" in AI, How to start a new job
We (humans) read hundreds of articles on company building, angel investing, and self-management and curate the best ones into a weekly summary—helping founders and operators stay on the top of their game.
Better thinking
Think about what to focus on (15 minute read)
We highly recommend reading this one. It tackles this idea of “exploring” and "exploiting" new ideas. That is, finding things to work on and actually working on things. Spend too much time exploring and you’ll get nothing done. Too much time exploiting and you might end up working on the wrong things. Intentionally design periods of time (on whatever time frame makes sense to you) to exploit and to explore. Don’t stay in one camp for too long.
IRL brain rot and the lure of the Labubu (14 minute read)
This is such an intriguing read, I enjoyed it a lot. If nothing else, it’ll give you a quick window into the GenZ brain. We’re seeing more strange products pop up everywhere - Labubu dolls, Dubai Chocolate, and matcha everything. Deliberately ugly, these things bring the internet brain to life in physical form. Our taste and preferences are being driven by an algorithm. And that algorithm is increasingly a little weird.
AI x Commerce (10 minute read)
This piece breaks down how AI might kill Google's cash cow by looking at different ways people shop. Google's probably fine for now since AI is mostly stealing the worthless "how many protons in cesium" searches, but once AI gets good at helping people buy stuff, especially the middle-tier purchases like laptops or skincare, that's when things get interesting.
Operational tactics
How to build a coding agent (20 minute read)
This workshop breaks down how coding agents like Cursor and Claude Code actually work. Just 300 lines of code running in a loop with LLM tokens. You'll learn the five core primitives (read, list, bash, edit, search tools) and discover that building your own agent isn't all that difficult, and in fact, its engineering any developer should probably understand in 2025.
10 lessons from a convo with the CEO of Flexport (2 minute read)
Ryan Peterson is the CEO of Flexport. Safe to say, scaling the company to >$1bn has been a bumpy right. Shane Parrish (Farnam St) sat down with Ryan, and drew up these 10 lessons from the conversation. This is a great example of a CEO in founder mode.
Refer and we’ll send you our favorite books as a “thank you” for spreading the word.
Venture investing
We have no idea how to value the AI app layer (4 minute read)
Boggled by all these “$100m ARR in 12 months” claims? Repeat after me, “throughput is not ARR”. It’s more than likely that the company making these claims is a closed source AI tool tool. Meaning you can’t bring your own model keys, and you have to use their underlying providers. Which means the payment for that usage passes through the company, and they’re booking it as revenue. It shouldn’t be considered revenue, come on.
Angel investors, a field guide (10 minute read)
As a first-time founder, cap table space will be limited, signaling matters a lot, and your network can go a long way in helping you signal. Jean Yang, founder of Akita (acquired by Postman), had little idea how much working with the right angels could help when she first became a founder. After partnering with and learning from angels like Jason Hong, Elad Gil, Kevin & Julia Hertz, Dan Boneh, and even Kevin Durant, Jean outlines everything she would tell a first-time founder (who is headed down the VC-funded) route about working angels.
OpenAI: Building the "Everything Platform" in AI (45 minute read)
Leonis Capital profiles OpenAI, one of the most dominant AI players. From its humble beginnings as a non-profit research lab to the world’s most valuable private company as it targets a $500B valuation, OpenAI has positioned itself as the industry’s ‘everything platform’; building an everything app to become the foundational layer upon which all AI-powered applications, services, and interactions are built. OpenAI envisions its platform as a convergence point for specialized AI capabilities across every domain of human activity. However, its ambitions aren’t unchecked. Google’s search dominance, Microsoft’s productivity suite franchise, and Meta’s social networks all face disintermediation if OpenAI successfully executes, leading the hyperscalers and other AI labs to challenge the front runner in what could prove to be the most transformative technology in human history.
Managing your career
Don’t keep climbing the wrong hill (3 minute read)
A short essay from Chris Dixon. Many people are too afraid to do something new because it places you at the bottom of a new hill. And so they cling to the hill they’re already on (like a career pathway or relationship), despite, deep-down, knowing they don’t really want to be on the top of it. Know when you’re on the wrong hill and be brave enough to climb back down.
How to start a new job (7 minute read)
After reflecting on her past 8 months as Harvey’s new Chief People Officer and 11.5 years skillfully observing new leaders onboard at Hubspot, Katie Burke shares useful advice and watchouts for people starting new roles. It’s a densely packed and compartmentalised guide with a parting note to give yourself some grace because as a new joiner, it’s unlikely that you’ll get everything right the first time. But don’t forget that you might find yourself on a rocket ship with people who are willing to help you get it right as you move forward.
Respectful performance management (3 minute read)
Performance management conversations are difficult, no matter what side you’re on. As the employee, it may feel unfair, surprising, and overwhelming. As the manager, you’re prone to doubting your judgment. Either way, it’s emotionally taxing but a normal part of working life. Ami Vora, Faire’s CPO, recommends incorporating six principles to handle such conversations with clarity and respect. One principle that stood out was giving the person a strong chance to demonstrate their skills. Ami prefers giving people 2 projects, one tactical and one strategic, to be completed in 4 weeks with regular check-ins. In many cases, it turns out people needed clearer feedback and ultimately crank out excellent work products. You want to bookmark this one for that inevitable conversation.