Weekly digest #88
The last temptation of Claude, The feature backlog has gone *poof*, The definitive guide to Forward Deployed Engineering
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 last temptation of Claude (6 minute read)
Today’s temptress is AI. It can help us with tasks big and small, from compiling a Twitter thread to vibe-coding a software application. It’s remarkable and its temptations will continue to pulsate. But these temptations have a darker side. By hitting snooze on deliberation with the underlying process, we limit our capacity to see our rationalizations clearly. Slowly but surely, one’s akrasia (weakness of will) grows steadfast. While it’s true that no one wakes up one morning and decides to stop thinking, it usually begins by outsourcing the stuff that doesn’t matter and suddenly, the boundary between trivial and meaningful becomes harder to spot. The email becomes the memo, the proposal becomes the argument. But one more time can’t hurt, can it?
Becoming an artist to outrun the machines (7 minute read)
In the AI era, the intangible is becoming more important. As LLMs progressively consume all the available writing, images, video, code, data, etc., what’ll be left for humans is art; intuition, feel, anything that’s difficult to put into words or systematize. There are three intangibles in the business world: taste, judgment, and influence. The people with the best taste know taste is about removing things, not adding them. Judgment is about prioritizing what to do when you can do everything. To build anything, be it a product or company, you’ll need to wield persuasion and influence. If you’re trying to become an artist, consume information intentionally, get many feedback loops, work in-person, seek out great mentors, and generalize on skills, while specializing on problems. Do this and maybe, just maybe, the machines won’t catch up.
This simple skill will put you ahead in the era of AI (6 minute read)
The machines are getting better at sounding smart. Which means we need to get better at actually becoming smart. We need the judgment to separate signal from noise. We need the discernment to know something seems a little off. We need the curiosity to not be satisfied with first answers. We need patience and discipline. We need wisdom. Now more than ever. Here’s Ryan Holiday of the Daily Stoic elaborating further in a punchy X post.
Operational tactics
My top 54 PM AI guides, in one place, for free (4 minute read)
Here’s Tal Raviv’s list of categorized tutorials to stay ahead of current AI developments. Most of these tutorials are a few minutes long and you’re encouraged to follow your curiosity and click on whatever brings you enlightenment or joy.
50 Claude Code tips to get you started (4 minute read)
Most people look at Claude Code and see an AI that writes code. That’s not wrong, but it’s incomplete. What Claude Code actually is, at its core, is a context engineering tool with a really smart agentic loop powering it. So here’s a Meta Staff Engineer’s full video walkthrough of 50 tips with live demos.
The feature backlog has gone *poof* (3 minute read)
Agentic coding tools like Claude Code are killing the feature backlog. There seems to be broad consensus in the Valley that something really changed in the last few months. The coding models got really good and engineers who have been leveraging these models to code are able to do more and more and more, with multiple agents running 24/7 to build features and products. That’s why the feature backlog on the product roadmap may no longer be a backlog for a set of companies. If your company isn’t already living in this new reality and realizing gains from agentic engineering, someone else competing against you is, and your product will be left behind.
Refer and we’ll send you our favorite books as a “thank you” for spreading the word.
Private investing
The greatest strategy game in the history of the world (2 minute read)
Next time you see a big AI headline, you’re witnessing one of the greatest strategic battles is playing out in real-time, with access to play-by-play updates. One of the reasons why it’s captivating is because the players are so powerful. The Mag 7 companies represent nearly a third of the stock market. These companies should be modeled as nation-states rather than corporate actors. The CEO’s the King/Queen, the board of directors is akin to parliament, shareholders represent the population, and employees play the role of the admin bureaucracy of the state. Each company has a unique game-winning strategy. And while it’s possible to see who is currently ahead, it’s difficult to forecast the end-game, and this is especially the case when those playing are grandmasters.
VC-backed startups are low status (9 minute read)
Before Tech became the dominant industry, jobs like investment banking (IB) were generated viewed as the best expected value career trade to make out of school in terms of money, training, and accrued social capital. As tech began to rise, fostered by cultural tenets of weirdness and meritocracy, IB within many younger circles went from high status to low status, signaling you’re smart but not that interesting. The mainstream VC-backed startup path is undergoing a similar transition. In many non-tech circles, people struggle to distinguish between another founder with another launch video building another (insert zeitgeist here) startup. The noise outpaces the signal as the system is optimized to produce as many founders as possible, as safely as possible. The legible thing becomes the average thing, and thus the average thing becomes the low-status thing.
Building an institutional-grade portfolio tracker with AI (10 minute read)
This guide will walk you through building one specific project - a portfolio tracker - but the real value is learning how to work with Claude. Once you get it, you’ll start seeing opportunities everywhere. “Wait, Claude could automate this.” “Claude could build me a tool for that.” It changes how you think about problems. Congratulations! You just built something that would cost you $24,000/year from Bloomberg, or $50-100/month from portfolio apps, or weeks of developer time if you hired someone.
Career management
OpenAI’s Sherwin Wu on why the productivity gap is widening, how his engineers now manage 10-20 parallel agents, and why you need to build for where models are heading. During his conversation with Lenny, Sherwin outlines how OpenAI cut PR code review times from 10-15 minutes to 2-3 minutes, the role of a SWE is shifting from writing code to managing fleets of agents, and audio will be the next frontier for multimodal AI. For this dialogue, you’ll want to be taking notes proactively.
The definitive guide to Forward Deployed Engineering (22 minute read)
If you work in tech, you should learn about Forward Deployed Engineering (FDE). It’s quickly become one of the most popular roles that nearly all fast-growing AI companies are looking for…at the same time, it’s the type of role very few people and companies seem to actually understand. FDE was originally popularized by Palantir (now a $275 billion public company) but it’s now become the most valuable credential in tech, and most people don’t understand why. So here’s the definitive guide by Vinoo Ganesh, the person who designed and built the program that turned software engineers into FDEs at Palantir (back in 2015).
Claude Code is the new Excel (5 minute read)
For several decades, Excel was the undisputed king of general-purpose knowledge work and became the lingua franca of finance and accounting before colonizing every domain. Claude Code is far more disruptive than Excel and will create a K-shaped divide along the fault line of AI literacy. While Excel automated calculations, agents can complete entire tasks end-to-end. You’re no longer one person working faster, you’re one person with an entire team. Every time Anthropic ships a new model, your agents will get better overnight. Excel never got smarter on its own. Not knowing Excel meant you were less competitive but not using AI agents means you’re almost bionically outmatched, since you’re competing with cybernetically augmented people.