Issue #667, 5th September 2025

This Week's Favorite


Developing a Powerful How-to-Win: From Pixels to Portrait
7 minutes read.

"the job of strategy is to paint a portrait, not to spew out an unconnected, if not random, array of pixels." -- Roger Martin's iterations on How To Win when directing your business are always a fascinating read: Drawing a bigger picture that you can zoom in and zoom out to explore critical pillars while connecting the dots (or pixels) to a meaningful outcome and story.

Read it later via Pocket or Instapaper.
Share it via Twitter or email.


Culture


The Age I Consider Old vs. My Age
1 minutes read.

My humble effort to help you start the weekend with a smile on your face. I’m 41 y/o and this graph feels spot on.

Read it later via Pocket or Instapaper.
Share it via Twitter or email.


The Post-Developer Era
8 minutes read.

It's hard to predict how (and when) the economy will shift into completely different jobs. Like an advanced version of the Turing Test, once Job Descriptions are posted and filled, and you won't be able to notice if it's a human doing the job or a machine (aka Agent), things will really change. We're not there yet. Practice the tools to develop your own opinion and narrative, setting your own path. Those who are brave and build strong fundamentals (seeking truth in how things work) will prosper, as they'll find new leverage points.

Read it later via Pocket or Instapaper.
Share it via Twitter or email.


5 Essential Things Every Leader Should Evaluate After a Reorganization
8 minutes read.

Many companies these days are undergoing reorgs, given changes in the economy - both the implications of macroeconomic changes (e.g. tariffs) and new GenAI technologies - having Allison McMillan post as a guideline to follow up on can be extremely helpful. Reorgs should serve a purpose to align the team around a new mission with a clearer focus, as Allison writes it: "Here's what I don't see as often... Actually getting people together to make this progress."

Read it later via Pocket or Instapaper.
Share it via Twitter or email.


AI Will Change How We Build Startups -- But How?
7 minutes read.

As long as users and buyers are humans, software distribution is both limited in capacity (how many apps should one use at work or for personal needs to get a job done?) and in emotional taste (humans tend to buy or install things beyond just numbers): "Obviously, some infrastructure layers, like foundation models, are very CapEx heavy. And theoretically, some apps _should_ be very easy to build. But growth and distribution cost a lot of money to get into customers' hands. One thing we've learned over the past decade is that even though it's relatively cheap to build a web app, acquiring users can still cost millions." We're still a tiny ant in the universe, and what I'm most excited about are space companies, pushing the envelope beyond another web or mobile app. Progress breach boundaries, expanding what we thought was even possible. So even if some jobs are going away, I'm thrilled to imagine a world in 20 years where engineers, product managers, or designers will define new ecosystems. Perhaps the "new mobile" or the "new internet" is space or even home management (humanoid robots). Builders and sellers will always push the economy to new levels.

Read it later via Pocket or Instapaper.
Share it via Twitter or email.


Peopleware


The Senior Developer LLM Babysitter Problem
5 minutes read.

A powerful observation of where we stand in September 2025: "This experiment made me think about a bigger risk: turning senior developers into babysitters for junior-level AI. LLMs can be clever, but easy doesn’t always mean fast. Offloading problems can feel like progress, yet often wastes time. The data echoes this. In the 2025 Stack Overflow Survey, 84% of developers said they use or plan to use AI tools, but nearly half distrust their accuracy (46% vs. 33% who trust them). Two-thirds cited “almost right, but not quite” answers as their biggest frustration, and 45% said debugging AI code takes longer.2 Google’s 2024 DORA report found only 24% of developers trust AI output “a lot,” with many worried about long-term software quality. In short: instead of taking work off our plate, used poorly, the tools can turn programmers into full-time babysitters." Perhaps we'll laugh at it in a few years or, more likely(?), a few weeks. I hope so.

Read it later via Pocket or Instapaper.
Share it via Twitter or email.


How to Digest 36 Weekly Podcasts Without Spending 36 Hours Listening (Video)
35 minutes read.

People like Tomasz Tunguz are a great example of finding new leverage points to improve their job performance, rather than questioning whether their role (in this case, as an investor) is needed.

Read it later via Pocket or Instapaper.
Share it via Twitter or email.


The Real Story in AI Right Now Isn’t Just Agents. It’s Agents + Deterministic Workflows. (Thread)
4 minutes read.

Determining ways to achieve deterministic outcomes (whether agents are involved or not) is a fundamental requirement when using software. Most people don't care about the magic behind the tools they use. Just Google "what happens when you search [http://google.com](http://google.com "smartCard-inline") into a browser" and read the responses to understand the complexity behind it. People care about themselves. Wade Foster is spot on here, and LLMs will make our jobs harder in that sense, as the models are not deterministic by nature, so our architecture around them will need to shift.

Read it later via Pocket or Instapaper.
Share it via Twitter or email.


Inspiring Tweets


@hnshah: Product is half the equation. Distribution is the other half. Ignore it, and you’ll spend years building something nobody wants.

@thejustinwelsh: The richest people I know are miserable. The happiest people I know figured out what "enough" looks like and optimized for it.

- Oren

P.S. Can you share this email? I'd love for more people to experiment and improve their company's culture.

Subscribe now & join our community!