Software Lead Weekly 
Issue #723, 2nd October 2026
This Week's Favorite
Simplification Is a Loan
5 minutes read.
Hew Suber noticed that a whole category of controls never came back after Pinterest automated campaign setup: "People conflate wanting to control something with wanting to understand it, and when the understanding arrives the control stops mattering." -- This is the playbook for agentic products too. Before you add another knob for your AI feature, ask if people want the knob or just a clear explanation of what the system did and why.
Read it later via Instapaper.
Share it via Twitter or email.
5 minutes read.
Hew Suber noticed that a whole category of controls never came back after Pinterest automated campaign setup: "People conflate wanting to control something with wanting to understand it, and when the understanding arrives the control stops mattering." -- This is the playbook for agentic products too. Before you add another knob for your AI feature, ask if people want the knob or just a clear explanation of what the system did and why.
Read it later via Instapaper.
Share it via Twitter or email.
Culture
Most Companies Adding AI to Their Products
1 minutes read.
My humble effort to help you start the weekend with a smile.
Read it later via Instapaper.
Share it via Twitter or email.
1 minutes read.
My humble effort to help you start the weekend with a smile.
Read it later via Instapaper.
Share it via Twitter or email.
How to Pivot Your SaaS Company in the Age of AI
6 minutes read.
Jon Yongfook offers 7 areas where vibe coding doesn't compete with you as the value moves up the abstraction levels. Hardware, hardcore foundations (email, hosting, etc.), and underserved audiences (plumbers, teachers, etc.) are great areas to explore for value today. I like most the nuance of building for agents-first: We will seek a way to go deeper and understand what happened and why, given we're responsible for the systems we operate and the value we provide. This might require some special UI optimized for humans. Most often, though, when things go well, I can guide my favorite AI app (or coding tool) to keep me updated on what's going on.
Read it later via Instapaper.
Share it via Twitter or email.
6 minutes read.
Jon Yongfook offers 7 areas where vibe coding doesn't compete with you as the value moves up the abstraction levels. Hardware, hardcore foundations (email, hosting, etc.), and underserved audiences (plumbers, teachers, etc.) are great areas to explore for value today. I like most the nuance of building for agents-first: We will seek a way to go deeper and understand what happened and why, given we're responsible for the systems we operate and the value we provide. This might require some special UI optimized for humans. Most often, though, when things go well, I can guide my favorite AI app (or coding tool) to keep me updated on what's going on.
Read it later via Instapaper.
Share it via Twitter or email.
Harness Engineering Is Not Enough: Why Software Factories Fail (Video)
19 minutes read.
Dex Horthy puts it well: "If you're drowning in PRs, you actually have too many bad PRs. Because a good PR is a joy to review." -- I agree, and the fix is upstream, not in more review bots. His team spends about 30 minutes on product, architecture and program design (types, method signatures, call graphs) before the agent writes code. We all like faster output, but alignment up front is what keeps code review fast. Slow down where it matters. You get no points for slop work.
Read it later via Instapaper.
Share it via Twitter or email.
19 minutes read.
Dex Horthy puts it well: "If you're drowning in PRs, you actually have too many bad PRs. Because a good PR is a joy to review." -- I agree, and the fix is upstream, not in more review bots. His team spends about 30 minutes on product, architecture and program design (types, method signatures, call graphs) before the agent writes code. We all like faster output, but alignment up front is what keeps code review fast. Slow down where it matters. You get no points for slop work.
Read it later via Instapaper.
Share it via Twitter or email.
Thinking in Systems, Shipping in Loops
3 minutes read.
Tomasz Tunguz uses Donella Meadows' lens (resilience, self-organization, hierarchy) to describe great agentic systems. The upside is that we might get a tier of really amazing software engineers to pay attention to business initiatives, learning more about the business and product than they could before.
Read it later via Instapaper.
Share it via Twitter or email.
3 minutes read.
Tomasz Tunguz uses Donella Meadows' lens (resilience, self-organization, hierarchy) to describe great agentic systems. The upside is that we might get a tier of really amazing software engineers to pay attention to business initiatives, learning more about the business and product than they could before.
Read it later via Instapaper.
Share it via Twitter or email.
Peopleware
The Manager’s Path in the Age of AI
10 minutes read.
"We should be spending some of our productivity gains on things that seem like slowing down: meetings, design reviews, conversations. These ways of maintaining alignment, sharing knowledge, and creating human connections are, in my opinion, more important than ever to creating good outcomes at scale." -- I like the idea of investing time in areas that feel like slowing down that Camille Fournier is suggesting, assuming there is leverage there for insights to pick up based on observations around bottlenecks: ownership of production, system understanding, value protection, value creation, incentives alignment, etc.
Read it later via Instapaper.
Share it via Twitter or email.
10 minutes read.
"We should be spending some of our productivity gains on things that seem like slowing down: meetings, design reviews, conversations. These ways of maintaining alignment, sharing knowledge, and creating human connections are, in my opinion, more important than ever to creating good outcomes at scale." -- I like the idea of investing time in areas that feel like slowing down that Camille Fournier is suggesting, assuming there is leverage there for insights to pick up based on observations around bottlenecks: ownership of production, system understanding, value protection, value creation, incentives alignment, etc.
Read it later via Instapaper.
Share it via Twitter or email.
My Favorite Paul Graham Essay “How to Do Great Work” in Under 200 Seconds of Video
3 minutes read.
I'm sharing this not only because Paul Graham's essay is great, but also because the video version is wonderful. Even better, you can easily steal the prompt Deedy shared and try it on content you like if you think learning from video would help you understand and remember the material. Consider also asking Claude to create a website and a system to help you practice your memory via Spaced Repetition.
Read it later via Instapaper.
Share it via Twitter or email.
3 minutes read.
I'm sharing this not only because Paul Graham's essay is great, but also because the video version is wonderful. Even better, you can easily steal the prompt Deedy shared and try it on content you like if you think learning from video would help you understand and remember the material. Consider also asking Claude to create a website and a system to help you practice your memory via Spaced Repetition.
Read it later via Instapaper.
Share it via Twitter or email.
Building an AI Hive Mind to Understand Systems as a Manager
13 minutes read.
Every engineering manager is currently trying to figure out a system to stay connected to the business, product, architecture, and code. Great take by Candost Dagdeviren that you can copy a few ideas from: "The main goal I established for myself was to eliminate the “gather 3-4 people in a room just to figure out how the system works” stage: find out how the applications, services, tools, and infrastructure connect together and combine it with the deterministic knowledge: the codebase and team structures. As a secondary goal, I wanted to build a mechanism to surface tacit knowledge—knowledge that’s not written in a document, shared, or made explicit."
Read it later via Instapaper.
Share it via Twitter or email.
13 minutes read.
Every engineering manager is currently trying to figure out a system to stay connected to the business, product, architecture, and code. Great take by Candost Dagdeviren that you can copy a few ideas from: "The main goal I established for myself was to eliminate the “gather 3-4 people in a room just to figure out how the system works” stage: find out how the applications, services, tools, and infrastructure connect together and combine it with the deterministic knowledge: the codebase and team structures. As a secondary goal, I wanted to build a mechanism to surface tacit knowledge—knowledge that’s not written in a document, shared, or made explicit."
Read it later via Instapaper.
Share it via Twitter or email.
Inspiring Tweets
- Oren
P.S. Can you share this email? I'd love for more people to experiment and improve their company's culture.
P.S. Can you share this email? I'd love for more people to experiment and improve their company's culture.