Read The Pragmatic Engineer with Scottie
Keep The Pragmatic Engineer. Lose the inbox pileup.
Scottie reads the newsletters you choose, keeps the useful differences, and attaches every original link. The Pragmatic Engineer stays in your reading life without becoming another email to manage.
An independent guide. Scottie isn't affiliated with, endorsed by, or sponsored by The Pragmatic Engineer. The publication owns its name, writing, and subscription terms.
From Scottie · July 31, 2026
See The Pragmatic Engineer in a Scottie brief
Scottie read The Pragmatic Engineer, ByteByteGo Newsletter, and Level Up Newsletter for a reader with the priorities shown below. Start with the rundown, open the full brief, or check every issue behind it.
3sources
7issues read
5stories included
2 of 2 The Pragmatic Engineer issues included
What shaped this brief
Reader priorities
A software engineer or manager who wants reported lessons about teams, tools, delivery, and how AI is changing the profession.
These are illustrative priorities, not a customer’s data.
Sources in this brief
- The Pragmatic EngineerThe publication this guide is about
- ByteByteGo NewsletterAdds concise system-design and architecture explanations.
- Level Up NewsletterAdds focused coaching on influence, leadership, and team development.
Scottie
Scottie example brief
July 31, 2026 · Executive brief
01 / The rundown
- Anthropic is restructuring software engineering teams around small two-person pods while AI tooling shifts developer time toward verifying generated code.
- Food delivery platforms are adding LLMs to production search by adapting existing knowledge graphs and vector towers rather than replacing legacy infrastructure.
Read the complete brief 5 stories · 2 action items
02 / The briefing
01 / main
Anthropic shifts workflows with AI tools and two-engineer project caps
The takeaway: AI tools are reshaping software delivery at Anthropic. Teams cap projects at two engineers, make prototyping more fluid, and spend more time verifying AI-generated output than writing code, while creator Jarred Sumner completed a 500,000-line Bun rewrite in 11 days.
Concrete details
- Bun creator migrated a 535,496-line codebase from Zig to Rust in 11 days using $165,000 in AI tokens.
- Anthropic spent six months architecting Claude Managed Agents after running a proof-of-concept for Claude Code.
Why it matters for this reader: As a software engineering manager, these workflow patterns show how AI-native teams structure small pods, automate language migrations, and shift developer focus from coding to system verification.
Original sourcesThe Pragmatic Engineer
02 / main
AI labs target lower task costs across agent loop architectures
The takeaway: Leading AI labs are competing on task efficiency rather than pure model capabilities. Recent releases like GPT-5.6 Sol focus on optimizing harness, API, and inference layers to lower the total cost per successful task completion.
Concrete details
- GPT-5.6 Sol with maximum reasoning outperforms Fable 5 on the Artificial Analysis Coding Agent Index at lower cost.
- Serverless fine-tuning platforms charge per processed token instead of billing for entire GPU machine hours.
Why it matters for this reader: Knowing that AI labs prioritize per-task execution costs helps you evaluate model choices based on operational budgets rather than raw benchmark scores alone.
Original sourcesByteByteGo Newsletter
03 / main
How food delivery apps rearchitected search with LLM integrations
The takeaway: DoorDash, Instacart, and Uber Eats integrated LLMs into search differently based on existing infrastructure. DoorDash enriches an offline knowledge graph, Instacart fine-tunes Llama-3-8B for tail queries, and Uber Eats uses a Qwen-powered two-tower embedding space.
Concrete details
- Instacart expanded query rewrite coverage from 50% to over 95% while keeping tail query latency under 300 milliseconds.
- DoorDash achieved a 30% lift in dish carousel triggers by using RAG to constrain LLM outputs to graph concepts.
Why it matters for this reader: These architectural patterns give your team concrete blueprints for inserting LLMs into production retrieval pipelines without tearing down legacy infrastructure.
Original sourcesByteByteGo Newsletter
04 / main
Solving request retry ambiguity with idempotency keys and deduplication
The takeaway: Network timeouts make it impossible to tell if a payment request failed or succeeded on the receiver. Implementing idempotency keys and clear delivery semantics ensures duplicate requests update state safely without double-charging customers.
Concrete details
- Idempotent operations produce identical system state whether executed once or repeated ten times.
- Deduplication schemes require defined time limits to bound state storage across producers, brokers, and consumers.
Why it matters for this reader: For your engineering team's transaction pipelines, designing idempotent endpoints prevents duplicate operations when network drops or client retries trigger redundant API calls.
Original sourcesByteByteGo Newsletter
05 / main
Why property-based testing beats full formal verification for most teams
The takeaway: In a conversation on formal methods, Hillel Wayne explains that while tools like TLA+ spot complex concurrency bugs, property-based testing offers the most practical path for most teams to catch subtle edge cases without writing complex specifications.
Concrete details
- AWS discovered a 35-step concurrency bug using TLA+ that conventional design and code reviews completely missed.
- Property-based testing evaluates thousands of generated inputs against defined system properties to uncover edge cases.
Why it matters for this reader: For your delivery quality strategy, adopting property-based testing provides high bug-finding coverage without requiring your team to master mathematical specification languages.
Original sourcesThe Pragmatic Engineer
Action items
- Audit payment and transaction endpoints this week to ensure idempotency keys safely handle client request retries.
- Evaluate property-based testing tools for your team's core logic to catch complex edge cases without formal specifications.
See every issue behind this brief
How Building Software is Changing at Anthropic
- How building software is changing at AnthropicThe Pragmatic Engineer · Included
Engineering Guide to Idempotency and Delivery Semantics
- A Detailed Guide to Idempotency, Delivery Semantics, and DeduplicationByteByteGo Newsletter · Included
Optimizing ChatGPT's Agent Loop Across Harness, API, and Inference
- How ChatGPT Optimizes its Agent Loop: Harness, API, and InferenceByteByteGo Newsletter · Included
Comparing LLM Search Integrations Across Delivery Platforms
- Why DoorDash, Instacart, and Uber Eats Integrated LLMs Into Search Three Different WaysByteByteGo Newsletter · Included
NVIDIA open models
- How NVIDIA Builds Open Models for the Age of AIByteByteGo Newsletter · Read, not included
Formal Methods and Testing with Hillel Wayne
- Formal methods with Hillel WayneThe Pragmatic Engineer · Included
Executive coaching and leadership
- "Am I Just a Loser?"Level Up Newsletter · Read, not included
About The Pragmatic Engineer
Should you add The Pragmatic Engineer to Scottie?
Make better engineering and management decisions using evidence from real software organizations.[1][2][3]
Who it’s for
A software engineer or manager who wants reported lessons about teams, tools, delivery, and how AI is changing the profession.[1][2]
What you’ll find in it
The issues linked below include “When AI writes almost all code, what happens to software engineering?”, “Formal methods with Hillel Wayne”, and “How building software is changing at Anthropic”. Open them to judge the publication in its own words.
- Byline
- Gergely Orosz[1]
How to read The Pragmatic Engineer with Scottie
Add its public feed for publicly available issues. If you subscribe to paid issues, forward those emails to your private @scottie.fyi address. Scottie does not need access to your inbox.
Scottie read 2 The Pragmatic Engineer items and included 2 for the brief. The sources covered different things, so Scottie kept their stories separate instead of forcing a connection.
Read three issues from The Pragmatic Engineer
Read the publication in its own words. Scottie keeps these original links attached; it does not replace the writing.
- When AI writes almost all code, what happens to software engineering?Public issue[2]
- Formal methods with Hillel WaynePublic issue[3]
- How building software is changing at AnthropicTue, 28 Jul 2026[4]