wayanjimmy
ENID

Newsletter #31

· 2026-W33 · 2026-08-10 to 2026-08-16

Why Go is an Ideal Language for AI-Assisted Software Engineering #

— Google Developers Blog

tl;dr: As AI shifts software engineering from writing to reviewing, Go’s strict compiler and unified toolchain make it unusually well-suited for AI-generated code. The compiler catches type errors before they ship, the toolchain (gofmt, go vet, tests) gives agents a tight feedback loop, and the language’s simplicity keeps generated code predictable. A strong argument for Go as a default language when agents write most of the code.

Go AIProgramming Google SoftwareEngineering AIAssistedDev #

Agentic Engine Optimization (AEO) #

— Addy Osmani

tl;dr: AI coding agents consume documentation fundamentally differently from humans — they scan, extract, and infer from structure rather than read. If you’re still optimizing docs only for human readers, you’re leaving a growing share of your “audience” (agents) underserved. Addy lays out how to structure docs, code comments, and APIs so agents can actually use them well. The doc-engineering counterpart to “how coding agents read your code.”

AEO AgenticAI Documentation DeveloperExperience AIAgents #

We replaced Redis with MySQL for inventory reservations—and it scaled #

— Shopify Engineering

tl;dr: Shopify replaced Redis with MySQL for inventory reservations and hit their scale targets using SKIP LOCKED, composite primary keys, and connection visibility. A masterclass in making an “ordinary” relational database handle a hot, contended workload — and a reminder that the boring choice often wins when you understand your access patterns.

MySQL Redis Shopify Databases BackendEngineering #

Use AI to make product decisions #

— Kent C. Dodds (YouTube)

tl;dr: Most people use agents to write code; Kent uses one to decide what to build. He walks through using his agent to evaluate Cloudflare’s “Computer” package: gather evidence, argue about primitives (“couldn’t an existing primitive handle this?”), check entitlements and cost, and peel cleanup into side-quest agents. The core lesson: agents are research partners, but you make the final call — and the product sense you demonstrate shapes what they return.

AIProduct ProductEngineering AIAgents DecisionMaking #

Etched Sohu vs NVIDIA: Transformer ASIC vs GPU #

— Spheron Blog

tl;dr: Etched’s Sohu claims 500k tokens/sec on Llama 70B from an 8-chip server. This breakdown compares what a transformer-only ASIC actually gives up versus H100, B200, and Groq’s LPU — the trade-offs of specialization in inference hardware, and why “ASIC beats GPU” isn’t the whole story.

AIHardware ASIC GPU Inference LLM Transformer #

celld: self-hosted, distributed Durable Objects #

— Flavio Copes / denoland

tl;dr: celld is Deno’s answer to Cloudflare Durable Objects that you can run on your own machines. Each “cell” gets its own SQLite database with S3-compatible storage, giving you distributed, stateful compute without the vendor lock-in. Worth reading if you’ve been eyeing Durable Objects but want self-hosting.

DistributedSystems DurableObjects Deno SelfHosting SQLite #

User agent strings to HTTP signatures — methods for AI agent identification #

— Arcjet

tl;dr: UA strings are dead as a way to identify AI agents — anyone can forge them. Arcjet walks through the evolution from user agents to HTTP message signatures: how to verify an agent’s identity cryptographically at the API boundary with TypeScript. The practical primer for building APIs that want to know which client is really calling.

AIAgents HTTPSignatures APIIdentification TypeScript Security #

Rise of the agents (in open source) #

— Sourcegraph

tl;dr: Sourcegraph charts how agentic development is reshaping open source — from AI-generated contributions to tooling that treats agents as first-class participants. A short, pointed thread on where open source is heading as agents do more of the work.

OpenSource AIAgents Sourcegraph AgenticDevelopment #

Subscribe