Reverse-engineering is cheap now
How coding agents have collapsed the cost of reverse-engineering and automating devices at home — tasks that previously weren't worth the effort are now trivial.
Who’s Afraid of Chinese Models?
Willison's take on Ben Thompson's argument about the hypocrisy of Western labs trying to restrict open Chinese models while relying on open research.
Quoting Sam Altman
A quote from Sam Altman on OpenAI's internal debate over open-source strategy, hinting at a possible near-term open-weights release.
AI Mania Is Eviscerating Global Decision-Making
Willison highlights Nik Suresh's biting essay on how AI mania is distorting organizational decision-making.
Claude Code uses Bun written in Rust now
Notes that Claude Code now ships the Rust rewrite of Bun, with a look at Jarred Sumner's reasoning and the debate it caused.
Voices · last 30 days
From engineers worth following
Latest posts from a curated roster — infra, devops, AI, backend.
The Pulse: Grok’s CLI caught uploading all your local files to the cloud
Grok's CLI was caught uploading all local files to the cloud, a serious privacy incident, plus other Big Tech notes.
Context engineering with Dex Horthy
An episode with Dex Horthy on context engineering — how to structure what you feed models for reliable agent performance.
What is “loop engineering?”
Explains 'loop engineering', the emerging practice of designing agent loops instead of writing one-off prompts, spotlighted by Anthropic and OpenAI leaders.
The Pulse: What can we learn from Bun’s rapid Rust rewrite with AI?
What Bun's rapid AI-assisted Rust rewrite teaches about the changing economics of large refactors.
The Pragmatic Engineer AMA
A reader AMA episode covering careers, Big Tech trends, and the state of engineering in the AI era.
The Archaeologist’s Copilot
On using AI copilots as 'archaeologists' to understand and safely change legacy systems.
DSLs Enable Reliable Use of LLMs
Argues domain-specific languages and strong abstractions act as a harness that guides LLMs to generate exactly what's intended.
Fragments: July 13
Notes from Thoughtworks' Future of Software Development retreat on how AI is reshaping the craft.
Experiences with local models for coding
Hands-on experiences using local models for coding, and where they hold up versus frontier cloud models.
Viability of local models for coding
The factors that determine whether local models are viable for coding work today.
The one thing every experienced engineer gets wrong
Why even strong engineers stumble on project management, and how that first big project became a rite of passage.
Theory of constraints, AI, and code review
Applies the theory of constraints to argue that code review, not code creation, is now the bottleneck in an AI-accelerated workflow.
Using vibes to rewrite my blog
A candid account of vibe-coding a full blog rewrite, and why knowing what you're doing still matters.
You get what you hire
The kind of engineers you hire shapes the kind of system you end up with — hire distributed-systems people, get a distributed system.
Have you heard? Clickhouse is winning the observability wars!
Amplifies Mat Duggan's argument that ClickHouse is winning the observability wars by making high-volume telemetry cheap to store and query.
In defense of AI mandates
A defense of top-down AI adoption mandates, drawn from Honeycomb's own AI journey and the tension between enthusiasts and skeptics.
Is it ethical to use AI?
Wrestles with the ethics of using AI tools that have real externalities, arguing for making AI boring and pragmatic again.
Kimi K3: The open-weights escalation
Analyzes Moonshot's 2.8T-parameter Kimi K3 and what its release means for the escalating open-weights race.
6 months to live for open models
Argues open models face their most serious viability test yet as enforcement and policy threats mount.
Latest open artifacts (#22): Zyphra, Cohere, and Poolside are expanding the breadth of the ecosystem
A roundup of recent open model releases from Zyphra, Cohere, and Poolside, showing a broadening ecosystem.
GLM-5.2 is the step change for open agents
Makes the case that GLM-5.2 is a step change for open agentic models, closing ground on closed labs.
Opaque, Interoperable Passkey Records (and a Go API)
Proposes an opaque, interoperable format for storing passkey records, with a Go API, framing passkeys as the principled fix for phishing.
Vulnerability Reports Are Not Special Anymore
Argues that, for open-source maintainers, vulnerability reports should be treated as gifts rather than obligations.
Cloudflare Internal DNS is now generally available
Announces general availability of Cloudflare Internal DNS, bringing managed private-network DNS onto Cloudflare's global control plane.
Cloudflare WAF protects WordPress applications from two high-severity vulnerabilities
Cloudflare deploys WAF protections for two high-severity WordPress vulnerabilities, including an unauthenticated RCE.
A broken DNSSEC rollover took down .al. Now 1.1.1.1 tells you when validation is bypassed
How a broken DNSSEC rollover took down Albania's .al domain, and how 1.1.1.1 now signals when validation is bypassed.
Introducing Precursor: detecting agentic behavior with continuous client-side signals
Introduces Precursor, which detects agentic and bot behavior using continuous client-side signals rather than one-off challenges.
Improving Smart Tiered Cache for Public Cloud Regions
Improvements to Smart Tiered Cache that pick the fastest upper-tier data center for origins hosted in public clouds.
Discord Patch Notes: July 7, 2026
A rundown of the more technical fixes recently shipped into Discord.
Discord is Now on Meta Quest: Reach Out to Your Servers While in VR
Discord launches natively on Meta Quest via the Horizon App Store, ending sideloading for VR users.
Cost Attribution in Discord’s API
How Discord tracks per-feature hosting costs across 1,700+ API endpoints and hundreds of Kubernetes deployments without restructuring.
Discord Update: June 25, 2026 Changelog
The June 25 changelog covering recent Discord app updates.