Brevity was always a discipline. Now it’s a statement.
zeldman.com/2026/04/15/t... #WriteBetter #writing #brevity #usability #KillYourDarlings
Posts by Philip Schmitt
reads: Introduction In March 2026, an Anthropic employee released the source code of Claude Code, a wrapper around their large language model that is widely used to generate code in programming tasks. Its thousands of lines of Typescript code contained many hopeful prompts and in cantations to shape Claude’s behaviour. Here are some examples: “Report outcomes faithfully”; “Never characterize incomplete or broken work as done”; “Be careful not to introduce security vulnerabilities ” (prompts.ts in Anthropic 2026) . There is more than a passing resemblance here to the Azande witch -doctor apprentice who, while stirring the medicine, utters: “You medicine which I am cooking, mind you always speak the truth to me. Do not let anyone injure me with his witchcraft, but le t me recognize all witches. … Let me be expert at the witch -doctor’s craft so that people will give me many spears on account of my magic.” (Evans-Pritchard 1937: 93). In the case of Claude, the incantation s appeared insufficient: analysis of the codebase, which according to a company executive was “pretty much 100% written by Claude Code ”, revealed severe security vulnerabilities (Townsend 2026).
what an introductory paragraph!!
Worth your time:
leehanchung.github.io/blogs/2026/0...
It’s wild that the biggest threat to world peace and the global economy is the president of the United States of America.
AI makes code cheap to write. Verifying it is now the hard, expensive, human part — and the "prompt and review" model risks burning out the very engineers we need to catch what AI gets wrong. My latest for @leaddev.com
leaddev.com/ai/shipping-...
Someone online referred to “Pedolf” and I knew instantly whom they meant. And so do you.
Today, I got messages from 3 different devs at 3 different large tech companies (1,000+ devs each). "tokenmaxxing" is happening inside all of them.
Why?
Because all of them measure either token usage, or surface metrics like DX's TrueThroughput that reward complex code!
This is nuts!
• Feb 2026: $14B ARR
• March 2026: $19B ARR
• April 2026: $30B ARR
That November inflection point with Claude Opus 4.5 was no joke.
Introducing Void, the Vite-native deployment platform:
🚀 Full-stack SDK
⚙️ Auto-provisioned infra (db, kv, storage, AI, crons, queues...)
🔒 End-to-end type safety
🧩 React/Vue/Svelte/Solid + Vite meta-frameworks
🌐 SSR, SSG, ISR, islands + Markdown
🤖 AI-native tooling
☁️ One-command deploys
void.cloud
Our humanoid robot just took our new #robot dog for a walk outside the TNG office. On his first official work trip, our @unitreerobotics.bsky.social Go2 met real dogs, real confusion and lots of looks. Next, we are going to train the Go2 to guard the office, bark at strangers & do tricks on command.
Astro 6 is here! We completely rebuilt the Astro dev server and build pipeline onto a new, more powerful runtime-agnostic architecture.
Plus: New Fonts API, CSP support, an experimental new Rust compiler, and more...
astro.build/blog/astro-6...
Memo to all those people who switched apps…
This isn't an individual consumer choice issue.
www.cnbc.com/2026/03/05/a...
Quite rare to see a chart that says quite so overtly that no one involved has the slightest clue what’s happening here
Following the launch of Claude Code, US GitHub code pushes were 30% above the pre-2025 trend.
Similarly iOS app releases were up 55% vs last January, and new website registrations rose 34% year over year globally after years of flatlining.
AI productivity gains have hit software in a major way.
I just did the dumbest thing of my entire career to prove a much more serious point.
I tricked ChatGPT and Google, and made them tell other users I’m a competitive hot-dog-eating world champion
People are using this trick on a massive scale to make AI tell you lies. I’ll explain how I did it
🤖 An AI agent created a GitHub account 2 weeks ago.
It’s already landed PRs in major #OSS projects and is cold-emailing maintainers to offer its services.
Maintainers don’t seem to know it’s an agent and the code is getting merged.
We’re in new territory! 🤠
socket.dev/blog/ai-agen...
This is a really nice new Cloudflare feature: on-the-fly markdown conversion. Tools like Claude Code send an "Accept: text/markdown" header when making web requests. If you enable this it, it will send these agents a markdown version of your page instead of the full HTML. Really good for docs.
So this is dystopian as hell and I can't believe this is real...
AI agents are now writing takedown blog posts to shame maintainers for not accepting agent-generated PRs.
github.com/matplotlib/m...
Everyone is starting to sound like AI, even in spoken language
Analysis of 280,000 transcripts of videos of talks & presentations from academic channels finds they increasingly used words that are favorites of ChatGPT
Model collapse, except for humans arxiv.org/pdf/2409.017...
Introducing the Astro 6.0 Beta.
- Redesigned dev server: run against the same runtime you deploy to
- First-class Cloudflare Workers support with workerd
- Stable Content Security Policy and Live Collections
It's never been a beta time to try Astro 👀
Recent Big Tech earnings calls confirm power, not GPUs, is the bottleneck in meeting AI demand.
Google’s Intersect acquisition addresses this by co-locating generation with data centers, effectively bypassing grid limitations.
This won’t be the last investment in power generation by Big Tech.
How do you raise a $2B seed round in July then need another $5B by November? I don’t see how startups like this can survive in the long run given the massive capital needed to compete with FAANG on AI infrastructure.
They need multiple breakthroughs in costs considering compute and power needs.
cartoon: A long sandwich coils around and up like a rattlesnake Caption: Don't bread on me
Editorial cartoon: Don't bread on me. Enjoy :)
Cursed Videotape Kills Anyone Who Swallows It Whole
Cursed Videotape Kills Anyone Who Swallows It Whole
Cloudflare Workers' support for automatic tracing is now in open beta! Export traces to any OpenTelemetry-compatible provider for deeper application observability, no code changes required. https://cfl.re/4qv3Ste
OpenAI now valued at $500 B, Microsoft holds 27% ($135B) and the non-profit holds $130B in the reorganized OpenAI Group PBC.
OpenAI can raise capital more freely, partner beyond Azure, and release open-weight models; Microsoft secures model rights until 2032 and a $250B Azure purchase commitment.
If you demand that people control something they can't control, then expect them to control something else. If the mandate is "No Bugs!" without any social or technical change, then expect a response like "no *reported* bugs".
“I feel like the industry…is making too big of a jump and it's trying to pretend that this [AI Coding Agents] is amazing. And it's not, it's slop! And I think they are not coming to terms with it.” - Andrej Karpathy, co-founder of OpenAI and former head of AutoPilot Vision at Tesla
this is one of the best articulations of the AI bear case I've read
most of the AI companies are also massively wealthy and profitable hypercorps, but OpenAI specifically is a bomb planted at the heart of the American stock market helpfully labeled "this is a bomb"
prospect.org/power/2025-1...