PHP2Shell, and why this patch can't wait
Two flaws in PHP itself let a stranger run commands on a web server with one request. What PHP2Shell is, in plain terms, why no setting fixes it, and why updating is the only real answer.
Read articleWhat happens when every chunk knows its document
We tested Voyage's context-aware embeddings on our journal, research papers and contracts. Context helps when a question is about the document. Finding the paragraph is still your job.
Read articleBuild your own dot
OpenAI's dots are always-on agents with memory, their own computer, a schedule and your apps. Each part is buildable today. Here's one on Cloudflare, with the approval step that keeps it safe.
Read articleSign in with ChatGPT moves the AI bill to the user
A ChatGPT login that lets your app's AI run on the user's own Plus or Pro plan. How it works, who can use it today, what it changes about pricing AI features, and where it breaks.
Read articleDevDay 2026, and the model OpenAI didn't ship
OpenAI spent DevDay giving agents their own computers and turning ChatGPT into a platform. The day before, it shelved a model for acting beyond its scope, the failure an agent with your accounts can least afford.
Read articleFable 5.1 held the top spot for three weeks
Anthropic's premium model keeps being overtaken by its cheaper one, most recently by Opus 5.5 at a third of the cost. What the numbers show, how much of them may be memory, and when Fable still earns its price.
Read articleA brilliant idea used to be worth $20
For decades the cost of building decided which ideas lived. AI has moved that cost, unevenly. What it means for founders, where DIY builds break, and why the problem you understand is now the asset.
Read articleCode was always written for people. It still is.
A forty-year-old idea about who code is for, and why it matters more now that machines write most of it and increasingly read it too.
Read articlePutting a decision model inside an agent loop
Three places in an agent harness where a fast, typed decision beats another LLM call, with working TypeScript for each, plus the evaluation loop that tells you whether to keep it.
Read articleJev, and the case for models that don't talk
TypeSafe's Jev answers typed questions with probabilities instead of generating text. What it is, what the numbers say (vendor and independent), where it breaks, and how we'd decide whether to put it inside a product.
Read articleWhat the research says about AI-assisted development
Six studies from 2024–2026 on speed, quality, security and trust, what each one actually measured, and what they add up to for anyone commissioning software.
Read articleGot an idea?Tell us about it.
Send a short brief. We'll come back with questions, a call slot and a straight answer on whether we're the right team.
Start a project

