A connected library of essays, sacred texts, career tools and working technology experiments — written and built by Paddy Iyer.
The changes worth knowing about — not every commit. New essays of note, sacred-text work, Interview Studio and JobSignal features, demos, corrections, and changes to privacy and security.
Last updated
Follow the changelog (RSS) Corrections log
Five journeys (Read, Learn, Prepare, Find, Build) now organise the site; a single search covers essays, sacred texts down to the verse, Interview Studio, companies and demos; the Atlas shows how they connect; and the homepage can pick up where you left off — from this browser's own history, never a server.
Every sacred-text app now has a Sources & Verification panel that says honestly where its text came from — including when that is not recorded. A public corrections log lists meaningful fixes. Every public statistic is now derived from the content itself, which corrected the question, data-model and sacred-text counts that had drifted.
Pages whose privacy promises went further than the code — claiming nothing was ever sent while loading analytics — were corrected; the Privacy Console gained a step-by-step data-flow section and dropped its tracking pixel; the analytics server now honours Global Privacy Control for every request. A plain-language privacy policy, terms, disclaimer and copyright page were added.
The site's API now only answers the site's own pages, compares the admin password in constant time and locks out repeated wrong guesses, meters the privacy scan proxy and re-checks every redirect it follows, and deletes expired leaderboard data daily. CI now scans for committed secrets and audits dependencies.
Search is now a proper keyboard- and screen-reader-accessible dialog; every page gets a visible focus ring, a skip link and reduced motion on request; sacred-text search boxes are labelled. An automated WCAG 2.2 check runs on every change and may not get worse.
A job search whose promise is that every role is still live on the employer's own careers page, re-checked every four hours. Reposts are labelled as reposts with their history, never as newly posted, and search keeps to the role you asked for.
About 9% of roles showed a place that does not exist. Multi-location postings now list every place, and each city has one spelling.
Seventeen verses gained word-by-word breakdowns, interpretation and cross-links to the Lalitha Sahasranama. Inherited English meanings found to be misaligned are labelled as under review rather than presented as settled.
A story in nine scenes about how professional discovery is changing when an agent, not a recruiter, reads your profile first.
The AI Executive Command Center, Revenue Leakage & Margin Intelligence Hub and Enterprise Resilience War Room join the Lakehouse demo — each on clearly labelled synthetic data.
A visual field guide to the 2026 hiring market: how to tell whether a role is real, who is behind it, and how to get from 'Apply' to a conversation.
A modelled look at sixty years of Bay Area housing against income, rebuilt after a source-by-source review; the article carries its own dated corrections note.
A new series, one song per article: Telugu text and transliteration, word-by-word meanings, ragam and talam where sources agree (and the disagreement where they do not), and curated listening.
A field guide from 'why is my job stuck?' to root cause — then hardened after review, which corrected an inconsistency about Adaptive Query Execution.
Questions filed under a topic the importer did not recognise were being silently skipped; they are now routed and published.
A working tool that finds where your personal information is published, checks each result is really you, and walks the removal through — with your profile kept in your own browser — alongside a local command-line agent.
A contact form, and testimonials that are never published automatically: every one is reviewed first, and only approved ones appear.
Workplace English and communication practice for interviews and the job, and an AI engineering track from data engineering to LLM-native systems.
Every essay, as it is published, is in the essay feed. Interview Studio's daily question additions are on its own What’s New page.
