LLM experiment: "Everyone's focused on AI that writes code. You know the usual suspects, Claude Code, Cursor, Copilot, all that. But that felt like missing the bigger picture. So I built something to test a different question: what if you skip code generation entirely? A web server with zero application code. No routes, no controllers, no business logic. Just an HTTP server that asks an LLM "what should I do?" for every request. The goal: prove how far away we really are from that future." And the interesting thing is it actually worked: "But despite all that, forms actually submitted correctly. Data persisted across restarts. The UI was usable. APIs returned valid JSON. User feedback got implemented. The AI invented, without any examples, sensible database schemas with proper types and indexes, parameterized SQL queries that were safe from injection, REST-ish API conventions, responsive Bootstrap layouts, form validation, and error handling for edge cases. All emergent behavio[u]r from giving it three tools and a prompt." It's not fast, it's not cheap, but certainly gives a hint at the future. Will LLMs become computers?

LLM FutureLove this LLM experiment: "Everyone's focused on AI that writes code. You know the usual suspects, Claude Code, Cursor, Copilot, all that. But that felt like missing the bigger picture. So I built something to test a different question: what if you skip code generation entirely? A web server with zero application code. No routes, no controllers, no business logic. Just an HTTP server that asks an LLM "what should I do?" for every request. The goal: prove how far away we really are from that future." And the interesting thing is it actually worked: "But despite all that, forms actually submitted correctly. Data persisted across restarts. The UI was usable. APIs returned valid JSON. User feedback got implemented. The AI invented, without any examples, sensible database schemas with proper types and indexes, parameterized SQL queries that were safe from injection, REST-ish API conventions, responsive Bootstrap layouts, form validation, and error handling for edge cases. All emergent behavio[u]r from giving it three tools and a prompt." It's not fast, it's not cheap, but certainly gives a hint at the future. Will LLMs become computers?

Software ArchitectureOne for the hardcore techies today 😁 If there is one single thing that any software developer should learn it's in my opinion the difference between designing a library vs designing a program. Their design and their architectures are completely different. The reason I think it is so extremely important is the fact that API's in particular have the visual traits of a library, but 99% of its use cases are for programs. This leads to extreme over complication of simple problems and one of the most bastardized practises in software: dependency injection. While it should be the art of abstraction it seems te be the Olympics of injecting everything into everything else. The majority of API's that I see do the following: Validate -> Transform -> Persist or Read -> Transform. Something that best can be represented as functional chaining with very limited requirements for dependency injection. Anyways tomorrow I am going to talk to a team about the sexy topic of pure functions and why understanding them is the key to understanding dependency inversion and simplifying your architecture. And most importantly, how to build programs instead of libraries. A topic anyone else cares about?

Hello everyone! I am looking for a new position and would appreciate your help. If you hear about any options or just want to talk to me, send me a message or leave a comment below. I look forward to hearing from you. hashtag#OpenToWork About me and what I'm looking for: 💼 I am looking for positions such as Technical Tester, Software Developer, Quality Assurance, Teacher and Tester. 🌎 I am open for positions in Greater Copenhagen ⭐ I have previously worked at the Danish Agency for Development and Simplification, nemlig.com, Q Nation, teacher at KEA, e-boks and others.

Wir suchen ab sofort eine:n erfahrene:n Test Engineer:in, um unsere Vision der offenen Interoperabilität im Gesundheitswesen Realität werden zu lassen! Folgende Tätigkeiten warten auf dich: Als Test Engineer:in übernimmst du bei uns die zentrale Rolle in der Qualitätssicherung unserer webbasierten Softwareprodukte im Gesundheitswesen. Du baust unsere Teststrategie von Grund auf auf – mit dem Fokus auf moderne Webtechnologien (React, TypeScript), End-to-End-Testautomatisierung und die Validierung von FHIR-basierten Datenflüssen innerhalb unseres Clinical Data Repository (CDR). In dieser Rolle arbeitest du eng mit Entwicklung, Produktmanagement und – wo nötig – auch direkt mit Kunden zusammen. Im Klartext erwarten dich folgende Aufgaben mit den Schwerpunkten Teststrategie, End-to-End-Automatisierung und Qualitätssicherung im Gesundheitsumfeld: Konzeption und Aufbau einer QA-Architektur für eine moderne FHIR-basierte Gesundheitsplattform Entwicklung und Implementierung von automatisierten End-to-End-Tests (z. B. mit Cypress oder Playwright) für Web-UIs und APIs Validierung komplexer Datenflüsse im Kontext unseres CDR unter Berücksichtigung des HL7 FHIR Standards Einrichtung und Pflege von Tests in unsere CI/CD-Pipelines (z. B. GitLab CI) Erstellung und Pflege von Testfällen für funktionale und nicht-funktionale Anforderungen Dokumentation von Testergebnissen, systematisches Bug-Tracking und Ableitung von Optimierungsmaßnahmen Zusammenarbeit mit Entwicklern, Produktmanagement und ggf. Kunden zur Anforderungsabstimmung und Qualitätssicherung Was dich auszeichnet Du bringst fundierte Erfahrung in der Testautomatisierung mit – idealerweise inklusive Aufbau einer QA-Struktur zusammen mit dem Entwicklungsteam On top dazu: Technisches Know-how in Webentwicklung (React) und Backend-Architekturen (Node.js, TypeScript, Docker) Kenntnisse im Testen von APIs (REST, idealerweise FHIR-Profile) und datenintensiven Anwendungen (z. B. CDRs) bringst du außerdem mit. Mit modernen Testframeworks wie Cypress, Playwright, Vitest, o. ä. bist du vertraut Du arbeitest eigenverantwortlich, strukturiert und mit einem hohen Qualitätsbewusstsein Deine Sprachkenntnisse sind im Deutschen und Englischen Bereich sehr gut in Wort und Schrift Nice-to-have wären für uns folgende Kenntnisse Praxiserfahrung mit dem HL7 FHIR Standard und Verständnis für klinische Datenstrukturen Erfahrung mit Testfällen in KI-basierten Modulen oder Algorithmen Kenntnisse im Umgang mit Testdatenmanagement, Mocking und Testkonzepten im Healthcare-Bereich ISTQB-Zertifizierung oder vergleichbare Nachweise Das macht uns aus Technologischer Stack: React, TypeScript, Node.js, KI-Komponenten – mit Fokus auf HL7 FHIR Spannende Projekte im Umfeld digitaler Gesundheitslösungen und klinischer Datenhaltung Enges, interdisziplinäres Team mit kurzen Entscheidungswegen Flexible Arbeitszeiten, Remote-Option und individuelle Weiterbildungsmöglichkeiten Dein Weg zu uns Wenn du den Aufbau nachhaltiger Qualitätssicherung in einem stark regulierten, technologisch anspruchsvollen Umfeld als reizvolle Herausforderung siehst, dann bewirb dich jetzt unter folgendem LINK!

The Global Dropdown Dilemma You have seen it. You go to sign up for a company’s software. The company is based in the United States. Ninety-nine point nine percent of its customers are also in the United States. Yet when you reach the field that says “Select your country,” you are greeted by a list that begins with Afghanistan, the Cook Islands, and Antarctica. You scroll for many seconds before finding the United States somewhere in the end of the alphabet, as though you are registering from a research station or a coconut shack in the South Pacific. It is absurd. It is not global awareness. It is bad design. Even worse, it is data blindness. Companies pretend that a global drop-down list somehow signals inclusivity or technical sophistication. In reality, it shows that no one in the product chain has ever looked at their own analytics. If your revenue is overwhelmingly American, why is the United States hidden behind a list that begins with countries that will never buy your software? Why not start with your top three? These are almost always English-speaking countries such as the United States, the United Kingdom, and Canada. Sometimes Germany or Australia. But certainly not Antarctica, which has no permanent residents at all. Some people joke that this is part of a “Marxist psyops” to undermine American exceptionalism. Maybe not. But if it were, it would be working, one sign-up form at a time. Still, the real cause is simpler and less interesting. It is ignorance. Or indifference. Somewhere, a designer copied an ISO country list and never looked back. Companies, if this is how you are collecting your metrics, you have work to do. Prioritize your users. Sort your lists by usage, not by alphabet. A sign-up form should not feel like a geography quiz. The world will not think less of you for making it easy for your actual customers to find themselves. I am Jorge de la Torre, a GRC professional at Fortress Solutions Group. I put the C in GRC for Compliance.

То что материя обладает необычными свойставами по-сравнении с не-материей спорить бесполезно да и вредно для горла. Горло это первое, что падает жертвой конфликта материи и не-материи. Падающие горла изображены на многих фамильных гербах и флагах не-материи. Надо ли говорить, что не-материя несет большие не-материальные потери в это конфликте. Не-материю перестали расщеплять в коллайдерах, убрали из табеля о рангах, забрали паспорт. Не-материю не узнать - твердят лодыри ученые, кидаясь пустыми бутылками. Не-материя перестала за собой следить, складки уже долго не разглаживаются и в них неконтролируемо растут хнехни. А хнехни это последние, что хотел бы увидеть продавец не-материи Овидий.

I just read this line from X’s latest update: “Grok will literally read every post and watch every video to match users with content they’re most likely to find interesting.” And then: “We’re aiming for deletion of all heuristics within 4 to 6 weeks.” On paper, that sounds like progress. In reality, it sounds like the moment we hand the keys over completely. When a company says it’s “deleting heuristics,” it means it’s removing the parts of its system that humans still understand. The parts you can explain, debug, or hold accountable. What replaces them is a black box that feels personal, because it knows you better than you know yourself, but ultimately serves goals you can’t see. And “Grok will read everything”? That’s not personalization. That’s total observation, framed as convenience. We keep being told this evolution is inevitable. But maybe the real innovation would be to slow down long enough to ask: At what point does personalization become prediction, and prediction become control? Because if every post, every thought, every frame becomes input for “Grok,” then maybe the system isn’t learning what we like. Maybe it’s teaching us what to like.