标签：#development

By Vilius Vystartas This is the follow-up to What Actually Changed in Two Weeks . That one was about setting up a project for AI-speed delivery. This one is about something I keep re-learning on every fast delivery. You start shipping faster with AI. The code works, the feature lands, it feels good. Then a few weeks later the feedback comes back, and some of it catches you off guard. Not because anything is broken — but because a few things that seemed obvious to you weren't obvious to the other side. No drama. It happens. Here are five things I'm learning to check earlier. 1. What does "done" look like from their side? To me, done means working software. To someone else it might mean pixel-match with a design. Both are valid. What helps: A quick "what does good enough look like to you?" before the work starts. One sentence can save a lot of back and forth. 2. When will they actually look at it? Sending something doesn't mean it gets reviewed immediately. It lands in a queue like everything else. What helps: Naming a review date alongside the delivery date. "I'll share this Tuesday — could you take a look by Friday?" Turns silence from a mystery into a signal. 3. What needs to be perfect vs what can be improved later? Not everything in the feedback is the same weight. A label change and a broken flow are different things. Without saying so upfront, everything looks like an emergency. What helps: Two buckets agreed early. "Here's what I'll get right before it ships. Here's what I'd revisit in a follow-up." Makes the first feedback session more productive. 4. Could they see something before the full delivery? The first time someone sees your work often sets the tone. Showing one page or one flow halfway through can catch mismatches before they multiply. What helps: A mid-point check-in. "First page is ready — want to see if this matches what you had in mind?" Five minutes that can save a round of revisions. 5. Do they have the full picture? You've been living in this

Hardwood, the project Gunnar Morling kick-started handling of Parquet files in Java, reached version 1. Its multi-threaded approach and zero mandatory external dependencies promise a simpler, more efficient alternative to the Apache Parquet Java implementation. For now, the library supports just reading; writing support is expected in the upcoming versions. By Olimpiu Pop

Oracle halved the Always Free Ampere A1 compute allowance from 4 OCPUs and 24 GB RAM to 2 OCPUs and 12 GB RAM with no public announcement. Support agents gave conflicting answers on whether PAYG accounts are affected. Documentation states the new limits apply to "all tenancies" while support emails say only free-tier accounts. By Steef-Jan Wiggers

Google has released A2UI v0.9, a framework-agnostic standard for AI agents to declare user interface intent across multiple platforms without arbitrary code. The update emphasizes alignment with existing design systems. It includes a new SDK for Python, improved error handling, and various transport methods. Migration guidance and evolution specifications are also provided. By Daniel Curtis

You merge the PR. The build passes. The code does what you expected it to do. You move on. That is review for most engineers. A final read. A feeling that things looked right before the branch closed. I did it the same way for years. Three phases had already run before this one. Think had scoped the work, Plan had written the requirements, Build had shipped a diff that matched the plan exactly. I trusted that the chain held. I had never actually checked. Then I ran the Review phase, and checking turned out to mean something specific: not does this work, but does this requirement hold up, and what is my evidence. I went in expecting to approve it or send it back. The phase gave me three answers instead: covered, partial, missing. I found out what they meant one requirement at a time, starting with the one I almost got wrong. I had been giving impressions, not verdicts The notification scheduler used a queue to manage dispatch. Every call to the external provider went through it. The provider was never exposed directly. The requirement said the provider must be notified. It was notified, exactly the way I had pictured it. I almost called it covered and moved to the next line. The Review phase stopped me there. But the requirement said must be notified , not how. The queue had introduced a call order and a timing the requirement never anticipated. Nothing was broken. Something had changed shape, quietly, and nobody had written that shape down. I sat with that for longer than I expected to. Not because the code was wrong. Because I could not immediately tell you whether the change mattered. The same pass gave the shim from Plan a different verdict on the same page: covered. Mapped to the requirement it existed to satisfy, no gap between what was promised and what was in the diff. One requirement held exactly the shape it was given. The other had quietly grown a new one. Same review. Same pass. Two verdicts. Partial is not a softer word for broken. It is the verdict for

HeroUI v3 is a redesigned React component library, previously NextUI, offering over 75 components, including 21 new ones, and a new React Native library with 37 components. Built on React Aria and Tailwind CSS v4, it emphasizes accessibility and customization. The library has experienced many updates since its release, and migration from the previous version is necessary. By Daniel Curtis

Sriram Madapusi Vasudevan discusses industry-converging patterns for securing autonomous AI agents in production. He explains the critical vulnerabilities hidden inside the ReAct loop across context, reasoning, and tool execution. He shares how to mitigate risks like memory poisoning and rogue tool execution using defense-in-depth strategies, LLM-as-a-judge critics, and MAESTRO threat modeling. By Sriram Madapusi Vasudevan

SokoFlow Build Log — Month 1 of 4 Last semester I set out on a new strategic plan to level up my software development skills through deliberate, project-based learning. That work produced one of the most ambitious things I've built so far: Sim-Pesa , a local-first transactional appliance that lets developers working in the M-Pesa ecosystem test and simulate STK Push workflows entirely on their own machines, without depending on the Daraja sandbox. I documented that build in 16 weekly posts, which you can find here . This semester, the focus shifts — from fintech foundations to cloud-native integration and real-world systems. The flagship project is SokoFlow , a conversational ERP for small Kenyan shopkeepers to track inventory and record sales entirely through WhatsApp chat. No app to download, no training session required — just natural language. Where Sim-Pesa lived in a controlled, predictable transactional world, SokoFlow steps into the mess of cloud-native reality: third-party API failures, webhook signature verification, the statelessness of HTTP, and container orchestration. The target audience shifts too — Kenyan SMEs operating on infrastructure that is often unreliable by design, not by exception. It's an ambitious project, but the goal was always to learn as much as possible from it. With the plan in place, I got to work. 1. The Vision of a Headless ERP The first real question I had to answer before writing a line of code: what does "headless" actually mean? Headless architecture decouples the frontend — the "head," or user interface — from the backend, the "body" that holds the data and business logic. A conventional ERP bundles both: backend plus a dashboard or UI on top. A headless ERP, by contrast, is just the engine. The brain. There's no built-in screen. So how do users interact with a system that has no interface of its own? SokoFlow doesn't actually care. It could be: WhatsApp SMS A web app A mobile app A voice assistant In this case, the "frontend

Remember when Artificial Intelligence (AI) felt like something from a science fiction movie? Well, it's not just for movies anymore! AI is here, and it's rapidly changing how businesses of all sizes operate. From making customers happier to solving tricky problems faster, AI is becoming a vital tool for success. But how exactly is AI making such a big difference? Many business owners wonder about the real-world uses of AI. That's why we've put together this comprehensive guide. We're going to explore 50 specific ways AI development is transforming modern businesses, helping them work smarter, grow faster, and serve their customers better. Get ready to see how AI isn't just a buzzword, but a powerful engine driving real change in the business world! Boosting Customer Service & Experience (CX) (1-10) AI is making customer interactions smoother, faster, and more personal. Instant Customer Support (Chatbots): AI-powered chatbots answer common questions 24/7, so customers get help right away. Personalized Recommendations: AI suggests products or services customers might like, based on their past choices, making shopping feel more personal. Faster Problem Solving: AI helps support agents quickly find solutions by sifting through information. Predicting Customer Needs: AI can guess what a customer might want or need before they even ask, allowing businesses to be proactive. Voice Assistants for Support: AI voice assistants can handle basic customer calls, freeing up human agents for more complex issues. Sentiment Analysis: AI understands how customers feel about a product or service by analyzing their feedback (reviews, social media posts). Automated Email Responses: AI can draft quick, helpful replies to common customer email inquiries. Targeted Customer Outreach: AI helps businesses send the right message to the right customer at the right time. Improved Loyalty Programs: AI personalizes rewards and offers, making customers feel more valued and increasing their loyalty.

Erik Steiger discusses the operational pain of legacy PDF generation in regulated banking and manufacturing. He explains how transitioning from resource-heavy engines like Puppeteer and LaTeX to a serverless Rust architecture powered by Typst can drop render latencies below 2ms. He shares how applying Git and Docker concepts to template registries ensures ironclad compliance and rapid debugging. By Erik Steiger

The AI Engineer World's Fair has attracted a lot of adults, and they brought their kids too, for...

GitLab's 2026 AI Accountability Report highlights an AI Paradox: although 78% of developers say they code faster, overall software delivery has not accelerated due to downstream testing and review bottlenecks and new challenges for enterprise governance and traceability. By Sergio De Simone

The AI Implementation Process I Use With Every Client Most AI projects do not fail at the model. They fail in the six weeks before anyone writes a prompt, and in the six weeks after the demo lands in a Slack channel and nobody knows who owns it. I have run enough of these now (from one-off automations to multi-agent content systems running unattended) that the process has converged into something stable. This is the version I actually use. It has five phases: scoping, POC, integration, evaluation, operations. Each phase has an exit criterion. If we cannot meet the exit criterion, we do not move forward. That single rule has saved more projects than any clever architecture choice. Phase 1: Scoping (1 to 2 weeks, fixed price) Scoping ends with a written document that names the workflow being automated, the system of record it touches, the success metric in hours or dollars, the data we have access to, and the smallest possible first slice. No model is chosen yet. No code is written. If we cannot produce that document, the engagement stops here and the client keeps the document. The hardest part of scoping is resisting the urge to solve the interesting problem. Clients almost always describe the AI-shaped fantasy ("an agent that handles all support tickets") when the real opportunity is narrower and uglier ("triage tier-1 tickets that mention billing, route to the right queue, draft a reply for human approval"). The narrower version ships. The fantasy does not. I run scoping as three sessions: Workflow walkthrough. Someone who actually does the work shows me their screen for an hour. I record it. I take timestamps. The point is to find the moments where a human is doing pattern matching that an LLM can do, and to find the moments where they are doing judgment that an LLM should not do. Data audit. Where does the input live? Where does the output need to go? What is the auth story? If the data is locked inside a SaaS product with no API and no export, that is the projec

Like most developers, I have a set of tools I use every day. Most of them work great, and I don't...

Hello Devs, for months, I felt confident about system design interviews. I'd watched endless YouTube videos. I'd studied architecture diagrams. I could explain how Netflix builds recommendation systems. I understood Kafka, Redis, load balancers, and microservices. I'd memorized the designs of Twitter, Uber, YouTube, and TinyURL. Then I sat down for my first real system design interview and froze. The interviewer asked: "How would you design a notification system?" I had memorized notification systems. I knew about push notifications, email queues, delivery workers, and retry logic. I could recite architectural patterns. But suddenly, none of that helped. I didn't know which questions to ask first. I started designing before understanding the actual requirements. I built architecture for problems that didn't exist. I missed obvious bottlenecks. I couldn't articulate why I made specific trade-offs. When the interviewer pushed back, I had no framework to adjust. I failed that interview. But that failure taught me something crucial: System design interviews aren't about knowing technologies. They're about knowing how to think. After that, I went back and systematically practiced 20 system design problems. Not passively watching solutions. Actually designing. Making mistakes and refining my approach. And somewhere around problem 12, a pattern emerged. The best candidates didn't know more technologies than anyone else. They had a framework . They asked the same questions in the same order. They structured their thinking consistently. They could handle curveballs because their framework was flexible. They reasoned through trade-offs explicitly. Here's the framework that finally made it click for me. The Problem with Memorization Before I share the framework, let me explain why memorizing designs fails. When you memorize " How to Design Twitter," you learn: Use relational databases for users and tweets Use NoSQL for timelines Cache with Redis Use message queues for fanout S

There is a popular misconception that if you do TDD, your design also stays correct. That if the tests pass, quality is guaranteed. In AI-assisted development, this misconception is the kind that quietly accumulates — the more tests you have, the more invisible damage builds up underneath. All tests passed. The design was still broken. Here is what happened today. A function called safe_post.py had its signature changed. Two arguments — notify_sh and doctor_sh — were removed. The test suite passed in full. But the callers were still using the old signature. They were silently broken. Why did the tests pass? Because the test code itself was using the old signature. The tests had been written (by AI) at a time when the design intent was already misunderstood. The misunderstanding was baked into the tests from the start. Tests passing and the design being correct are two different things. "All tests pass" tells you only one thing: the implementation matches what the tests expect. Whether the tests express the right design intent is a separate question. TDD verifies "implementation against tests" — nothing more Let me restate the TDD definition. Red → Green → Refactor. Write a test. Write the implementation that passes the test. Refactor. In this loop, what the test verifies is whether the implementation meets the test's expectation. That is one verification — and only one. What TDD does not verify is whether the test itself correctly expresses the design intent. The structure looks like this: Design intent → Tests (← this link is not verified) ↓ Implementation (← this link is verified by tests) If the person writing the tests misunderstands the design intent, the tests will pass and the design will still be wrong. Machine learning engineer Hamel Husain calls this the "Gulf of Specification" — the gap between what you intended to measure and what your metric actually measures. Optimize hard against a flawed metric and you optimize hard in the wrong direction. The same d

Michael Webster discusses the rise of headless AI agents and their impact on software delivery pipelines. He shares how massive, AI-generated pull requests create a severe bottleneck for human reviewers and introduce persistent technical debt. Learn how engineering leaders can leverage test impact analysis and automated validation pipelines to verify agentic output without sacrificing stability. By Michael Webster

Why adaptability beats perfection in startup software development The Startup Trap: Building for a Future That Doesn't Exist Yet Many startup founders make the same mistake. They spend months building the "perfect" product architecture. The code is clean. The design patterns are flawless. The test coverage is near 100%. The infrastructure can scale to millions of users. There's just one problem: They don't have any users. In the startup world, survival depends on learning faster than competitors, not on creating the most elegant codebase. Product-market fit is uncertain. Customer needs change weekly. Business models evolve. Features that seemed critical last month become irrelevant the next. In that environment, the biggest advantage isn't perfect code. It's malleable code . Code that can bend, adapt, and evolve as the business learns. What Is Malleable Code? Malleable code is software that is easy to change. It isn't necessarily perfect. It isn't over-engineered. It isn't designed to solve every future problem. Instead, it's designed to support continuous experimentation. Malleable code allows teams to: Launch MVPs quickly Test assumptions rapidly Respond to customer feedback Pivot when necessary Add new features without major rewrites Remove failed features with minimal effort Think of it this way: Perfect code optimizes for certainty. Malleable code optimizes for uncertainty. And startups operate almost entirely in uncertainty. When you're still searching for product-market fit, the ability to adapt is often more valuable than technical elegance. Why "Perfect" Code Often Hurts Startups Software engineers love solving technical problems. It's natural. Building a scalable architecture feels productive. Refactoring code feels productive. Designing the perfect system feels productive. But startup success isn't measured by code quality. It's measured by business outcomes. Questions such as: Are customers using the product? Are they paying for it? Are they returning? A

Cloudflare has recently shared how they uncovered an issue in their Rust implementation of CUBIC, a congestion controller algorithm, which prevented it from recovering from a scenario of heavy packet loss at the start of a connection. By Gianmarco Nalin