For the past year, I've been building visual prototypes as a product manager without a design background. What I've learned is that the tools have recently shifted in a way that fundamentally changes how we can approach this work—and it's a complete game
If you’ve ever built a chat interface, activity feed, or onboarding flow where content keeps growing over time, you’ve probably run into animation issues. Things start out fine with a handful of components, but as the UI grows, animations get noisy, scroll behavior becomes unpredictable, and performance starts
What is the router ? Every message into Velo™ goes through one agent: the router. The router's job is to examine the conversation and select exactly which downstream agent should respond. It could be checking order status, retrieving context from the knowledge base, concluding the conversation, or any other
Context Google's recently open-sourced A2UI, a protocol that lets AI agents generate declarative UI (as JSON) rendered by host applications using native components. The goal is to let agents “speak UI” safely across platforms and trust boundaries. Teams building product-owned agent experiences are starting to ask how A2UI
I built a custom text-to-sql agent.. That’s a lie. I just told Cursor how to query tables and that’s good enough. If you’ve used an agentic coding tool, you’ve seen how well agentic tools work with command line tools. With the right rules (or skills for
There’s a persistent myth that building AI products requires throwing out everything we know about Agile. That AI is fundamentally different. That "AI Success" demands large teams, big budgets, longer runways, bigger bets, or fully formed platforms before users ever see value. That wasn’t our experience
Anthropic’s Claude helps me with a lot of tasks. I write techblogs about the things I do so people can see how cool, smart, and busy I actually am, so why shouldn’t Claude get to do the same as it does more and more of the work for
The "Side-Scanning" Revolution (Agentless Security) The core problem with traditional security is that it tries to run "inside the house." It sits on the server, fighting for resources. Our approach separates the duties: * For Scanning (Agentless Side-Scanning): Instead of installing software on the live servers processing
Every AI product team eventually hits the same crossroads: just because you can put something in a chat interface does not mean you should. Chat is fast, flexible, and incredibly capable, but it is not a universal solution. Sometimes chat accelerates the user. Other times it slows them down, disrupts
The Theory of Constraints (TOC) teaches a simple truth: every system has a limiting factor that determines its overall performance. Improving anything other than that constraint is wasted effort. In manufacturing, that constraint might be a slow machine. In software, it is often the people who carry the highest cognitive
Almost every complaint about AI agent tools going off the rails reduces to one problem: if AI is misbehaving, you haven’t built your project with enough rigor and structure to enable autonomous development. The irony is that building effectively with agents forces you to face fundamentals many engineers avoid.
The Rise of Malicious Bot Traffic The internet is now dominated by automated activity. According to the 2025 Imperva Bad Bot Report, bots generate over 51% of all web traffic, and nearly 37% of that is malicious. In other words, more than two-thirds of automated requests online are from “bad