Watch, Read, Listen
-
There’s a lot of excitement around AI writing code. And that excitement is justified. AI can produce code faster than any individual developer ever could. It can scaffold services, write tests, generate APIs, and even refactor existing systems. But the best way to understand AI isn’t as a replacement for engineers. It’s as something else
-
When I was working for Palm, I came to them because they bought out startup AnyDay.com. If you believe in patents, which I don’t, we invented the online calendar, which is why they bought us. After the dotcom crash at the end of 2001, Palm decided in their shortsighted way that their future was as
-
—
by
I made a open source project on GitHub called “Gambler” that uses SpecKit plus the AI agent of your choice to assist you in making gambling decisions. It prompts your for all the relevant information you gave give it — either URLS or you can cut and paste stuff. Tell it all the factors you
-
AI-assisted programming has quietly created a new development style: vibe coding. You describe what you want.The model produces code.You skim it, maybe run it, tweak a few things, and keep moving. Sometimes it’s astonishingly productive. A feature that might have taken an hour appears in seconds. But vibe coding changes something fundamental about software development:
-
I got a SMS message from a friend I loosely know. We’re FB friends and we’ve met IRL twice. They are also a *serious* player in the startup space — and I have a startup looking for financing.Here’s how the conversation went. This is where I think they are probing me to find out if
-
There’s a narrative forming in tech that AI is a young engineer’s tool. The assumption is that younger developers will adopt it faster, integrate it more naturally, and ultimately dominate in an AI-assisted development world. I’m not convinced. Older engineers may be uniquely equipped to use AI effectively. AI Makes the Same Mistakes as Junior
-
There’s a belief in our industry that building software is incredibly hard. It isn’t. Writing code that works? That’s increasingly easy. Even “simple” AI systems can generate functional microservices with decent structure and solid test coverage. Given a clear task, modern AI can scaffold APIs, wire up a database, add authentication, and produce a respectable
-
If you only open some developer tools occasionally, you’ve seen this: Open the tool → giant update → wait before you can work. This happens because Windows does not centrally update developer software. Many tools ship their own updaters, and they only run when you launch the application. Windows already has the solution: winget, the
-
If you build on AWS long enough, you’ll eventually face the same question: Where should we store secrets and configuration? Keeping stuff in a local env file or applications.properties means either those files are going to get checked in, or even worse, casually shared between engineers. And what if your builds are made in a
-
Executive Summary On-demand cloud workstations unlock powerful, cost-efficient computing for engineering teams. But as organizations scale beyond a handful of machines, a new challenge emerges: consistency. Manually configuring Windows systems through portals, scripts, and remote desktop sessions introduces drift, delays, and operational risk. This paper explores how Terraform and Infrastructure as Code (IaC) transform Azure-based