Claude Code’s agent teams feature lets you spin up multiple Claude instances that work in parallel, each in its own visible terminal pane. On Windows, visualizing this properly requires WSL2 and tmux — here’s how to set it up from scratch.

Prerequisites

• Windows 10/11 with WSL2 available
• A Claude Code subscription (Pro or Max) with Opus 4.6 access
• Your repos somewhere accessible (e.g. C:\dev\github\reponame)

Step 1 — Enable WSL2 and Install Ubuntu

Open an elevated PowerShell and run:

AI coding agents are only as good as the information you put in front of them. Every byte of context you load into an agent’s context window is space that could have been used for the code, conversation, or task at hand. Getting this balance right requires the same design instinct that makes good UIs, great games, and well-structured APIs: Progressive Disclosure.

What Is Progressive Disclosure?

Progressive Disclosure is the principle of revealing complexity incrementally rather than all at once. Instead of bombarding an audience with everything they might ever need to know, you give them just enough to act now, with clear pathways to more detail as it becomes relevant.

“Software developers fail in two ways: they build the thing wrong, or they build the wrong thing.”

These two separate failure modes are worth exploring independently, because they have different root causes and very different consequences. And Agentic AI “developers” face the same challenges, but can do both much more rapidly than their human counterparts.

The Two Ways Developers Fail

1. Building the Thing Wrong

Building the thing wrong means the technical execution is flawed. The requirements may have been understood correctly, but the implementation was flawed.

One of the foundational promises of cloud computing is elastic, near-infinite scalability. You pay for what you use, and when demand spikes, the cloud scales with you. It was supposed to be the end of capacity planning, the death of “we need to buy more servers” conversations. So why is Anthropic—one of the best-funded AI companies in the world, running on the infrastructure of Amazon and Google—throttling its users during peak hours?

If you want to stand out, you need to have your own identity. This is true whether you’re talking about you, the person, individually, or your business. You can’t really have your own distinct identity if everything about you is just vanilla, off-the-shelf stuff that anyone else could easily do if they chose to do so. I’m mainly focusing on businesses and their branding and positioning here, but you’ll see parallels that apply to personal brands as well.

DotNetConf (.NET Conf) is a long-running virtual conference hosted each November with the release of a new version of .NET. Its sessions are published to YouTube each year. What have been the top sessions each year? I wrote a tool to pull the stats year by year (since 2021) to see.

Also, why do I keep giving (updated) versions of the same session, Clean Architecture, each year (like last November as covered by Jeremy Sinclair)? There are some who have found it a bit… repetitive (though I do mix it up year to year):

In 2026 we can all agree that AI and agentic development are certainly exciting topics which many see yielding great productivity gains. But as the investor-subsidized pricing of these services gives way to realistic and profitable business models, where will the real costs land?

As many businesses downsize staff or pause hiring to see how these new models and tools actually perform in the real world, the trillion dollar question is:

Introduction

I recently became aware of Asciinema when I saw that the Aspire docs are using it to document installing their CLI. As someone who creates a lot of content (or used to, anyway) and docs, this seems like a really useful tool! But it doesn’t “just work” on Windows, so I figured I’d document how to get it working on Windows for future me.

Asciinema Itself

There are basically two parts to Asciinema:

Introduction

Today’s LLMs have already ingested basically all of the publicly available information they can to build their models. In order to improve further, they’re going to need to seek out additional sources of information. One obvious such source is the countless interactions they have with their users, and while privacy concerns are certainly relevant here, for the purposes of this article I want to focus on another issue: quality signals. How will these LLMs know whether a given exchange led to a solution (defined however the user would define it)? Without this knowledge, there’s no way for LLMs to give more weight to answers that ultimately were fruitful over answers that were useless but the user gave up and ended the session.

Better late than never, as I’m writing this in early February 2026…

First, here are a few links to past years in review:

• My 2017 Year in Review
• My 2018 Year in Review
• My 2019 Year in Review
• My 2020 Year in Review
• My 2021 Year in Review
• My 2022 Year in Review
• My 2023 Year in Review
• My 2024 Year in Review
• My Year in Review Post Checklist

Now on to the various categories…