Adron Hall

Weekend views! Thanks Pacific Northwest!

Shroomies! Rather smurfy like eh!

https://www.elcorazonseattle.com/shows/eluveitie-27-feb Buy tix.

Essential Context Management for AI Agents in Development Explore the importance of context management and precise prompts for effective AI interaction in project workflows using Herdr. https://compositecode.blog/2026/10/09/spaces-tabs-panes-and-precise-words-why-i-break-agent-work-out-the-way-i-do/

Essential Context Management for AI Agents in Development

Explore the importance of context management and precise prompts for effective AI interaction in project workflows using Herdr.

compositecode.blog

I'm not sure about everybody else. But having an "AI Bot" that attempts to sound human - EVEN WHEN SELF-IDENTIFYING AS AN AI - is the absolute worst red flag you can put forward when attempting to recruit me or follow up with me about anything. Just don't have robots call me. Any other method of communication is a better first impression.

Optimize Your Workflow with Herdr: AI Coding Agent Manager Discover Herdr, the terminal workspace manager that transforms how AI coding agents communicate and operate within your terminal. https://compositecode.blog/2026/10/09/herding-agents-how-i-actually-use-herdr-on-a-ridiculously-wide-screen/

Optimize Your Workflow with Herdr: AI Coding Agent Manager

Discover Herdr, the terminal workspace manager that transforms how AI coding agents communicate and operate within your terminal.

compositecode.blog

Gotta admit something. I just kind of got a non-stop low key chuckle at how much teams are integrating Gen AI but then they're just doing the same maddening workflows and pull requests and just trying to do "parts" of it faster. It's so wildly error prone, nobody can really manage what is being done, it's just a wild west clusterfuck of code getting shoved into repos. Some people also chuckle and keep saying eventually that's going to cause a boom for people fixing it. It isn't going to get fixed though, it'll eventually just get thrown away. Nobody is going to comb throw the pixels with a comb and fix that trash. The other wild truth is, it isn't really the Gen AI writing "bad code". It's engineers, and specifically leadership, pushing absurd demands onto Gen AI without any threshold for quality or thinking through how to apply quality to the product under development. This rests on the lack of discipline of leadership - NOT on engineers - they're largely just being forced into absurd situations and thus shoving garbage in and getting garbage out because of this. All of it, on so many levels is maddening but hilarious insane. Anyway, with that said, back to the mines.

https://www.instagram.com/reel/DeDOby4PbNW/?stkn=MXJjd2JlZGs4YzBlMA==

Do you use transit in the Seattle metro area? Tacoma, Everett, Bellevue, Redmond, Kirkland, Bainbridge, Bremerton, Anacortes, etc... Link/Sounder/ST Express, Pierce Transit, Community Transit, King County Metro, Kitsap Transit, washington state ferries, whatever your daily combo looks like. What do you think of this site? What would you want added to it (cuz I'm wanting to add things to it) — real-time arrivals, better route/transfer info, service alerts, something else entirely? Lay it on me. https://www.dashingarrivals.com/

Do you use .NET? Do you have .NET Framework projects of yesteryear that you're still working on? Do you still have VB.NET or VB6 or even COBOL.NET projects? What older, possibly ripe for migration, Microsoft stack developed solutions do you have? Would you mind taking ~2-5 minutes to tell me about it? Bonus points if you're in the Seattle or Portland area, secondary bonus points if you're in San Francisco, San Jose or - ya know - the "Bay" area. Special bonus point if you're in Spokane. Hit me up, let's discuss. Happy to get ya coffee, beer, or whatever beverage of choice you'd like! 🍻

When the laggards and luddites don't realize they're the laggards and luddites, and the minds of those laggards and luddites aren't open enough to even learn that they are, what is the path forward with the likes of that resistance? You see it everywhere: the team that still insists on manual spreadsheets while the rest of the industry automates, the exec who dismisses new tools as "just a fad" for the third cycle running, the colleague who confuses skepticism with wisdom and calls it principle. They're not malicious, just blind to their own blind spot, and that's the harder problem to solve. You can't out-argue someone who thinks the debate is already settled in their favor. So do you wait for the market to force their hand, do you route around them and build without their buy-in, or do you keep trying to open a door that was never unlocked from the inside?

Transport Fever 3 released! It has begun!

Streamline Coding Agent Setup for Maximum Productivity Unlock your coding agent's potential with a repeatable setup process. Learn key strategies for efficient local environments and collaboration. https://compositecode.blog/2026/09/24/agent-ready-local-environments-and-mcp/

Streamline Coding Agent Setup for Maximum Productivity

Unlock your coding agent's potential with a repeatable setup process. Learn key strategies for efficient local environments and collaboration.

compositecode.blog

Fireproof Gloves

Quick proposal: it's time to give up pull requests. PRs were built for a world where a human wrote code slowly, thought hard about every line, and needed a review gate before merge. That world's gone. Now half our diffs are LLM-generated, shipped in minutes, and iterated on faster than any reviewer can meaningfully read them. We're running 2024 process on 2026 output in a world every month is the equivalent of a decade of the 90s or 00s. Why it's broken: - Reviewers are rubber-stamping 400-line (or 1600-line) AI diffs because there's no time to actually understand them - The PR queue has become a bottleneck that punishes speed instead of enabling it - "LGTM" now means "I trust the model" not "I read this code" - We're optimizing for a review ritual, not for actual correctness or safety What to do instead: 1. Replace pre-merge review with continuous, automated verification — tests, type checks, behavior diffs, and LLM-based reviewers that actually read every line (they can, we can't) 2. Merge to main by default, gate releases instead of commits — ship fast, catch issues at the release boundary with real monitoring 3. Reserve human review for architecture decisions and irreversible changes, not routine feature work 4. Treat code as disposable and regenerable — if an AI wrote it in 10 minutes, the fix is often just regenerating it, not painstakingly patching a diff The goal isn't less rigor, it's rigor that matches the actual speed and nature of how code gets made now. PRs made sense when writing code was the bottleneck. It isn't anymore so maybe let's stop pretending our process should look the same? But hey, throw a proposal this way I'm open to alternate solutions to the current mess.

Got on the Link to cross the lake today and a few things came up that just reminded me why I love two things: 1. Pairing transit with bikes. 2. Transit. When you pair your bike, i.e. bring it along with you, it dramatically increases the distance transit can take you, because then you're not dependent on Uber or walking or whatever other such nonsense when you arrive somewhere. But starting out it gives you more options too. But then, today there was a third thing that reminded me how useful transit can be. Beyond the obvious trip one takes. I needed to do two things and instead of standing around working on them at home, I rode down to Westlake to get on Link. I boarded and then pulled out two specific wrenches. One for my bike seat height, and one of the clips on my shoes. Both needed adjusted. So I put my bike on the rack once I boarded the 2 line to Redmond and adjusted the seat right there on the rack. Super easy, wasn't in anybody's way, done! Then I sat and adjust my shoe clips with the other wrench: one shoe finished and then the next, done! Two tasks done without twiddling about having it consume other activities, I used trip time instead. Then I needed to do some banking, also something that is relative good to do with a quick tether connection for internet access. Boom, got that done too! All that and I was enjoying the beautiful Lake Washington Bridge crossing al the while. Now I'm delving into some other work and studies, all while taking care of errands via transit - and bike! It's a great pairing, love it! Thx Seattle for being awesome like this!

Help me out here if you would, I've got a question for you! Prescient Cerebration Mill (PCM) — a factory of foresight, grinding raw signal into forecasts, patterns, and proactive strategy to get your software built with integrity, consistency, speed, and quality. Vision Cognition Plant (VCP) — the sister facility, converting perception into understanding: seeing not just what is, but what it means and where it's headed. A pivotal way to get your software built with integrity, consistency, speed, and quality. Two factory terminologies, one pipeline: sense → interpret → predict → act. Of the two which sounds more enterprisey + businessy + logical to you?

BrickCon 2026! So much going on in this cityscape!

BrickCon GI Joe! 🧐🤓

Also so many folks riding the light rail to BrickCon! 💙 to see that! Good job Seattle!