Drafts

I’m tracking results of 2026 Predictions

🌱 This post is still growing

a scroll through of https://github.com/WaylonWalker/markata-go/pull/1021

🌱 This post is still growing

Late last year I started writing I'm Out On Agents. Agents sucked, the models were good, but there was still something missing between the harnesses and the models. They could write good code, they could do some debugging and exploring, but they were too good at fucking up the whole project to be useful. They could crank out Green Field POC’s like nobody’s business, but they created so much mess in brown field projects that it was easier to chat and edit yourself.

The Beautiful Glitch - Gemini

...

I’ve been thinking about this for awhile, the AI wars are so much worse and burnout prone than the framework wars of the 2010’s.

I remember really starting my professional programming journey during the framework wars. It was a time when there were new and exciting js things every single month. Frameworks and meta frameworks came and went, the ones that lasted changed best practices yearly or so, often flip flopping on technique.

I was deep in python and data engineering at the time and only experienced it adjacently. I was into webdev. I did a bit of react, gastby, vue, gave all the big ones a try in a demo level.

Session ID: ses_3444b09b1ffeXTEr3ELFg5qqtg Created: 3/4/2026, 7:55:24 PM Updated: 3/4/2026, 8:52:32 PM

I want to spec out a cli for tmux session switching in go.

keymap

...

Ansible Vault keeps secrets out of sight, but the ciphertext changes on every encrypt. That turns Git diffs into noise and makes it hard to tell if anything actually changed. Decrypting, editing, and re-encrypting often leaves uncertainty about whether any plaintext changed. This is amplified when secret repos are tightly coupled to dependent repositories. A typical cycle includes decrypting, adding a key, updating a value, applying changes, and returning later with little clarity about what changed while secrets were in plaintext.

Today a new workflow was created with @gpt-5.2-codex to keep diffs clean and avoid re-encrypting when the plaintext is identical.

Re-encrypting a file with the same content produces different ciphertext. A simple decrypt/encrypt cycle can look...

...

I don’t know if you know this, but the web is a beautiful platform that allows you the freedom to create things and put them out there. Its not tied to four major platforms. You don’t have to post your thoughts, ideas, and apps to a platform, you can just make it. This is a beautiful thing that seems to have been forgotten. I was inspired this morning from @scotthanselman’s tinytooltown. Looking through all of the tiny tools that people have built for themself, as personal software, not answering to anyone but themself, it was inspiring.

Agents have gotten a lot better, like seriously better. The ai bros that were ai pilled too early that said SWE is over in six months called it too early. It wasn’t time. Now since Nov 2025 we have had agents that can do some damn work. Proving the point some of the greatest devs I’ve ever looked up to have not written a line of code since. Not hype bros or someone not good at the craft, but...

...

Is AI making us more productive yet, more faster yet?

probably not

I’ve seen this question hitting all over the Internet lately, and often points to people not writing code. Copilot turns prompts into emails, emails back into summaries that look a lot like prompts. I think there’s a place for this, making rambled thoughts sound more coherent, summarizing notes and meeting minutes. All good stuff but does it make us more productive, probably not by an amount that you can put $ $ behind, unless you are reducing headcount. thats not what we are doing right???

...

drafted by kimi

How I write and build this site. A personal style guide.

Casually self-deprecating with technical credibility.

...

drafted by kimi

Things I’m passionate about and why they excite me.

I’ve fallen deep into the custom keyboard rabbit hole. Building keyboards combines electronics, ergonomics, and 3D printing into one satisfying hobby.

...

Are developers really running 5 agents in parallel? How the Heck do they keep up with the changes? This seems Impossible.

I was listening to Syntax.fm this morning and heard this question, and thought I’d throw in my take, which is probably pretty similar to Wes and Scott’s.

https://www.youtube.com/watch?v=NrBQI9So5lM&list=PLLnpHn493BHHNUfHN5lDf11UD8jQ5Bpzl&index=1&t=99s

...

A running list of blog post ideas to strengthen underrepresented topics on the site.

Building out Docker/containerization content to match the llms.txt claims.

Lightest coverage area. Need to expand beyond the few short posts.

I’ve been dabbling with go this week using

In early 2026 I’m trying to lean more in to agentic workflows. The tools are not only better than they were a year ago, but available to me now unlike they were 6 months ago.

At home I’m using opencode, the tooling here is fantastic. LSP incide of these things is incredible, mcp is fine when its needed. The free models it gives you are impressively good for free models, but they are still not the big models from the big providers. I’m using what work gives me, they give me a tool with access to good models, the models are great, the tools kinda suck. I’m being vaugue here because I dont share real work details.

I’ve been skeptical the whole way, I see ai being a very useful tool. I remain January 19, 2026 2 min read

I’m late on this one so I’ve seen everyone else’s. I’ll try my best to make some bold predictions I’ve not seen elsewhere

I predict that this is the point in time that we see an explosion of open source, but more projects than ever going the way of sqlite, Livestream, android, unreal engine, aesprite. It is getting so easy for agents to generate 10k lines of code in a few hours. This makes it easy for folks who have ideas they want to create to create them. It also makes it much harder to accept outside contributions when they make such massive changes. Thirdly it will be easier for folks to just make their own fork to do what they want.

Right now it’s hard to name 5 big open source, closed contribution projects by the end of the year there will be 5 more that are recognizable.

...

A new approach to agentic workflows.

This is probably news to no one else, I’m sure I’m behind on this one. You can’t one sentence prompt and expect to get what you want.

It’s the start of 2026 and agents are getting a lot better than they were. I’m using opencode at home, free mode with Zen and big pickle. At work I have access to a wider variety of models including what seems to be the gold standard 3 from anthropic opus, sonnet, haiku.

Around Aug 2025 I wrote I'm Out On Agents. I saw others in the space having such great success I gave it a solid shot, but found it to egregious edit more than I asked, make massive unneeded changes, and make more small bugs hidden in the details than was...

...