FLUXparticle

Visualizing XML Schema: A Picture is Worth a Thousand Words xsdvisualizer.fluxparticle.com/permalink/ABFl…

Because reviewing AI-generated code becomes more and more important, I wanted to create a tool to help... This video shows my prototype. What do you think? wearedevelopers.com/en/videos/2102…

NEW: Apple’s AirPods with cameras for AI are launching next year alongside the 20th anniversary pro iPhones and a second-generation foldable iPhone. It’s all part of Apple’s biggest new product period ever. bloomberg.com/news/articles/…

@unclebobmartin Or maybe... Let the agent write the script 😉

if anything you want your agents to do is deterministic, such as managing a queue, or sorting a list, create a script for the agent to use. Agent's don't do deterministic things well. It's our job to force determinism upon them.

#SiriAI does perfectly read and write German already.

Writing code gets cheaper with #AI. Understanding architecture does not. That’s why teams drown in systems nobody fully understands. Sven Reinck shows how dependency visualization can make large codebases navigable again: javapro.io/2026/06/10/bey… #CleanCode #Java @FLUXparticleCOM

@AKBrews Meanwhile on a Mac: caffeinate -d

Human ingenuity never ceases to amaze me.

@MatthewBerman I am working on something to tackle that 😁

there's no way to review all code produced by ai and there's no way you're actually reviewing all the code explanations

@_toolone_ @garrytan I use MCP to teach Codex about my CLI tool with help parameter. Works like a charm and the model only needs to look at commands it needs.

@FLUXparticleCOM @garrytan Just skills, take a look at the linked GitHub repo. Skills are powerful enough to instruct Claude to run scripts, CLI commands… whatever you want. Also important in gstack, the additional instruction in CLAUDE.md to not use the Chrome MCP.

MCP sucks honestly It eats too much context window and you have to toggle it on and off and the auth sucks I got sick of Claude in Chrome via MCP and vibe coded a CLI wrapper for Playwright tonight in 30 minutes only for my team to tell me Vercel already did it lmao But it worked 100x better and was like 100LOC as a CLI

@garrytan How do you tell Claude about your CLI tool?

I just open sourced my whole Claude Code setup here

AI shouldn’t replace buttons. It should replace complex exploration. I asked my AI to search years of archived tasks, use binary search to locate events, and detect patterns in my workflow - all through a simple CLI toolset.

And when I add a new command, I don’t need to restart the context with a new tool list. I just tell the AI to execute the help command again.

They work even better combined. I've built a MCP with just ONE command. THIS is the description: Runs todo-api commands with the provided CLI options array. Use ["--help"] for global help or ["<command>", "--help"] for command help. So AI can figure out the commands on-demand.

I asked codex to apply a clean architecture solution (components and dependency inverted boundaries). It wrote up a nice plan describing the components and the boundaries. Then I asked it this: › would the result of executing this plan make it easier for you to work with this system? Why? • Yes. Because it would reduce ambiguity and hidden side effects, which are the main things that slow safe changes: 1. Single mutation boundary With apply-command!, I can trace state changes in one place instead of chasing swap! across many files. 2. Clear dependency direction If domain/app no longer depend on atoms/UI/filesystem, changes are localized and lower-risk. 3. More pure functions Pure transitions are easier to reason about, test, and refactor without breaking behavior. 4. Better failure isolation Ports/adapters make bugs easier to classify: domain rule bug vs UI adapter bug vs persistence bug. 5. Faster, safer edits When invariants run after commands, regressions are caught immediately, so iteration speed improves. So it helps me work faster, but more importantly it makes changes more predictable for everyone on the codebase.

AI doesn't feel the pain of bad code

@LeakerApple Maybe they offload some of the rendering to achieve 120 Hz as a suggested for years.

A19 chip is overkill for a monitor. Why does Apple do this?

@rqobela @Timofey_Khmelev It's the basis of Java 8 streams and everything reactive.

@Timofey_Khmelev I have never heard about it

Programming language you learnd but never used again is...?

@Bill_Prin @MatthewBerman Or like Bart Simpson 😂

@MatthewBerman When I tried to get it to stop, it kept saying things like “Got it — I’ll stop using the em-dash in my responses.” Felt like an elementary school bully or some Austin Powers bit.

AI's desire to use em dashes is too strong for system messages, memory, and even direct prompt instruction.

@LeakerApple A good improvement would be the color

It has been a month since the launch of the iPhone Air and I can confidently say that Apple absolutely nailed this product. They understood that people didn't want a thin phone just for the sake of it being thin. They wanted a thin phone because it would be a new and fresh way to interact with a smartphone. The smartphone market is all the same now. Flat screens with rounded corners and flat sides; big cameras and big batteries to suit. But it was getting boring, and people just wanted a phone they could have fun with again. The iPhone X in 2017 brought an end to the decade of monotony after the original iPhone, and I feel the same way about the iPhone Air. Yes, it made sacrifices. A single camera and single speaker in 2025 are huge compromises for a $999 device. But, when you actually pick up and use the device, you almost forget about its shortfalls. This is wow factor is something that was desperately missing from Samsung's S25 Edge (which by the way was rushed into existence due to rumours of Apple developing an ultra-thin iPhone). The Edge feels thin for sure, but it also looks and behaves exactly like a "normal" phone. There weren't any engineering miracles like Apple's camera plateau, or magic hardware optimizations like Apple's C1X modem. In the end, the S25 Edge simply felt stale and uninspiring, and it is reflected in its poor sales. In fact, Samsung is now discontinuing the Edge lineup... Apple has found success with the iPhone Air due to nothing other than it finally feeling like an iPhone again. It truly is something Steve would have strived to will into existence if he had modern-day technology, and I don't say that lightly. An iPhone Air 2 is slated for September 2026. We don't know what the improvements will be, but I do think that the Air will eventually overcome its downsides and continue to be loved.

@MatthewBerman Obviously!

Typing is slow. Speech is the closest to how quickly humans think. That's why I am increasingly bullish on Voice as the ultimate UI for humans x AI. Plus, you can speak even when on the go, driving, etc.