sh03 @sh03.dev · 10/06/2026just updated macos and now i have to live with these blue slices in every angle of the screen 100
sh03 @sh03.dev · 27/05/2026People complaining about the new Google icon set are delirious. Should I remind you of the last icon set? 000
sh03 @sh03.dev · 17/03/2026Another example: told it to fix a conflicting warning in VSCode (medium reasoning) it simply replaced the entire VSCode configuration file (tbf while also fixing the issue) 100
sh03 @sh03.dev · 17/03/2026This is codex with gpt 5.4 (paid plan) on high reasoning: it just needed to remove the useViewer reference (it's just a react query) instead it left a useless useViewer call and removed unrelated code. You guys are joking right? This is the best LLMs have to offer? 120
sh03 @sh03.dev · 02/12/2025Still can't believe the UI/UX community allowed floating action buttons to be popular. There should be Nuremberg trials when the reign of terror, this anti pattern has brought upon us, ends. I bet they are the same people that are now trolling us with the liquid glass design. 110
sh03 @sh03.dev · 30/11/2025Gemini recommending "module": "commonjs" by default for tsconfig.json is wiiiiild. 010
sh03 @sh03.dev · 01/11/2025We have officially gone too far. Yet another directive. Nobody writes code like this, come on. This is a solved problem: you create a task queue and then have a scheduler execute these tasks. Why is the ecosystem moving more and more towards pure magic? 020
sh03 @sh03.dev · 08/10/2025@infrequently.org Corporate needs you to find the difference between these two pictures 100
sh03 @sh03.dev · 17/09/2025😅 My OCD is triggering me so much with this (profile icon on Bluesky) but now I have much more empathy for the dev team. 010
sh03 @sh03.dev · 17/09/2025I absolutely despise this way of presenting filters. It's so chaotic and difficult to implement correctly. It's a prime example of a lose-lose UI component. You, the company, the users... Everybody loses. 010
sh03 @sh03.dev · 12/08/2025This is prototype-level code. Less than 330 lines of code. 1 single objective. Using a library that is battle tested to solve that objective. With extensive documentation and examples. Somehow it managed to create a non trivial bug. This is the latest (released 56 days ago) "Pro" model. 100
sh03 @sh03.dev · 12/08/2025... surprise, surprise! The bug was not fixed. At least it fixed the height issue while dragging items. Now it's the width that shrinks instead. 100
sh03 @sh03.dev · 12/08/2025I asked it to fix it. He confidently told me he knew how to fix it. So I followed the instructions and... 100
sh03 @sh03.dev · 12/08/2025Not only that but it seems like the "active" tiers are chosen based on the first item that you drop. 100
sh03 @sh03.dev · 12/08/2025I kept prompting it the errors to try and fix them but there was just no way and I was getting irritated so I slapped some `!` to make it all compile (not modifying the resulting JS). I got this: 100
sh03 @sh03.dev · 12/08/2025Tbf this is the Flash version, so let's try 2.5 "Pro". It gave me code that didn't compile. 110
sh03 @sh03.dev · 12/08/2025I started off using 2.5 flash. It gave me this. Nothing worked. The tiers were not colored, nor were they ordered properly. 100
sh03 @sh03.dev · 12/08/2025The project? Creating one of these tier lists that everyone loves. I told it to use dnd-kit, react, typescript. 120
sh03 @sh03.dev · 10/08/2025TIL one of my favorite Italian snacks growing up actually contains 8% liquor in them. It explains so much. 🤌 110
sh03 @sh03.dev · 10/06/2025Now that React is moving towards having a compiler I wonder if conditional hooks are possible. IIRC the reason the Rules of Hooks™ exist is that conditionally calling hooks break because of different ordering between renders. 🤞 300
sh03 @sh03.dev · 11/04/2025Any designer that wants to comment on this amazing color contrast on Youtube? 😅 000
sh03 @sh03.dev · 19/02/2025Newer and more modern AI/LLMs instead just add a new step to explicitly remove the parenthesis as they don't understand that the code is correct and/or ask questions to better understand the problem. 100
sh03 @sh03.dev · 19/02/2025AI/LLMs cannot reason, part 475: When ChatGPT is presented with a non working Postgres function (which actually works correctly and the issue is elsewhere) it incorrectly spouts misinformation, lies and refactorings masked as solutions. 110
![Futuristic image used as a meme in "The world if we could [good thing]"](https://cdn.bsky.app/img/feed_thumbnail/plain/did:plc:dkqcftotcpzy3dcpe7a7yz34/bafkreigmgfsnp3mpvwf3ttmi6mqddw4q2cnv24re25pigphj5dvhlcex4y)
