Sadjow Medeiros Leão

Web dev. Father of 4. Passionate about tech, innovation and making a positive impact. Always learning. Remote worker.

[Boris Cherny @bcherny]

Fable 5 is the biggest step up I’ve felt in our models since Opus 4.5 back in November. After 4.5 came out I uninstalled my IDE when I realized that I’d been doing 100% of my coding in a terminal for a few weeks. With Fable, it’s felt like Claude has stepped up from being a coding agent to a thought and design partner in building the product. Fable has judgement, taste, and dimensionality in a way that previous models didn’t, leading me to trust it more with the most complex work.

I think the first time I had this realization was when I asked Fable to debug something. It is the first model I have used that was so methodical and precise, taking measurements and adding logs then verifying that it truly fixed the issue before declaring victory.

There’s nothing in claude code’s prompting telling the model to do that, it’s just part of its personality. It really has this “big model smell” that I haven’t felt before.

[shadcn @shadcn]

I want the following in Codex, Cursor, and OpenCode...

1. Pinned Messages: Let me pin assistant messages to the sidebar for things I want to keep track of but am not ready to address yet. Render as a checklist & jump navigation.

2. Notes: Give me a scratchpad for thoughts while working.

[Sadjow Medeiros Leão @sadjow]

@cursor_ai you need to support steer messages!

[Sadjow Medeiros Leão @sadjow]

Eu acho que tenho um faro aguçado para reconhecer boa qualidade de código. Faz um tempo que venho notando um desempenho melhor do GPT-5.4 e do 5.5 no @codex, e hoje vejo muitos devs muito reconhecidos falando bem deles também.

[DHH @dhh]

I've had more "I can't believe it's this good" moments with GPT5.5 than any other model since Opus 4.5. It's shockingly, scarily capable. Days and days of amazing progress. All steering, no handwriting. Yet utterly delightful to conduct its coding. So, so good.

[Ilyas Turki @ilyasturki_]

AI CLIs ship a new version basically every day. nixpkgs can't keep that pace, so the package you get is always a week or two behind.

claude-code-nix and codex-cli-nix fix it the boring way: a flake that checks upstream every hour and rebuilds, binaries cached so you compile nothing.

nix run, and you're on today's release. No npm, no pinning. Thanks to @sadjow for these packages.

https://t.co/UF8syDSXMg
https://t.co/UgURkayi5V

[Sadjow Medeiros Leão @sadjow]

The missing feature from these new agentic windows is better git control and UI/visibility. I know we can ask the agent to go to the state, but it would be helpful to know this ahead of time and to be able to change it manually as well before starting to prompt.

[José Valim @josevalim]

Alright, after a full week with Codex, I understand why folks like @chris_mccord and @antirez have been praising it.

It’s far more thorough than Claude Code. When you ask for a change, it does a better job of understanding the system and the different areas that will be affected.

The only downside is that it can sometimes overreach on smaller changes, but that’s partly because I’ve been feeding it piecemeal requests. I’m starting to trust it with larger tasks now, something I didn't feel much comfortable before.

[Sadjow Medeiros Leão @sadjow]

@evanyou I noticed this as well and I think the issue is not at the model level because the gpt-5.4 at @cursor_ai is also weak than what we have in the Codex. I think the system prompts we have in Codex are helping a lot with the outcomes. Using both xhigh.

[Domen Kožar @domenkozar]

I used to merge Claude Code updates into nixpkgs within an hour since it was released, now it sits there for quite a bit :)

Looks like someone automated it: https://t.co/dZbGytc9pQ

[Sadjow Medeiros Leão @sadjow]

@vussyviz I have an unique name and the magic was that I have a google alert for new indexed pages. :) glad the project was the solution! Thanks for sharing :)

[Vis (SF Jun 9-17) @vussyviz]

ok i got it working
.
.
with sadjow/claude-code-nix
https://t.co/eqMVpGfABn

[Google Research @GoogleResearch]

Introducing Vibe Coding XR, a new rapid prototyping workflow that empowers Gemini Canvas w/ the XR Blocks framework to turn user prompts into interactive, physics-aware WebXR applications, allowing creators to quickly test intelligent spatial experiences →https://t.co/suwxBMoMvD

[Anthropic @AnthropicAI]

We partnered with Mozilla to test Claude's ability to find security vulnerabilities in Firefox.

Opus 4.6 found 22 vulnerabilities in just two weeks. Of these, 14 were high-severity, representing a fifth of all high-severity bugs Mozilla remediated in 2025.

[Hugo Baraúna @hugobarauna]

OpenAI built a cool agentic project using Elixir! 🔥

#ElixirLang

[Theo - t3.gg @theo]

Elixir?? You have my attention OpenAI

[Google @Google]

Gemini 3.1 Pro be like

[Alvaro Cintas @dr_cintas]

Figma meets Claude Code!

Pencil is an infinite design canvas that runs parallel design agents locally.

→ Turn designs into code
→ Runs locally with Claude Code
→ Works with VSCode, Cursor

100% free, no subscription

[Boris Cherny @bcherny]

Love seeing how Spotify is shipping with Claude Code.

Their best developers haven't written a single line of code since December, they fix bugs from their phones, and they shipped 50+ features from Slack during morning commutes

https://t.co/rYTVJBHE0s

[José Valim @josevalim]

Here is my take on why Elixir is the best language for AI: immutability, documentation, stability, and tooling for coding agents.

It builds on the recent study in which Elixir had the highest completion rate across models among 20 different languages.

Link in the thread below.