CI is one of the MOST important parts every company should be concerned today.
An agent can generate custom workflows easily, and this will create a safe environment for AI-first development.
3
Replying to @leerob
@leerob @elonmusk @cursor_ai Would be nice to make Cursor and Grok Bot communicate each other. Many ideas I have by talking to Grok Bot, but I want to execute in a Cursor session.
5
47
It's just me or nobody can scroll-x the tables on desktop? @bcherny @ClaudeDevs
13
Mais uma aquisição de time de engenharia. Um dos maiores assets que existe hoje na área de tecnologia. Assim como a anthropic fez com o Bun. O artigo diz que eles vão continuar evoluindo o Deno, mas com certeza (aqui é minha opinião) parte da equipe vai trabalhar em produtos internos.
We are excited to announce that Deno is joining Cloudflare!
deno.com/blog/cloudflare
1
54
Essa é uma ideia bem boa na verdade. Não tinha pensado nisso
Here's my technique for adversarial code review if I'm driving from Codex: "Review this with claude".
And when I'm in Claude: "Review this with codex".
Models know how to to kick off a review using the cli. Know how to take turns. Know how to settle an argument.
No magic.
21
Isso me lembra minha mãe falando “se seu amigo se jogar da ponte, vc tb se joga?”
Ai eu falava: “eu não”
As empresas de IA: “SIM”
Replying to @ClaudeDevs
We’re rolling out monthly Claude Platform API credits for Max and Team plans:
Max 5x: $100
Max 20x: $200
Team: up to $500, pooled
Works on any model, including Haiku 5.5, in your code or third-party harnesses.
How it works and full terms: support.claude.com/en/articl…
22
Anthropic faz um Tweet sobre uma novidade: então 99% da minha timeline é sobre outras pessoas falando desse mesmo lançamento. Que cansativo.
15
Eu estou impressionado que existe fanboy de empresas de IA. É só entrar na thread da anthropic e vc vê “oh an open ai está tão morta”, e isso se repete nas threads da open ai e por aí vai.
Surreal
15
Quanto menos texto melhor. Regras claras, simples e diretas, coisa de bullet points de 1 linha, juntamente com contexto.
Mais informação passada ao modelo = mais chance dele ignorar.
I am surprised that people are surprised this is how I prompt Claude.
Talk to Claude the way you would a coworker. There's no secret to prompting. There's no need to be overly scaffolded or prescriptive for most tasks -- give Claude a goal, and it will figure it out.
Back in the Sonnet 3.5 days, your prompt mattered a lot. Nowadays, it's much more important to communicate to the model:
1. What you want it to do
2. How much effort you want it to spend
3. How it should verify that it did the right thing
12
Algo que não gosto no Grok 4.7 e que não acontece no Opus 5.5. Quando um erro 500 ocorre:
Grok: "A página retornou 500, isso não funcionou" [fim]
Opus: "A página retornou 500 porque X está errado e Y não vai receber a interação Z. Corrigindo agora..."
É fato que o Opus é mais proativo.
18
My first Claude Code Mod - Simple and Powerful. It generates a "[Copy]" button after pathnames, and "!" commands. In tables, plain text or any place.
1
1
67
This is interesting... GLM-5.3 (which is Chinese) has hacking capabilities on par with Anthropic's Mythos, but with none of the safeguards.
@DarioAmodei framing is: "see, this is why our approach (lock it down, pace the frontier) is the right one."
Maybe... But it's also a pretty convenient narrative for the company calling for AI regulation while still racing to ship the most capable model first.
A good story is taking shape, just not sure who's writing it.
anthropic.com/research/glm-5…
29
Is Sonnet 5.5 Really Faster and Cheaper? I gave 3 exact tasks for both models for testing with REAL Web Dev Cases
1
31
Just post my first video starting with a bench of sonnets 5 vs 5.5 - interesting the results:
youtu.be/sueyfH6k47A?si=KqAY…
Replying to @claudeai
Sonnet 5.5 improves on Sonnet 5 across benchmarks, in some cases dramatically.
It’s a faster, lower-cost complement to Claude Opus 5.5, strongest at well-scoped everyday tasks, fixing bugs, and creating polished documents, slides, and spreadsheets.
59















