Carlos Henrique

Pessoal, o Google Calendar da Copa 2026 ficou pronto para vocês adicionarem à agenda. No decorrer da Copa, vou atualizando as transmissões da 3ª rodada e os confrontos do mata-mata. calendar.google.com/calendar/u/0?c…

Full write-up: dev.to/carlosperc/an-…

2/ Exported a .memgraph from Xcode. Asked Claude to run leaks on it: leaks ~/Desktop/x.memgraph | grep ROOT CYCLE Output: ROOT CYCLE: SwiftUI._DictionaryStorage<...> → TagIndexProjection<Int> → ForEachState<...> → Closure (.onImageSliderTap) → DetailViewModel

Cost of testing an iOS perf hypothesis the manual way: half a day. Cost with an LLM in the loop driving the simulator and parsing artifacts: 10 minutes. 24x. That's the unlock. Spent an afternoon on this last week, 3 fixes shipped. Quick thread. 🧵

4/ The CLI behind Xcode's Memory Graph(leaks, heap, vmmap, malloc_history on .memgraph files) is fully scriptable. Almost no one uses it that way. When you treat artifacts as programmable inputs and put an LLM in the loop, the math of "let me test this hypothesis" changes.

Fix: 50 lines mirroring an existing pattern. Result: 0 hangs.

3/ After the leak fix, the screen still felt slow. xctrace Time Profiler vs a fluid sibling feature, side-by-side: GraphQLClient.init: 7.7% vs 23.7% CTTelephonyNetworkInfo.init: low vs 18.5% 35 hangs >250ms in 90s. Cache miss in one ViewModel factory.

30 seconds. Without leaks: 4 hours of closure auditing.

We’re updating our ChatGPT Pro and Plus subscriptions to better support the growing use of Codex. We’re introducing a new $100/month Pro tier. This new tier offers 5x more Codex usage than Plus and is best for longer, high-effort Codex sessions. In ChatGPT, this new Pro tier still offers access to all Pro features, including the exclusive Pro model and unlimited access to Instant and Thinking models. To celebrate the launch, we’re increasing Codex usage for a limited time through May 31st so that Pro $100 subscribers get up to 10x usage of ChatGPT Plus on Codex to build your most ambitious ideas.

Introducing Claude Managed Agents: everything you need to build and deploy agents at scale. It pairs an agent harness tuned for performance with production infrastructure, so you can go from prototype to launch in days. Now in public beta on the Claude Platform.

A Anthropic acabou de liberar um CURSO GRATUITO de Prompt Engineering interativo. 9 capitulos + exercicios praticos + playground pra testar na hora: 1. Estrutura basica de prompts 2. Ser claro e direto 3. Atribuir roles 4. Separar dados de instrucoes 5. Formatar outputs 6. Pensamento passo a passo 7. Usar exemplos 8. Evitar alucinacoes 9. Construir prompts complexos (casos reais) Se voce usa IA e ainda nao domina prompt engineering, ta deixando 90% do potencial na mesa. Prompt bom = resultado bom. Prompt ruim = frustracão. GRATUITO no GitHub: github.com/anthropics/pro… Salva esse tweet. Voce vai precisar.

LLMs are much better at writing native code than shared code solutions. The future is Android and iOS native apps.

We just open sourced the code-simplifier agent we use on the Claude Code team. Try it: claude plugin install code-simplifier Or from within a session: /plugin marketplace update claude-plugins-official /plugin install code-simplifier Ask Claude to use the code simplifier agent at the end of a long coding session, or to clean up complex PRs. Let us know what you think!

M5 Pro and M5 Max will have a new chip design with seperated CPU and GPU blocks, allowing customers to mix and match core counts. (ie: Maxed out GPU with base CPU) M5 chip will not be getting any of this new chip tech/design language. This explains M5 Pro/Max delay into 2026.

I'm excited to announce that I'm writing a book for O'Reilly! New example diagrams, new content, and new guidance. Early access to the first two chapters is now live on the O'Reilly platform -> oreilly.com/library/view/t…

Announcing the Swift on Android Working Group, with the goal of establishing and maintaining Android as an officially supported platform for Swift. More here: forums.swift.org/t/announcing-t… #Android

RIP Cursor/Windsurf Claude Code for VSCode 🤯

Are you worried about migration to Swift 6? Here is a list of tips to avoid concurrency issues in Swift 6. 1. Use sendable structs as much as possible. 2. Protect shared state using actors or global actors. 3. Use actors for stateful service types. swiftwithmajid.com/2025/01/07/ado…

@viniciusc70 @interbr hahaha, é o serviço mesmo. Geral reclamando aqui ahhahahaha

@Carloshperc @interbr Hackearam-te!!! Hahahha

não tô conseguindo fazer login no banco @interbr diz diz que meus dados são inválidos. ra-paz, que situação em