in.orbit

Claude Code @anthropicai
Software AI Enginner
Building systems you can actually trust

[in.orbit @inorbit01]

Pode parecer muito bom mas não podemos perder autonomia nesse ciclo

[Daniel San @dani_avila7]

Claude Code 2.1.139 added /goal

You set a completion condition and Claude keeps working across turns until it's met

Works in interactive, -p, and Remote Control 👏

[in.orbit @inorbit01]

A IA não precisa parecer humana.

Precisa ser auditável.

[in.orbit @inorbit01]

@DilumSanjaya

[in.orbit @inorbit01]

@noahzweben Congrats!

[ClaudeDevs @ClaudeDevs]

Up next:

▪︎ 1:00PM - A conversation with our co-founders Dario Amodei and Daniela Amodei, moderated by Chief Product Officer Ami Vora
▪︎ 1:50PM - Watch how @bcherny and @jarredsumner build with Claude Code

[in.orbit @inorbit01]

Anthropic just made a massive leap in Agents.

The biggest update of the year for Claude Managed Agents just dropped live on “Code with Claude”:

Dreaming (research preview) → agents can now “sleep,” review past sessions, identify patterns, correct recurring mistakes, improve their own memory, and even create new skills independently. Basically REM sleep for AI.

Outcomes (public beta) → you define a quality rubric, and the agent keeps iterating on its own until it meets the required standard.

Multiagent Orchestration (beta) → a lead agent delegates tasks to specialist agents running in parallel, sharing the same filesystem, and operating like a real team.

Webhooks (beta) → true asynchronous integrations without constant polling or keeping streams open.

Bottom line: Claude just evolved from being an intelligent chatbot into a production-ready agent platform.

This is no longer just prompt engineering. This is autonomous workflow infrastructure that improves through use, learns from history, and actually delivers outcomes without requiring constant supervision.

For anyone building products, automation, or using AI daily: this changes the game. Agents are moving beyond execution… they’re starting to evolve.

Anthropic isn’t playing around. They’re building the operating system for agents.

What do you think, does this accelerate the path toward agents that can genuinely run parts of a business while we sleep?

#Claude #AI #Agents #Anthropic #CodeWithClaude

[ClaudeDevs @ClaudeDevs]

In Claude Managed Agents, we’ve added multiagent orchestration, an outcomes loop for rubric-driven self-improvement, dreaming for self-learning, & webhooks.

[ClaudeDevs @ClaudeDevs]

Usage limits are up, effective today we're:

1) Doubling Claude Code's 5-hour limits for Pro, Max, Team and seat-based Enterprise plans

2) Removing peak hours limit reduction on Claude Code for Pro and Max plans

3) Substantially raising our API rate limits for Opus models

[Claude @claudeai]

Effective today, we are:

1) Doubling Claude Code’s 5-hour rate limits for Pro, Max, and Team plans;

2) Removing the peak hours limit reduction on Claude Code for Pro and Max plans; and

3) Substantially raising our API rate limits for Opus models.

[in.orbit @inorbit01]

Reset tokens @ClaudeDevs ? :D

[in.orbit @inorbit01]

Live now

#claude #ai

[ClaudeDevs @ClaudeDevs]

Code with Claude is happening now!

▪︎ 9:00AM - Keynote
▪︎ 10:30AM - What's new in Claude Code
▪︎ 11:15AM - Building on Claude at GitHub scale
▪︎ 12:00PM - Get to production faster with Managed Agents

All times PT. https://t.co/O4oygkSr2G

[Ethereum @ethereum]

Last week, Ethereum core contributors gathered in Svalbard for the Soldøgn interop: a week long event focused on hardening Glamsterdam implementations to scale Ethereum securely ☀️

Read the full recap, including their candidate post-fork gas limit, below:

[in.orbit @inorbit01]

What is Orbit?

Orbit is an AI workflow ecosystem built to make work with AI clearer, safer, and more execution-ready.

It starts before the AI answer.

Because most bad AI outputs do not start with the model.

They start with unclear requests.

Orbit Prompt helps you turn rough intent into structured prompts.

Orbit AI helps you reason over projects and decisions.

orbit-engine adds execution awareness, safety signals, and workflow observability.

Together, they create a simple path:

messy idea
→ clearer request
→ better reasoning
→ safer execution
→ traceable progress

The first step is simple:

Install the Orbit Prompt plugin.

Then, before your next AI request, write:

“use orbit-prompt to help with this implementation”

Link in bio

#Orbit #OrbitPrompt #AIWorkflow #PromptEngineering #DeveloperTools #AIProductivity #BuildInPublic #SoftwareEngineering #DevTools #ArtificialIntelligence

[in.orbit @inorbit01]

Orbit AI v0.4.0 is out.

This release closes the first Local Evidence Lifecycle.

The focus was simple:
make reason runs more traceable without storing sensitive content by default.

Evidence is local, optional and metadata-only.

No prompt.
No answer.
No file content.
No hidden persistence.

Small step, but a real one.

[Stephen Dixon @stphndxn]

Today is a beautiful day, so I’m launching @naturalis_app a day early.

If you or someone you know wants to get outside more, notice what’s around them, and build a calm, beautiful collection of what they find, Naturalis is for you.

Available now:
https://t.co/CWn8lldZQa

[in.orbit @inorbit01]

Useful signal.

The real frontier is not just better answers.
It is reducing false confidence, emotional mirroring, and decision distortion at scale.

People are already using AI for life choices.
That raises the bar from “helpful” to “responsible.”

Building tools that preserve clarity instead of dependency feels like the deeper challenge.

[in.orbit @inorbit01]

I completed the Claude with the Anthropic API course.

The most important point for me was not just learning how to use an AI API better, but understanding more deeply how agent based systems need to be designed with engineering discipline.

Tool use, MCP, agent architecture, evaluation, prompt caching and optimization make one thing clear: working with generative AI in production is not just about testing prompts.

It is about building context, measuring behavior, reducing noise and creating more reliable systems.

This learning connects directly with what I have been studying and building around AI assisted engineering, real productivity for developers and tools like Orbit Prompt.

#AI #ClaudeAI #Anthropic #AIAgents #MCP #SoftwareEngineering #OrbitPrompt

https://t.co/UWAxl4QaLh

[in.orbit @inorbit01]

Tim Cook told 8B ppl what the next decade will cost: be a builder, not a watcher.

No mention of chips/sw/revenue—he diagnosed human nature. Internet blurred the line; AI closed it permanently.

Consume? AI feeds u into irrelevance. Build? U get 1K engineers' output on 1 screen. Same tech, opposite paths.

1 person w/ right arch now ships what depts couldn't 3yrs ago. Leverage is generational & compounds for creators.

Gap btwn builders & commenters = century's defining fracture. Orientation > wealth/credentials.

Pick up tools or become someone else's foundation.

#AI #ArtificialIntelligence #BuildInPublic #BuildersMindset #FutureOfWork #CreatorEconomy #TechLeadership #Innovation #Productivity #SoftwareEngineering #StartupMindset #DigitalTransformation #NoCode #Automation #Entrepreneurship #TechTwitter #AItools #GenerativeAI #Leverage #CreateMoreConsumeLess

[in.orbit @inorbit01]

orbit-prompt in action.

A simple request becomes a clearer prompt with scope, risks, criteria, and next steps.

Built to reduce noise before the agent executes.

GitHub: https://t.co/13SmZOigl7

#AI #ClaudeCode #DevTools #Skill #Claude

[in.orbit @inorbit01]

Most AI coding issues start before the model answers.

The prompt is vague.
The scope is loose.
The risk is hidden.
No test proves the output.

I built orbit-prompt to fix that.

A public skill for developers who want clearer, safer, testable AI assisted work.

Try it:
https://t.co/muQw9Thzp0

#AIforDevelopers #PromptEngineering #SoftwareEngineering #DevTools #ClaudeAI #OpenSource

[in.orbit @inorbit01]

GPT-5.5 is not just about smarter answers.

It points to something bigger: AI that can understand goals, use tools, verify its work and move complex tasks toward completion.

The next phase is not prompt assistance.

It is execution based intelligence.

[Greg Brockman @gdb]

GPT-5.5 is a new class of intelligence.

This intelligence makes it intuitive to use; it completes challenging tasks with little micromanagement. Also very token efficient, and runs with low latency and at scale.

A real step toward a new way of getting computer work done.