Rodrigo ☕︎ 🤷🏻‍♂️ 👇🏻

@izzynobre Anti IA? Meu irmão, por mim usava IA até para me dar banho. Ser averso a IA na altura do campeonato me remete aos datilógrafos que eram contra os PCs assim quando chegaram. Qual firmware é este? Se puder me passar qual guia você usou, eu agradeço.

É só pedir pro chatGPT fazer crachás de instituições fictícias NASA, SHIELD, MIB, etc (A galera anti IA ficou putaça. É interessante essa dicotomia. Uns esculachando, outros perguntando como fazer gahahaha) Depois é só botar numa pasta chamada SLEEP na raiz do microSD. Mas configurações, é só trocar o que aparece na tela durante o sleep. A opção padrão é a capa do livro

Usar Kindle em 2026 = mico O futuro é isso aqui. Open Source, mil firmwares pra escolher, ler livro sem DRM, puxar do bolso a todo momento em vez de doomscrolling no celular Minhas impressões do XTeink X4: youtu.be/aOkZtBtGcNk?si…

A Microsoft tinha Windows, Office, Azure e OpenAI. Mesmo assim, o Copilot virou uma dor de cabeça. O problema não é só ter IA. É fazer o usuário enxergar valor real nela. Leia em: bitflowtech.com.br/artigo/microso… #Microsoft #IA

⚠️Claude Code's Network Sandbox Vulnerability Exposes User Credentials and Source Code Source: cybersecuritynews.com/claude-codes-n… Anthropic’s Claude Code AI coding assistant harbored a critical network sandbox bypass for over five months, allowing attackers to exfiltrate credentials, source code, and environment variables from developer systems, and the company issued no public advisory for either incident. A second complete bypass of Claude Code’s network sandbox, marking what he describes as a consistent implementation failure rather than an isolated bug. The vulnerability, a SOCKS5 hostname null-byte injection, affected every Claude Code release from v2.0.24 (sandbox GA on October 20, 2025) through v2.1.89, spanning approximately 130 published versions over roughly 5.5 months. #cybersecuritynews

The new @code release adds Mermaid diagram rendering directly in the Markdown preview. The Integrated Browser also now supports HTML file previews and drag-to-select for pulling multiple page elements into chat context.

VMs dedicadas 24/7 são o novo runtime do Gemini Spark. Ao integrar suporte nativo a MCP e processamento persistente, o Google move a IA de janelas efêmeras para identidades digitais autônomas. A infraestrutura de agentes escalou, mas a governança e a auditoria ainda não.

Did you know that if you don't mention AI every 3.7 seconds at Google I/O, security rushes in and escorts you off the stage? @passivelywealthydad/post/DYhWutmkbZz" target="_blank" rel="nofollow noopener">threads.com/@passivelyweal… #googleio #google

Reasoning layers now manage pixel physics in Gemini Omni. By blending logic with Veo, Google prioritize character consistency and lighting over raw motion fluidity. This marks a shift toward functional editing. Does stability matter more than flow in production-grade AI video?

@yeahfortommy @NousResearch Is it to be dedicated full time or is there an opportunity for part time?

looking for someone to help us bring Hermes agent to the world. if you're a hermes agent power user and want to help others become one as well, hit my dms or email us to apply! more info on the forward deployed engineer role: nousresearch.com/forward-deploy…

@elonmusk Como a tradução automática do @grok piorou nos últimos dias. Dos textos simples até os mais complexos. Acho que está precisando fazer em um reset.

The murderous lunatic left

TeamPCP alega ter invadido sistemas internos do GitHub e exfiltrado código-fonte proprietário. Sem confirmação oficial — mas se verdadeiro, é um dos incidentes mais sérios na infra de dev já vistos. A supply chain inteira respira pelo GitHub. Todo pipeline no raio de risco.

🚨 GitHub Source Code Breach - TeamPCP Claims Access to Internal Source Code Source: cybersecuritynews.com/github-source-… A notorious threat actor operating under the alias TeamPCP claims to have breached GitHub's internal systems, allegedly exfiltrating proprietary organization data and source code. According to the threat actor's post, the compromised data encompasses approximately 4,000 private repositories tied directly to GitHub's main platform. In a statement released via X (formerly Twitter), the company acknowledged the unauthorized access but sought to reassure users regarding the safety of customer data. "We are investigating unauthorized access to GitHub’s internal repositories," GitHub stated. #cybersecuritynews

GitHub just confirmed unauthorized access to its internal repositories. Read that again carefully. Not customer repos. Not enterprise orgs. Not your private codebase. Their internal repositories. But here’s the part most people are missing: “We are closely monitoring our infrastructure for follow-on activity.” That line matters more than the breach itself. Because modern attacks rarely stop at initial access. This is usually how it unfolds: 1/ Initial foothold attacker gets access to an internal system, repo, token, or credential 2/ Recon phase map infra, identify secrets, CI/CD pipelines, deployment configs, internal tooling 3/ Privilege escalation move laterally into higher-value systems 4/ Persistence backdoors, stolen tokens, automation abuse, shadow access 5/ Follow-on activity the real attack starts after defenders think they contained it And GitHub explicitly mentioning “follow-on activity” means they know this too. What makes internal repo breaches dangerous: internal repos often contain: infrastructure configs deployment scripts staging credentials internal APIs undocumented services feature flags monitoring hooks employee tooling Even without direct customer data access, internal code can expose the blueprint of the entire system architecture. This is why mature security teams assume: source code access = potential infrastructure intelligence leak Now the interesting part: GitHub said they currently have: “no evidence of impact to customer information stored outside internal repositories” That wording is extremely deliberate. “No evidence” ≠ “impossible” It means: investigation still ongoing logs are being reviewed blast radius not fully finalized containment likely still active internally This is also a reminder for every startup shipping fast with AI and automation: Your GitHub repo is not “just code.” It’s: secrets infra assumptions auth flows deployment logic business intelligence One leaked token + weak IAM + over-permissioned CI pipeline can become a full production compromise in hours. Security isn’t optional once you have users. Even for startups. Even for MVPs. Even for “just testing.”

@github Estou curioso qual será o ressarcimento? Aumentar as taxas do Opus para 50x?

If any impact is discovered, we will notify customers via established incident response and notification channels.

We are investigating unauthorized access to GitHub’s internal repositories. While we currently have no evidence of impact to customer information stored outside of GitHub’s internal repositories (such as our customers’ enterprises, organizations, and repositories), we are closely monitoring our infrastructure for follow-on activity.

Centralizing the world's code creates a single point of failure that no isolation layer can fully patch. The $50,000 price tag on 4,000 internal GitHub repositories signals a shift in targeting. Does code isolation matter when the platform holding it all is for sale?

O terminal virou luxo com o Grok Build. A xAI escala agentes paralelos com execução local e patches diários por 300 dólares mensais. É a troca da estabilidade do SaaS pela força bruta na iteração. Velocidade justifica o prêmio ou é apenas um beta perpétuo pago pelo usuário?

Four letters earned 126k likes while complex marketing cycles yield diminishing returns. This minimalist reset signals a return to core identity after years of brand dilution. Nostalgia buys sentiment but won't solve the content gap. Can vibes sustain the ecosystem until 2026?

Agentes autônomos morrem no gargalo da latência da CPU. O chip Vera da NVIDIA resolve isso com 88 núcleos Arm e o dobro de largura de banda com metade do consumo. Oracle e OpenAI já sinalizam a escala. O hardware removeu a trava técnica ou apenas expôs a fragilidade do software?

1470 Elo. Qwen 3.7 Previews are now the top-ranked Chinese models on LMArena. The architecture isn't a ground-up rebuild but an aggressive optimization of the 3.6 series. It excels in bug fixing yet trails the elite tier. Can refinement alone bridge the gap?