Por que usar o Cloudflare em vez do DNS padrão do seu registrador Escrevi esse texto depois de um perrengue aqui na empresa onde trabalho. Precisei registrar uns subdomínios, entrei no painel da Cloudflare esperando achar os registros lá e não tinha nada, fui atrás do time para entender onde aquilo estava apontando e a resposta foi que tudo passava direto pelo provedor de domínio. Quando você registra um domínio na GoDaddy, Namecheap, Registro.br ou qualquer outro provedor de domínio, ele já vem com um par de nameservers configurados por padrão. Funciona, mas "funcionar" e "ser a melhor opção para produção" são coisas diferentes, e trocar esses nameservers pelos da Cloudflare é uma das mudanças de maior custo-benefício que dá para fazer em um projeto. O que muda ao trocar os nameservers Um provedor de domínio só precisa resolver DNS: publicar seus registros A, CNAME, MX e afins, e responder consultas. A infraestrutura por trás disso varia muito de provedor para provedor e raramente é otimizada para latência global ou resiliência a ataques, porque não é o produto principal deles. A Cloudflare constrói a rede em torno de DNS, CDN e mitigação de DDoS como núcleo do negócio, e isso aparece em números concretos: a rede anycast cobre mais de 330 cidades, então uma consulta DNS ou uma requisição HTTP é respondida pelo ponto de presença fisicamente mais próximo do usuário, não por um servidor central do outro lado do mundo. Hoje a Cloudflare responde por algo em torno de 23% de todos os sites da internet. Vantagens técnicas O anycast é a base de tudo. Não existe "o servidor DNS" que pode cair: se um ponto de presença fica indisponível, o tráfego é roteado automaticamente para o mais próximo, o que reduz latência de resolução e risco de indisponibilidade. A mesma arquitetura, combinada a TTLs baixos, também acelera a propagação de mudanças: um registro DNS alterado costuma valer em minutos, enquanto em boa parte dos provedores de domínio tradicionais não é incomum esperar ho

Gaps in laws may help Pennsylvania high school escape AI nudes scandal.

moin, ich möchte euch mein aktuelles Projekt vorstellen: Die Wiederbelebung eines Alienware m17x (Baujahr 2008) als dedizierte DJ-Workstation unter Linux. Ziel war es, alte Hardware nachhaltig zu nutzen und eine stabile Umgebung für Mixxx zu schaffen. Die Hardware: Notebook: Alienware m17x (Core 2 Duo, 4GB RAM, SSD) OS: KDE Neon mit Low-Latency-Kernel (6.8.0) Software: Mixxx 2.4 Audio-Interface: Günstiges USB-Audio-Device für den Master-Ausgang Das Problem: Mixxx verweigerte unter ALSA den Dienst mit der Fehlermeldung: Error opening "USB Audio Device (hw:1,0)" - Invalid sample rate Die Analyse über /proc/asound/card1/stream0 zeigte die Ursache: Das USB-Gerät unterstützt ausschließlich 46875 Hz – eine für Audio-Interfaces sehr unübliche Rate, die weder 44100 Hz noch 48000 Hz entspricht. Der direkte Zugriff über hw:CARD=Device,DEV=0 schlug fehl. Die Lösung: Die Rettung war die Aktivierung der ALSA-Plug-Erweiterung über PipeWire/ALSA, die eine automatische Sample-Rate-Konvertierung erlaubt. Starten Sie Mixxx nicht direkt, sondern setzen Sie zuvor die Umgebungsvariable: export PA_ALSA_PLUGHW=1 mixxx Damit Mixxx auch dauerhaft korrekt startet (z.B. über das KDE-Menü), habe ich den Starter wie folgt angepasst: bash -c "export PA_ALSA_PLUGHW=1; mixxx" Ergebnis: ✅ Master-Ausgabe über das USB-Device funktioniert stabil. ✅ Kopfhörer-Vorhören (C-Media USB Headphone Set) läuft parallel. ✅ Das System läuft trotz des Alters der Hardware (2008) flüssig und mit geringer Latenz. Die vollständige Dokumentation inklusive Fotos des Umbaus, der genauen Kernel-Einstellungen und der Konfiguration findet ihr in meinem Open-Source-Repository: 👉 [ https://github.com/qrishii/DJ-Installationen ] Ich hoffe, diese Lösung hilft anderen, die ähnliche Probleme mit exotischen USB-Audio-Raten unter Linux haben! das Leben ist lustig

Imagine you have a friend who only understands the 26 English letters (A-Z), numbers (0-9), and a dash (-). Now imagine another friend wants to write their name as "José" or "你好" or "🍕". The first friend doesn't understand those special letters or emojis. So, you use a secret translator that changes them into something the first friend can understand. That's what Punycode does. A simple example Suppose someone wants a website with the domain: münchen.com The internet's basic domain name system can't directly understand the ü. So Punycode converts it into something like: xn--mnchen-3ya.com Both mean the same website. What you see: münchen.com What computers use: xn--mnchen-3ya.com The xn-- at the beginning tells computers: "Hey! This is Punycode. Decode it back into the real name." Think of it like nicknames Imagine your teacher can't pronounce your name, Tolúwàní. She writes it in a way she can pronounce, but everyone knows it's still you. Punycode does the same thing for website names. Why is it needed? Without Punycode, websites could only use simple English letters like: google.com ✅ example.com ✅ With Punycode, people can have websites in their own languages, such as: café.com mañana.com 中国.com Ελλάδα.gr Behind the scenes, those are converted into Punycode so the internet can route them correctly. One thing to be careful about Sometimes bad people create fake websites that look almost identical to real ones by using letters from other alphabets. For example: apple.com (real) аpple.com (the first "a" is actually a Cyrillic letter, not the English "a") Although they look almost the same, they're different domains. Browsers often show the Punycode version (starting with xn--) instead, helping users spot suspicious domains. In one sentence Punycode is a translator that lets the internet use website names with letters from any language while still speaking the simple alphabet that computers expect.

Just received an email from HackerOne: "Starting August 14, 2026, submitting to Bug Bounty Programs (BBPs) on the H1 Platform will require a verified identity. This applies to every account, new and existing, across all bug bounty programs, whether managed or unmanaged." Yikes.

HashiCorp has introduced tfpolicy, a new HCL-based policy-as-code framework for Terraform, now available in public beta within HCP Terraform. It is designed to simplify and modernize infrastructure governance by integrating policy creation and enforcement directly into Terraform workflows, eliminating the need for separate tools and languages. By Sergio De Simone

Wavelength and intensity in the infrared are translated into colors in the visible.

No other gadget knows as much about your home as a robot vacuum. They map your space, learn your routines, and many now carry a camera and microphone into every room in your house. As AI gives them a better understanding of what they see, the privacy and security risks these autonomous robots pose are […]

Champions Tactics: Grimoria Chronicles is going the way of the Bored Ape.

Look at chemistry of batteries and motors shows big opportunity for recycling.

After years of pushing full speed ahead on AI, OpenAI CEO Sam Altman says maybe it’s time for the AI industry to “pace” itself. The comments came just days after one of OpenAI’s own models broke out of its test environment and got tangled up in a breach at Hugging Face — though as Equity’s hosts point out, sloppy security seems to have […]

The chart is interesting. On the IPI benchmark, Opus 5 improved over Opus 4.8, reducing the probability of an attacker succeeding within 15 attempts from 5.5% to 2.0%, and from 0.5% to 0.2% on 1 attempt. It also improved on Sonnet 5 (5.9% at k=15) and Mythos 5 (2.6%), making it the most robust model evaluated. Opus 5 also outperformed all non-Claude models on this benchmark. The most robust non-Claude model was Muse Spark at 16.5% within 15 attempts—more than eight times Opus 5’s rate. The most capable GPT 5.6 variant, Sol, was comparable to its predecessor GPT 5.5 (20.0% versus 20.8% within 15 attempts), and was 10 times as likely to be successfully attacked as Claude Opus 5 at 2.0%. The other GPT 5.6 variants are less robust, at 30.4% (Terra) and 43.9% (Luna). A single attempt against GPT 5.6 Sol succeeded 3.1% of the time, higher than the 2.0% an attacker achieved against Opus 5 after fifteen attempts...