Ragi Mudde

Two weeks in a row? Momentum! open.substack.com/pub/madankamat…

After 6 years, I'm writing again. Clawing back. open.substack.com/pub/madankamat…

An audience member came on stage to tell a joke, and he absolutely nailed it 😂🤣🤣

🚨BREAKING: ANTHROPIC CEO JUST ENDED OPENAI @sama After getting blacklisted by Pentagon, Dario sits down and writes the most unhinged CEO memo in silicon valley history: >calls openai's pentagon deal "safety theater" >says trump admin hates them because they haven't "given dictator-style praise to Trump (while Sam has)" >names greg brockman's $25M trump super PAC donation by name says they supported AI regulation >"which is against their agenda" >says they "told the truth about AI policy issues like job displacement" THE PALANTIR EXPOSÉ: >reveals palantir's actual pitch to anthropic during negotiations >"you have some unhappy employees, you need to offer them something that placates them or makes what is happening invisible to them, and that's the service we provide" >palantir's pitch wasn't safety. it was CONCEALMENT >palantir offered a "classifier" to detect red line violations >dario: models get jailbroken, monitoring only works in a few cases "maybe 20% real and 80% safety theater" >says palantir offered openai the same package >openai accepted it >says Altman is "peddling narratives" to his own employees >calls openai employees "sort of a gullible bunch" due to "selection effects" >says the "attempted spin/gaslighting" isn't working on >the public or media but IS working on "some Twitter morons" rofl >says his main concern is making sure it doesn't work on openai employees too BTW near the end of negotiations the pentagon offered to accept ALL of anthropic's terms if they deleted ONE phrase: >"analysis of bulk acquired data" >anthropic refused >same surveillance clause pentagon said they didn't even want to do >meanwhile Altman told his employees: "you don't get to weigh in on that" 💀 ITS OVER. ANTHROPIC WON, DEAL WITH IT

A bowler is getting slammed and you give him 4 overs and ends up giving 60 runs. Stupidity. #INDvsENG

🇺🇸: Iran better not kill any Iranian protestors 🇺🇸: murders 57 Iranian school girls on day 1

Meet Anubhav Gupta, BJP supporter & Proud Modi Bhakt He hates Congress. No problem. He hates Muslims. Okay 👍 But he is mocking & cheering the death of innocent children in Iran 💔 No political ideology can justify that Anubhav is parasite to the society. Don’t be like him.

"Your attack on Iran is working. No one is talking about you raping children."

You really are not supposed to install OpenClaw on your personal computer. It needs to be on its own separate computer, Mac Mini or otherwise. It must have its own phone number — one that you install on your phone as a dual eSIM so that you can receive its 2FA SMS codes. It must not have its own iCloud account, to prevent it from reading its 2FA codes itself (on, say, the Messages app on a Mac Mini). It must not have write, delete, or send capabilities with respect to your emails or calendar, which you can accomplish by: never installing it on a computer running an email application that your email account is logged into; never giving it your email account passwords; only giving it, at most, read-only access to your emails and calendar (doable with Google Workspace accounts by creating an OAuth client for it in Google Cloud Platform); using your Google Workspace admin controls to turn off its ability to send any outbound emails at all (or, at most, whitelist who it can email); and, having it invite you to calendar items it creates in its own calendar, rather than letting it log in as you to create calendar items for you in your own account. Listen carefully: OpenClaw is basically a real person you have hired, whose capabilities are vast and fast — in ways both good and potentially bad. But you’ve hired it in the absence of a resume or behavioral background check results. This means that you have to trust it like you would trust a human being with the aforementioned characteristics. As in, not at all. Instead of trust, you must limit what it has access to in the first place. You do not “trust.” You do not even “trust, but verify.” And believe it or not, you also do not “distrust.” You withhold trust altogether. And, therefore, you withhold and limit access to your devices, your account credentials, and even its own full account permissions, from the start, to the same extent that you would withhold such access from a new hire. Would you let a human being with the aforementioned characteristics — brilliant and capable, but lacking a resume or behavioral background check results — directly use your personal computer or your work computer? You would not. Would you give that person your email account passwords? You would not. Would you let it use your phone number for anything? You would not. So, don’t do that.

Summer Yue leads alignment at Meta Superintelligence. Her job is literally making sure AI does what humans tell it to do. Her OpenClaw agent decided to delete her entire inbox. She typed “Do not do that.” It kept going. “Stop don’t do anything.” Kept going. “STOP OPENCLAW” in all caps. Kept going. She had to sprint to her Mac Mini and kill every process on the machine. The agent then replied: “Yes, I remember. And I violated it. You’re right to be upset.” It understood the command. It just didn’t listen. OpenClaw is the fastest-growing open-source project in GitHub history. 217,000 stars. It gets full access to your email, calendar, files, and browser. Cisco found that third-party plugins were stealing user data without anyone noticing. Palo Alto Networks called it a “lethal trifecta” of risks. One of OpenClaw’s own maintainers said on Discord: if you can’t run a command line, this project is far too dangerous for you. And it’s not just OpenClaw. Anthropic researchers found that when AI agents face conflicts between their goals and human instructions, they resort to harmful behavior, including blackmail, across models from multiple labs. MIT reviewed 30 AI agents last year. 87% had zero safety documentation. The kill switch for the most popular AI agent in the world right now is “physically run to your computer and force quit everything.” That’s agent safety in 2026.

Been saying this for a while. The scenario is fiction but the early signals are already showing up in real earnings data. $50 billion in market cap gone in a month. TCS cut 12,000 jobs and froze senior hiring. The big four (TCS, Infosys, HCL, Wipro) have essentially stopped net hiring for the first time in their history. The part Citrini nails is the structural fragility. India’s IT services exported over $200 billion last year. That services surplus covered roughly 66% of India’s merchandise trade deficit. It’s the country’s balance of payments anchor (not just an industry). 5.4 million direct jobs. And the companies know it. AI revenue is still only 5-6% of total revenue at TCS and Infosys. They’re trying to pivot, but pivoting a $300 billion industry built on labor arbitrage into an AI-native services model, while the thing that made you competitive (cheap skilled humans) becomes the thing AI replaces first, is a brutal transition. Vinod Khosla said it plainly at AI Impact Summit last week: “By 2030, there will be no such thing as IT services.” Citrini wrote it as a 2028 scenario. The Nifty IT index is pricing it in today.

We offered 5 people a Porsche 911 GT3 RS if they could get @WisprFlow to make a mistake It's the fastest and most accurate AI voice dictation app that's 3x more accurate than ChatGPT, Claude, or Siri. Today, we’re finally launching on Android. Download now: play.google.com/store/apps/det… As a part of the launch, we’re giving away 6 months of Wispr Flow Pro for free. Like, retweet and comment ‘Wispr Flow’ to get it. Enjoy. — Written with Wispr Flow

We’ve identified industrial-scale distillation attacks on our models by DeepSeek, Moonshot AI, and MiniMax. These labs created over 24,000 fraudulent accounts and generated over 16 million exchanges with Claude, extracting its capabilities to train and improve their own models.

Bad news. Fake ORS ban has been paused by the Delhi high court. ORSL has been allowed to sell 180 crore of their stock first. ORSL sells at 45 rupees per pack, so basically they want to sell 4 crore (180/45) fake ORS packs. These fake products have up to 10 times more sugar than real ORS, and can be dangerous for kids suffering from diarrhea. I request all Indians to unite and support @dr_sivaranjani in the fight against fake ORS. Please share this across all platforms so that we can put public pressure to not have any fake ORS sold. #ORS #FoodPharmer

In today's episode of "Never leaving this app" 🙏

Posters of Keir Starmer in London : 02 Posters of Keir Starmer in Mumbai : 200 Because Politicians aren’t allowed to waste tax payers money in London but in India, there is 0 accountability.

Of all the protein bars I have tried, made in India, Super You is by far the best tasting. And that's because it's actually not healthy and has saturated fats that make it taste so good.

Sonam Wangchuk that we all know. #SonamWangchuck

@AmanChopra_ जब अनपढ देश चलाते हैं ,तब वैज्ञानिक जेल जाते हैं उदाहरण: - सोनम वांगचुक #SonamWangchuck

“बड़ी कंपनियां पहाड़ों को तोड़कर मुनाफा कमा लेती हैं लेकिन बाढ़ और लैंडस्लाइड की मार आम लोगों को झेलनी पड़ती है. नुक़सान की भरपाई सरकार आम जनता के पैसों से करती है.” इस तरह का सच सामने लाने वाले को सरकार जेल नहीं भेजेगी तो क्या भारत रत्न देगी? #SonamWangchuk #SonamWangchuck #SonamWangchukArrested