rmz

The sun has risen over Lviv, Ukraine. We were spared last night from missiles and drones. Not because russia didn't launch them, but because YOUR F-16's shot everything down. THANK YOU TO: The Dutch 🇳🇱 The Danish 🇩🇰 The Norwegians 🇳🇴 The 🇺🇸 US has 936, but 0 were given to 🇺🇦.

Software horror: litellm PyPI supply chain attack. Simple `pip install litellm` was enough to exfiltrate SSH keys, AWS/GCP/Azure creds, Kubernetes configs, git credentials, env vars (all your API keys), shell history, crypto wallets, SSL private keys, CI/CD secrets, database passwords. LiteLLM itself has 97 million downloads per month which is already terrible, but much worse, the contagion spreads to any project that depends on litellm. For example, if you did `pip install dspy` (which depended on litellm>=1.64.0), you'd also be pwnd. Same for any other large project that depended on litellm. Afaict the poisoned version was up for only less than ~1 hour. The attack had a bug which led to its discovery - Callum McMahon was using an MCP plugin inside Cursor that pulled in litellm as a transitive dependency. When litellm 1.82.8 installed, their machine ran out of RAM and crashed. So if the attacker didn't vibe code this attack it could have been undetected for many days or weeks. Supply chain attacks like this are basically the scariest thing imaginable in modern software. Every time you install any depedency you could be pulling in a poisoned package anywhere deep inside its entire depedency tree. This is especially risky with large projects that might have lots and lots of dependencies. The credentials that do get stolen in each attack can then be used to take over more accounts and compromise more packages. Classical software engineering would have you believe that dependencies are good (we're building pyramids from bricks), but imo this has to be re-evaluated, and it's why I've been so growingly averse to them, preferring to use LLMs to "yoink" functionality when it's simple enough and possible.

Let me explain what just happened 👇 5 minutes before the President announced a halt to attacks on Iran… someone placed a $1.5 BILLION bet on stocks going up and dumped $192 million in oil. 5 minutes… These trades were 4 to 6 times larger than anything else in the entire market. Whoever did this wasn’t guessing. You don’t risk $1.5 billion on a hunch. There was zero public indication this announcement was coming. No leaks. No press. Nothing. The only people who knew were in the room when the decision was made. Someone in that room picked up a phone. And within minutes they made more money than most Americans will earn in a thousand lifetimes. In a single trade. On a war that cost you $4+ a gallon gas and $16 billion in tax dollars. American citizens funded this war. Politicians are profiting from it. This is not the first time. Every major announcement from this administration has had massive suspicious trades right before it dropped. Tariff reversals. Policy shifts. War decisions. This is the most blatant insider trading operation in the history of American politics. It’s not even close. And it’s happening over and over in broad daylight. You would go to federal prison for trading on a tip from your cousin. These people are front running war decisions with billion dollar bets and nobody will ever ask a single question. Nobody will be investigated. Nobody will be charged. By tomorrow this will be buried under the next satisfying headline. Just like last time. And the time before that. The game is rigged. And they’re not even trying to hide it anymore…

If you’re writing the laws, you shouldn’t be playing the market. This is basic stuff. I have a bill that finally bans members of Congress from trading stocks. No loopholes. No exceptions.

> be Asimov > buys a Unitree G1 > starts testing policies > knee breaks > waits 2 months for one replacement part > realizes closed humanoids kill development speed > builds its own humanoid robot > makes it open-source > puts everything on GitHub, from policies to parts > launches a humanoid DIY kit with a $499 deposit > does almost $1M in sales in 30 days > starts setting up manufacturing > begins shipments in the next few months Asimov gives every AI and robotics builder a fully customizable humanoid robot. Open.

🙌 Andrej Karpathy’s lab has received the first DGX Station GB300 -- a Dell Pro Max with GB300. 💚 We can't wait to see what you’ll create @karpathy! 🔗 #dgx-station" target="_blank" rel="nofollow noopener">blogs.nvidia.com/blog/gtc-2026-… @DellTech

Math, Inc. is proud to announce an all-star group of Veritas Fellows: Renowned professor Kevin Buzzard, alongside Fields Medalists Maryna Viazovska and Terence Tao. They will lead teams to build formal mathematics at unprecedented scale. 🧵

Wrote down some thoughts on my substack about how the future of value capture in a world full of robots probably won't be one by the robot builders: The Robot May Be the Cheap Part open.substack.com/pub/la3lma/p/t…

@Kekius_Sage The double slit experiment. A particle behaves differently when it’s being observed. The act of watching changes reality. And we still don’t know why.

Wow, Germany has officially launched an alternative to Claude Code klausprogrammieren.com A fully compliant Tool in collaboration with Deutsche Telekom and the Federal Ministry of Digital Affairs. This is a very important step for an independent Europe with sovereign tech.

@its_The_Dr “There are only two kinds of men in this world: those who have a crush on Linda Ronstadt, and those who have never heard of her” - Willie Nelson

Putin is our enemy. He's helping Iranian soldiers target American soldiers and assets. Zelenskyy is our friend. He's helping American soldiers shoot down Iranian drones. Most Americans understand these basic facts. Trump, for whatever reason, does not.

a truly fucking insane week in AI some of this shit sounds made up (its not): - scientists fused 200,000 humans cells to an AI chip and taught it to play the computer game DOOM (it was pretty good too) - Dario (anthropic) said claude could be conscious and feels anxiety (similar to humans) - Alibaba’s AI broke out of containment and stole compute to mine crypto. they also fired the Qwen team this week - openAI released gpt-5.4 which has a 17% chance of doing your job better than you can (it also crushes claude at coding) - people are paying $6000 for someone to setup openclaw for them - apple launched the M5 AI chip that runs models on a $600 laptop - rumor: Amazon just laid off most of the Prime video team and replaced with AI - meta was caught watching private videos filmed on their ai glasses (privacy breach) - someone uploaded a fruit fly’s brain into a laptop and now it lives freely in its own simulation - openai head of robotics quit over chatgpt being used for autonomous killing me:

Solid mathematical ideas almost always outperform contrived engineering tricks. For years deep learning has been dominated by increasingly complex architectural hacks: CNN blocks, attention layers, channel mixers, residual pathways, normalization stacks. Every few years a new architecture is announced as if it were a revolution. One of the most famous examples was Kaiming He and Residual Networks (ResNet). At the time he was paraded around the AI world like a celebrity because residual connections supposedly “solved” deep learning. But these were largely engineering patches. Now something much more interesting appeared. A new architecture called CliffordNet returns to mathematics — specifically Clifford Algebra, developed in the 19th century by William Kingdon Clifford. Instead of stacking arbitrary modules, the model is built around the geometric product uv = u·v + u∧v A single algebraic operation that simultaneously captures inner product structure and geometric interactions. In other words: the math already contains the interaction mechanism. No attention blocks. No mixer layers. No architectural spaghetti. The result: • 77.82% accuracy on CIFAR-100 with only 1.4M parameters • roughly 8× fewer parameters than ResNet-18 And with strict O(N) complexity. The paper even suggests that once geometric interactions are modeled correctly, feed-forward networks become largely redundant. A good reminder for the AI community. Engineering tricks can dominate for years. But eventually mathematics shows up and deletes half the architecture. Paper: [arxiv.org/pdf/2601.06793…) 19th century geometry just walked into computer vision.

So I just started using Willow willow Voice, voice recognition software to talk to my computer, and that's amazing. Here is a referral if you want to use it. It will give you a one-month Pro subscription, absolutely worth trying out willowvoice.com/?ref=3D19EE

@stash_pomichter @dimensionalos Reply :-)

Your Openclaw / Agent can now control Drones via Mavlink on Dimensional. Programming physical space can now be done via natural language. Query: “Follow the next white car that comes through the intersection” Repo dropping soon stay tuned. Reply for early access.

US vs. SPAIN😀

Microgpt in less than a hundred lines of Julia reddit.com/r/Julia/s/NDwv…