CHERAY

STARHORSE ✨✨✨✨🐎 coming this year; love is forever ✨✨🐎✨✨🌟 Cheray 🌌 ☄️🪐 #starhorse #cheray #ai #space #Texas

Why is taking a goat like that 😒

🚨 Holy shit… Columbia University just dropped one of the most unsettling papers on AI inference I’ve read in a long time. They proved that the entire private AI inference industry built the wrong thing. Prior methods: encrypt the full transformer. 280GB per query. 60-second latency. Enterprise-grade security theater. GPT, Gemini, Qwen, and Mistral independently converged to nearly identical internal representations. One linear equation connects them. > Sub-second inference. 1MB of communication. Same security guarantees. > The private AI inference problem is real. Hospitals can't send patient data to OpenAI. Banks can't send transaction records to Google. Legal firms can't send case files to Anthropic. The solution the industry built: encrypt everything every layer, every attention head, every weight using homomorphic encryption and secure multi-party computation. The result: 280GB of encrypted communication per query. 60-second latency. > Infrastructure costs that make production deployment practically impossible. > Columbia University found the shortcut everyone missed. The Platonic Representation Hypothesis the observation that large models trained on enough data tend to converge toward a shared statistical understanding of the world turns out to be exploitable. GPT, Gemini, Qwen, Mistral, and Cohere, trained independently on different data with different architectures for different objectives, developed internal representations with CKA similarity scores between 0.595 and 0.881. That's not close. > That's essentially the same space. > If the spaces are the same, you don't need to encrypt the model. You learn a single affine transformation one matrix that maps your model's internal representations into the provider's space. Encrypt that matrix. > Send it. The provider runs one linear classification operation on encrypted data and returns the encrypted prediction. You decrypt locally. The transformer never gets encrypted. The weights never get exposed. The query never leaves your control in readable form. > HELIX is the system they built on this insight. During training, the client encrypts their embeddings from public data and sends them to the provider, who computes the alignment map under encryption and returns it. During inference, the client applies the alignment locally, encrypts the transformed representation, and sends it. The provider applies a linear classifier homomorphically and returns the encrypted prediction. > Multiplicative depth of one. No bootstrapping required. 128-bit security by CKKS standard. → Prior methods communication cost: 280.99GB per query (Iron), 25.74GB (BOLT), 68.6GB (MPCFormer) → HELIX communication cost: less than 1MB per query → Prior methods latency: 20-60+ seconds per query → HELIX latency: sub-second → Cross-model CKA similarity: 0.595 to 0.881 across GPT, Gemini, Qwen, Mistral, Cohere → Text generation quality: 60-70% of single-model baseline for high-compatibility pairs → Tokenizer compatibility predicts generation quality with r=0.898 The finding that should end careers: models above 4B parameters with tokenizer compatibility above 0.7 exact match rate can generate coherent text across model families using only a linear transformation. Qwen encoding. Llama decoding. No fine-tuning. No weight sharing. No data transfer. Just matrix multiplication applied to the boundary between two independently trained systems that accidentally became the same thing.

Grok Imagine

Edinburgh’s luxury hotel scene has flourished over the past decade, with new standouts like Gleneagles Townhouse and 100 Princes Street joining longtime grand dames like The Balmoral. trib.al/470BB3u

@bayc8819 @MilkRoadAI Love this ♥️

You can create your own collar using a simple hardware and a raspberry pi monitor. I have it for my four dogs — not paying subscription and I did it using Claude code. Dog collars already have vibration control to guide movement and google geotags can already monitor physical space. $2 billion? And people actually pay for this?

This is WILD. Peter Thiel just bet $2 billion on a collar that wraps around a cow’s neck. The company is called Halter and it has a proprietary algorithm that runs the entire operation. They actually trademarked the name for it and called it the Cowgorithm and here's how it works. A farmer opens an app, taps a button, and 600,000 cows across three countries start walking toward the milking station on their own. No farm dogs, fences or physical labor, it's just a solar-powered GPS collar sending sound and vibration cues to each animal. The collar does more than move cows around. It monitors digestion, fertility cycles, and health patterns in real time, 24 hours a day, using machine learning trained on the behavior of hundreds of thousands of animals. Halter was founded by a rocket engineer who built spacecraft at Rocket Lab before deciding that farming was the bigger unsolved problem. US ranchers alone have already used the technology to build over 11,000 miles of virtual fencing, roughly the full perimeter of the continental United States, saving an estimated $220 million in physical fencing costs. Halter's previous funding round valued the company at $1 billion. This new round, led by Thiel's Founders Fund, doubles that valuation to $2 billion before the new money even hits the account. And they charge farmers between $5 and $8 per animal per month on a subscription model, meaning the more cows they collar, the more locked-in the revenue becomes. The most powerful venture capitalist on earth just decided that the future of food and farming runs through an algorithm named after a cow. He might be right.

@JamesLucasIT ♥️🐾

The ancient Romans were known to create tombs for their dogs and to give them epitaphs to remember them by. One such inscription reads: “I am in tears while carrying you to your last resting place as much as I rejoiced when bringing you home with my own hands 15 years ago.”

@cryptorover @alvinfoo We are already in the US recession 🔥 at least in California

💥BREAKING: Goldman Sachs lifts US recession probability to 30% - WSJ.

@CuriosityonX @star1965 ♥️🐟🌊

This is incredible. This machine is capable of cleaning up 100 million kg of plastic ocean waste, and as of 2025, it has already collected about 500,000 kg of plastic. It aims to remove 90% of ocean plastic by 2040.

@TiffanyFong A perfect story for a Disney movie 🎥 ♥️🐾

Seven dogs stolen from their owners are going viral for escaping their captors and finding their way home. The group was reportedly led by a corgi and traveled 17 km to return to their families. 😭

@DowdEdward ♥️🐂 love this Herd innovation ✨🔥✨

I see…

Private capital firms are starting to swap software systems for hard hats as the artificial intelligence boom forces the industry into a quick rethink of its priorities bloomberg.com/news/articles/…

@feckingm @domdyer70 The woman owned a fancy Tesla but couldn’t even drop her dog at a shelter ?

This is an older case. The defendant signed a confession of judgement to avoid criminal charges. She had to pay fines and give up owning dogs. "Henry’s owner, identified as Maria Bruce, 63, of Portland, has admitted to abandoning him the night of Aug. 6, according to Clark County Animal Protection and Control." I personally feel that she is a complete piece of shit and I shun her. columbian.com/news/2020/aug/…

A woman who’d raised Henry, a 13-year-old Lab, since he was a puppy was caught on camera removing his collar and abandoning him at a Washington park, then driving off in her Tesla.

@feckingm @domdyer70 Never forget this cruelty 💔

@domdyer70 💔

@EvaLovesDesign 🩵

French blue

Under the leadership of @POTUS, @DFCgov will work with @USTreasury to provide up to $20 billion of maritime reinsurance in the Gulf region. This DFC program will – working alongside @CENTCOM – provide shippers hauling oil, gasoline, LNG, jet fuel, and fertilizer with a level of security no other policy can provide.

@SantaCarotaBeef @FFT1776 ♥️🐂🔥 I support you

This is one of the hardest things we have ever had to share. We are not the kind of people who like to ask for help. We have always believed in putting our heads down, working hard, trusting God, and doing everything we can to carry the weight ourselves. But there comes a point where the truth is bigger than pride, and our customers, followers, supporters, and everyone deserves to know what is really happening. Right now, we are in a legal battle with a major meat processor. And while this fight has our name on it, it is much bigger than our family alone. Small producers, family ranchers, and farmers spend generations building something they are proud of, only to come up against an industry that too often protects power over people, profit over principle, and control over transparency. The effects do not stop with the people raising the food. They reach every person purchasing meat because corruption and lack of transparency in the beef industry affect the food system as a whole and the trust families place in what they buy and feed their loved ones. This fight has cost us deeply. Between personal health struggles and the weight of this battle, we have had to make sacrifices we never wanted to make. We have had to cut back on our restaurants and e-commerce. We have sold cattle to help pay attorney fees. We have carried stress, heartbreak, and pressure that, at times, have felt impossible to explain. But we are still here, and we are still fighting. We are fighting for our family, for our ranch, for the values we were raised on, and for every small rancher and farmer who has ever felt crushed under a system that was never built to protect them. So today, we are asking for help. If you believe in family ranches, quality food, hard work, and a more transparent, healthy, and clean food system, please stand with us. One of the best ways you can support us right now is by purchasing our beef at santacarota.com. We started a GoFundMe for those who want to be part of something bigger than our family alone. If you want to help us keep fighting, please consider donating and helping us fight for farmers, ranchers: gofund.me/5f9dbc127

AI is coming to farms. Halter makes AI-powered cow collars. Valued at $2B+, the NZ startup helps farmers: ▫️track GPS location ▫️monitor cow health ▫️draw virtual fences on an app to herd cows via a “cowgirithm” (guide to grazing areas with vibrations and audio cues in collar)