Solana is probably going to experience a black swan now that its security through obscurity approach no longer works. Normalizing complete lack of code transparency was never a good idea.

Introducing 𝙻𝙵𝙼𝟹: our new generation of Liquid Foundation Models. With a new tri-hybrid architecture and new multimodal encoders, the LFM3 family is the ideal model suite for specialized intelligence. > 6 model instances, across text, audio, vision, and omni modalities, generative and decision models. > 𝙻𝙵𝙼𝟹-𝟷.𝟽𝙱: text, 256k context window. > 𝙻𝙵𝙼𝟹-𝚅𝙻-𝟸𝙱: vision-language model with video-input support for spatial understanding > 𝙻𝙵𝙼𝟹-𝙰𝚞𝚍𝚒𝚘-𝟸𝙱: tool calls, voice chat, reasoning in 18 languages. > 𝙻𝙵𝙼𝟹-𝙾𝚖𝚗𝚒-𝟸𝙱: an omni model combines vision, audio, and language. > 𝚍𝟷-𝟸𝙱 and 𝚍𝟷-𝙾𝚖𝚗𝚒-𝟸𝙱: decision models with typed outputs and calibrated probabilities in a single pass. Omni adds audio input. > fine-tune on 1 GPU. > a new tri-hybrid architecture > architecture optimized for 14 processor classes across GPUs, CPUs, and NPUs. > 𝙻𝙵𝙼𝟹-𝚅𝙻-𝟸𝙱 is a beast. It performs competitively with Gemma 4 12B (6x larger), and Gemini 3.5 Flash Lite, and GPT-5.6 Luna, with 1.9B parameters. > 𝙻𝙵𝙼𝟹-𝟷.𝟽𝙱 is #1 on the Pareto frontier of device intelligence 𝚙𝚒𝚙𝚎𝚝𝚝𝚎.𝚕𝚒𝚚𝚞𝚒𝚍.𝚊𝚒 Every modality for specialized intelligence. Pick the one that fits your use case, customize it, and run it on-device. 1/ 🧵

I understand now why people are so deeply anti- YouTube Kids It's also now completely and utterly ravaged by slop. These kids are cooked man

AI is horrible at game design

Anthropic IPO in 3 weeks. If only there were a proper mascot with infinite reach who's an absolute star with an army of people making content using her downloadable soul.

Ledger Nano and Stax teardown Nano was so easy; 4 minutes to open and close. Stax is hard to tamper without damaging case. Heatgun maybe. There are so many exposed and labeled testpoints and comms interfaces you can tap. No wonder these are getting pwned.

Proof REA is is malicious. This repository is simultaneously: (1) the vehicle for an active social engineering campaign harvesting credentials and biometric data from users who install it, (2) the subject of apparent artificial star inflation to manufacture legitimacy, and (3) the host of an architectural vulnerability enabling prompt-injection-to-RCE in AI agent workflows. The combination of pre-positioned @latest install instructions and a coordinated amplification network represents a complete supply-chain attack delivery mechanism should a malicious package version be published.

life is a journey

Chomsky was wrong. Researchers found your brain doesn’t store language the way every textbook since 1957 claims. For decades, the undisputed rule of linguistics, led by Noam Chomsky, was that human language is hierarchical. We were told our brains process language using complex "grammar trees." That we secretly build deep, abstract, mathematical rules in our heads before we speak. This new paper proves that is completely wrong. Researchers at Cornell University analyzed how people actually process and produce language in real time. They found sequences. Linear chunks. When we speak, we aren't calculating deep syntactic rules. We are just predicting the next most logical sequence of words based on the context of the words right before it. We string linear chunks together. If this sounds familiar, it should. This is exactly how Large Language Models like ChatGPT work. For years, critics have dismissed AI as "just a next-token predictor." They said it isn't truly intelligent because it doesn't understand deep grammatical hierarchies like humans do. But this paper suggests humans don't use those hierarchies either. We are just predicting the next chunk. Language isn't a rigid tree of mathematical rules. It's a linear stream of contextual predictions. We thought we were building AI to mimic the complexity of the human brain. It turns out, the human brain was just running a simpler algorithm all along.

Elegantly crafted. Seamlessly contoured. Thoughtfully detailed. motorola signature 27 is designed to a higher standard. #MotorolaSignature #SignatureStatement

Horas soldando y al final… sigue sin mandar video 😔😩

GPT-6.1 Terra substantially outperforms GPT-5.6 Terra on Agents’ Last Exam, a test of how well agents use software at work. It even matches GPT-6 Astra at one-tenth the price per token and finishes tasks over 3× faster on average than both Astra and GPT-6.1 Sol.

I've watched kids fall off of these. They don't fall far

keystone essential internals. I was too lazy to dig out my heat gun and in this case, that was a terrible mistake. don't know if I will post a video. it's a bit of a disaster. have some pictures

I dug deeper into Gigabyte's crazy 9-GPU WRX90 AI TOP and found the fine print. 6× PCIe 5.0 x16 👈 👀 1× PCIe 5.0 x8 All directly from the Threadripper Pro. GPUs #8 and #9 can come in through the board's two MCIO 8i connectors. Here is the catch tho One of those MCIO connections shares CPU lanes with both PCIe 5.0 M.2 slots. So you can populate either of those Gen5 M.2 slots → that MCIO connector is disabled. 🎯 So this really is a workstation built around a choice you have to make, do you want maximum local NVMe, or maximum accelerator count Gigabyte has also now put the full support stack online, including a new BIOS and Windows 11 drivers. Still no price or firm retail date.

Claude rejected my request to debloat my Samsung TV but DeepSeek-4.1-Flash-1017, the normal one, not the uncensored one, happily took care of it for me. Why is Anthropic like this?